Aircon Thermostat vs Thermistor: What a Split System Has

A thermistor is a sensor and nothing else. Its electrical resistance changes as its own temperature changes, so a control board can read that resistance and turn it into a number. It reports only, and holds no view on what should happen next.

A thermostat, in the sense most people mean it, senses and then acts. The familiar version is a wall box with a dial or small screen. You set a target on it, it measures the air in front of it, and it switches the system on and off to hold that target. Setting, sensing and deciding all sit inside one object.

Anyone who has lived with a wall box in an office, a hotel room or a house overseas learned that the temperature is controlled from one place you can walk up to. Bring that picture into a Singapore flat and it stops matching the hardware on the wall.

Residential splits here keep the three jobs apart and put them in three different places. Nothing mounted on the wall does what a thermostat does, because those functions were never collected into a single part.

The mismatch would be harmless if it stayed a naming quirk. It does not. Someone convinced the thermostat has failed goes looking for one, finds listings for wall units that suit colder countries, and arrives with a part in mind that will never be fitted. The real candidates were never on the shortlist.

The setpoint lives on the remote, and the remote is a transmitter rather than a controller. Pressing a button sends an infrared instruction across the room, and the handset screen shows what it sent. The handset has no authority over the compressor. Take the batteries out and the unit carries on at the last setting it received.

The sensing happens at thermistors buried in the machine. One sits in the indoor unit's return air path, one is clipped to the indoor coil, and at least one more is fitted outdoors. Each reads a different temperature for a different reason, and not one of them sits where you sit.

The deciding happens on the control board. It stores the target it was last sent, reads what the sensors report, and works out what the compressor and the fans should do about the difference. Every start, stop and speed change comes out of that comparison.

So the three functions a thermostat bundles into one box are, on a split, in your hand, inside the indoor unit, and on a circuit board behind the front cover. A fault in any of the three produces a room that will not behave. They are separate parts with separate failure modes, and one word hides which is being discussed.

Room temperature alone cannot run a machine safely, which is why several thermistors are fitted. The air in the room answers one question: whether to cool. It says nothing about whether the equipment is fit to keep running, which needs readings taken inside the hardware.

The return air thermistor is the one closest to a thermostat's job. It sits in the path of air drawn back into the indoor unit, and reports what that air is at. The board holds that figure against the target and decides whether cooling continues. On units that display a room temperature, this is the sensor behind the number.

The coil thermistor watches metal rather than air. An evaporator runs below room temperature by design, and if airflow drops or refrigerant runs low it keeps dropping until moisture on the fins turns to ice. A reading taken off the coil lets the board ease back before that happens, a protective call the room reading could never trigger.

The outdoor sensors report the conditions the heat has to be pushed into. Outdoor air temperature tells the board how hard the system must work to shed heat, and a sensor on the compressor discharge line watches for the machine cooking itself. That last one exists to stop the unit, which is the reverse of what the room sensor exists to do.

A thermistor rarely fails in a way you can see. Its wiring can go open or short, and most boards catch that and put a code on the display, because a value outside any believable range is obviously not a temperature. The harder failure is the one that stays believable.

Drift means the sensor still reports, still reports something plausible, and is wrong. A sensor reading 3°C cooler than the air around it hands the board a figure the board has no way to challenge. It does exactly what it should with the information it was given.

The system therefore behaves logically and the room still ends up wrong, with nothing on the display to account for it. The unit reaches a target it never actually reached, eases off, and leaves the room short. Or it chases a room already at temperature and runs on well past the point it needed to.

That silence is what lets a sensor get skipped and over-claimed in the same week. A quote for a compressor or a control board on a unit that cools in bursts deserves a simpler question first. Has the reported temperature been compared against a separate thermometer sitting in the same air? Until that comparison exists, the sensor has not been ruled out, and the expensive parts it imitates have not been ruled in.

The same discipline runs the other way. A sensor swap proposed before the filter and coil have been looked at is the mirror image of the same shortcut. A restricted air path cuts a unit off early with a sensor that is reading perfectly, and fitting a new sensor changes nothing about a blocked one. Ask which was checked first, and in what order.

Wall controllers do exist, and they turn up where the system is too large or too concealed to run from a handset. Ducted systems, cassettes in offices and shops, and most commercial fitouts get a wired panel mounted at eye level. That panel is what people are picturing when they say thermostat, and there the word is closer to the truth.

The panel is still not the whole of a thermostat. It is a wired interface with a temperature sensor built into it, talking to a control board in the unit above the ceiling. It carries the setpoint, and often a sensing point as well. The deciding continues to happen somewhere you cannot see.

The difference earns its keep the moment something stops working. A blank panel is not automatically a dead panel, because it depends on power and a data link coming from the indoor unit. A panel that lights up but will not start the system points somewhere different from a panel that stays dark. Both sound identical on the phone.

Landed homes and larger condos running concealed ducted units belong in this group too. Anyone with one should say so up front, because the diagnosis, access and parts differ from a wall-mounted split.

A wired panel also moves where the room gets measured. Sensing at the panel means the system targets air at chest height near a wall. That is a fairer sample of an occupied room than a grille near the ceiling. A panel sitting beside a door, above a printer, or on a wall that takes afternoon sun feeds the board a local reading and calls it the room.

Naming a part narrows the diagnosis before anything has been measured, and the name is usually inherited rather than tested. Someone reporting a broken thermostat has described a room that will not hold temperature, which is accurate and useful. The part attached to that description is a guess, and on a split a guess about hardware that is not fitted.

What the room actually did is the part nobody else can reconstruct later. Note when the unit stops, and whether it stops early or refuses to stop at all; note what the display claimed the room was, and what a hand-held thermometer showed at that same moment. Note whether anything in the flat changed around the time the behaviour started. Those details survive the retelling, and a part name quietly replaces all of them.

The same complaint carries very different weight depending on which form it arrives in.