Set temperature vs room temperature: what the gap means
The number on the remote is a target for one sensor, not a promise about the air around you. That sensor sits high on a wall and reads air the room has already finished with. A large share of not-cold-enough complaints lives in that gap.
By Team Snowflake | Updated 5 Aug 2026
Where the machine takes its reading
The number on the remote is a target for a sensor, and the sensor is nowhere near you. It is a thermistor inside the indoor unit, sitting in the stream of air being pulled back in. It reads the air arriving at the machine, and the control board weighs that reading against the target.
Every decision the unit makes follows from that one reading. When the returning air is warmer than the target, the board calls for cooling. When it matches, the board eases the compressor back or stops it. Nothing in that loop knows where you are sitting, what the floor feels like, or whether the sofa is in the sun.
A wall unit is mounted high because that is where it can draw air freely, and the sensor goes where the air is. The machine therefore forms its idea of the room at the top of the room. On a cassette recessed into a ceiling the sampling point is higher still, and on a ducted system the sensor can sit in a return grille somewhere else in the flat.
The room temperature shown on many displays comes from that same sensor. It is an honest reading of the air at the intake grille. It was never a claim about the air at your pillow, and the distance between those two places is what the rest of this comes down to.
The room is not one temperature
A closed room separates into layers, warmer near the ceiling and cooler near the floor. That separation is ordinary and it happens in every flat. Anyone who has stood up from the sofa on a warm afternoon has felt the change across the height of their own body.
The layer the unit serves best is the layer it measures. Cold air leaves the louvre near the ceiling, runs along it, and finds its way back to the intake a short distance away. The upper band of the room gets the most contact with that stream. The lower band, where people sit, sleep and work, is served last and sampled never.
So the sensor can arrive at the target while the occupied half of the room is still behind it. The board reads a satisfied room, eases the compressor down, and lets the airflow drop. The lower band stops improving at that moment. Nothing has failed. The machine finished the job it was able to measure.
Mixing is what closes that gap, and mixing does not happen on its own. A ceiling fan, a higher fan speed on the unit, or a louvre angled down into the room rather than flat along the ceiling all pull the two layers together. The reading then describes the room instead of one slice of it.
When the unit reads air it just cooled
A sensor sitting in the unit's own discharge produces the same complaint in a sharper form. If the louvre is angled so the stream curls straight back into the intake, or a curtain or high shelf sits close enough to bounce it back, the sensor reads air the machine chilled seconds ago. The rest of the room never enters the calculation.
The signature is a unit that stops early and often while the room stays warm. It runs hard, cuts out sooner than it used to, goes quiet, then starts again. A unit that began behaving this way straight after a service visit is worth mentioning, because a sensor knocked out of position during cleaning reads the wrong air from then on.
The same number does not feel the same twice
Air temperature is one of several things your body is adding up, and it is not the loudest of them. Two evenings at an identical setting can feel completely different, with the sensor reading correctly on both. That is comfort behaving normally, not an instrument lying.
Moisture is the largest of the other variables. Sweat carries heat off skin, and damp air slows that down, so 24°C on a wet evening does not land the way 24°C does after a dry stretch. Indoor humidity is a subject of its own with its own targets, and it deserves reading separately rather than folded in here.
Air movement shifts the same number again. Moving air strips heat from skin faster than still air, which is why a fan makes an unchanged setting feel cooler. It is also why a room feels warmer the moment an inverter settles down and the airflow drops, even though the temperature has not moved at all.
Surfaces work on you separately from the air. A window that took the afternoon sun keeps radiating into the room long after the sun has gone, and so does a wall that backs onto an unshaded corridor. The sensor cannot feel any of that. Your skin can, which is why a seat near glass feels warmer than a seat further in at the same air temperature.
| What changed in the room | What the sensor sees | What you feel |
|---|---|---|
| What changed in the roomFan speed dropped to low | What the sensor seesNo change | What you feelWarmer, because less air is moving over skin |
| What changed in the roomWet evening after heavy rain | What the sensor seesNo change | What you feelWarmer and stickier at the same number |
| What changed in the roomAfternoon sun landing on the glass | What the sensor seesRises slowly, well after the glass is hot | What you feelWarmer on the side facing the window |
| What changed in the roomCeiling fan switched on | What the sensor seesFalls as the layers mix | What you feelCooler almost straight away |
| What changed in the roomTwo more people in the room | What the sensor seesRises slowly | What you feelWarmer before the number moves |
Loads the setpoint cannot see
A setpoint is a target, not a capacity. The unit removes heat at whatever rate it can manage, and the room takes heat on at whatever rate the day dictates. Where those two rates meet is the temperature you actually get, and the remote has no vote in that meeting.
Sun through glass is the largest load in most Singapore flats, and it arrives on a schedule. A west-facing living room gains heat all afternoon and hands it back through the evening. The same room at the same setting is a different problem in mid-afternoon than it is late at night.
People and appliances supply the rest. Each person in the room is a steady heat source, and so is a television, a laptop on charge, an oven that was used before dinner, and every downlight in the ceiling. A doorway open to a kitchen in use feeds warm air in about as fast as the unit takes it out.
The open door is the one to test first, because it is the largest of these and the easiest to remove. A room open to a corridor, a stairwell, or an unconditioned kitchen is not the room the unit was sized for. Shut it, let the unit catch up, and watch whether the number and the feel both move.
The pattern names the load
Loads announce themselves by when they show up. A room that only disappoints in the afternoon is losing to the sun. A room that only disappoints when the whole family is home is losing to occupancy. A room that disappoints in every condition, including a closed bedroom at night with nothing running in it, is not losing to a load at all.
That last case is the one worth reporting. Note what the remote is set to, what a thermometer beside you reads, and what was going on in the flat at the time. The pattern is the diagnosis. It cannot be seen in a single reading, and it is the part no technician can reconstruct after the fact.
What separates normal physics from a fault
Three patterns move this out of physics and into diagnosis. Everything above assumes the machine reaches its target and then eases off. When it never gets there, or gets there dishonestly, or used to get there and no longer does, the room stops being the explanation.
The first is a unit that never reaches the setpoint at all. Not slowly, not only under the afternoon sun, but never, including overnight in a shut bedroom with no sun and nobody else in it. The compressor runs on and the reading sits above the target. Capacity is missing, and the reason for that sits inside the machine. Confirming it on site means comparing the air the unit produces against the air it draws in, which is a separate measurement with a method of its own.
The second is a reading that contradicts a thermometer in the same place. Put a separate thermometer near the intake grille, let it settle, and compare it against what the unit says the room is. Agreement means the sensor is honest and the argument is about comfort. A wide disagreement either way means the board is working from a false number, and adjusting the setpoint cannot correct that.
The third is behaviour that changed on a unit that used to hold the room fine. Same flat, same furniture, same afternoon sun, different result. A room that has not changed paired with a unit that has is the cleanest signal available without tools, and it carries more weight than any single reading taken today.
If a gas top-up is proposed on a comfort complaint alone, ask what was measured first. Low refrigerant shows up as a unit that runs and runs without reaching the target. It does not show up as a room that reaches the target and still feels warm. A unit hitting its setpoint and easing off on cue has not lost gas, and adding some will not change what the room feels like.
Setting for comfort instead of chasing the number
Aim the air before touching the setpoint. A louvre angled flat along the ceiling serves the layer nobody is in. Angled down into the room, the same stream reaches the occupied half, stirs the two layers together, and hands the sensor a reading that describes the whole space. Where the unit has a swing function, sweeping mixes better than any fixed angle.
Fan speed is the lever most households never touch. A higher fan speed moves more of the room past the coil and past the sensor, which spreads the cooling and makes the reading more honest at the same time. It is the right answer to a room that feels stuffy at a number that looks correct. Auto fan hands that decision back to the machine, and the machine is optimising for the layer it can measure.
A ceiling fan running alongside the aircon does the same work for less. It moves air over skin, which lowers what any given temperature feels like, and it breaks the layers up so the lower half of the room stops lagging behind. Most households find they can sit comfortably at a higher setting with the fan on than without it.
Look at what sits around the indoor unit as well. A tall wardrobe, an open shelf, or a curtain hanging close to the unit can starve the return path or bounce the supply straight back into it. Leave the intake face clear air to draw from, and leave the discharge path open for as far as the stream has to travel.
Then set for the room you are in rather than the number you had in mind. There is no reward for holding a particular figure, and asking for a lower one does not produce colder air. The unit makes the same cold air at every setting. A lower setpoint only makes it run longer before it eases off, which shows up on the bill and does nothing about the layers or the moisture.
The setting that works is the one where the room feels right with the air moving. Find it once and note it down. Treat any change in it as information rather than preference, because when the number that used to be comfortable stops being comfortable, something has changed, and that is the observation worth passing on.
Ready to get started?
Tell us what’s going on. Symptoms, setup, photos, anything we should know. We’ll assess and come back with the right next step.