Dew point: the temperature that decides where water forms

The dew point is a temperature, and it belongs to the air rather than to any wall it is measured near. It names the point at which that air is carrying every bit of water vapour it can hold. Cool the air down to that figure and it starts handing water back. Cool any surface to that figure and the air touching the surface does the same thing on it.

A percentage moves all day. This figure does not. Warm a sealed room and the percentage falls, cool it and the percentage climbs, while the water in the air never changed by a gram. The dew point ignores all of that. It shifts only when moisture is genuinely added to the air or taken out of it, which is what makes it the reading that predicts behaviour.

That stability is the whole reason it is worth knowing. One figure means one thing everywhere in the flat at once: anything below it will be wet, and anything above it will be dry. The number draws a single line across every surface in the building. Nothing has to be measured about the surface itself for the rule to hold.

The demonstration everybody already knows is a car left out overnight. No rain fell, and the roof is soaked at first light. Sheet metal loses heat to the open sky through the night and drops under the line, so the air resting on it unloads. By mid-morning the panel has warmed back above the line and dries with nobody touching it.

What relative humidity measures, and where a Singapore home should sit on it, is covered by the indoor humidity target guide. Dew point is the second half of that pair. It answers where water lands rather than how a room feels.

The air over Singapore is warm and wet, and neither condition lets up long enough to matter. The Meteorological Service Singapore records overnight minimums that usually do not fall below 23 to 25 °C, against daytime maximums that do not rise above 31 to 33 °C. Air that never gets properly cold is never forced to shed what it is carrying.

The calendar does not rescue it either. The same service notes that temperatures here vary little from month to month and from day to day, unlike the temperate regions. There is no dry season during which a building gets a break from condensation, and no cold spell that wrings the moisture out for a few weeks. Whatever a surface has to survive in March, it survives again in November.

Outdoor air comes closest to its limit shortly before sunrise. Relative humidity outdoors passes 90% at that hour on the service's figures. It falls to around 60% during the afternoon on days without rain. Air sitting above 90% is only slightly warmer than its own dew point. That is why railings, windscreens and window glass go wet in the small hours with no rain involved.

That thin margin is the part which does not transfer from elsewhere. A cold surface in a dry climate has a wide gap beneath it before anything condenses, so it can run cold and still stay dry. Here the gap is narrow. A surface temperature that is comfortably safe in a dry country is wet in this one. That is how detailing copied straight from an overseas manual fails on a local job, with nothing wrong with the equipment.

Indoors is not automatically better, only different. Cooling pulls the room's dew point down, because the coil takes water out of the air as it works. A shut, well-cooled room genuinely holds a lower line than the corridor outside its door. Open that door, run a shower, hang laundry on a rack, or leave a window ajar through a squall, and outside air walks back in with its own dew point intact.

An aircon cools by holding one surface under the dew point of the room. The indoor coil is that surface. Room air crosses the fins, gives up heat, and gives up water in the same pass, because the fins are under the line the whole time the compressor runs. Drying is not something the machine offers separately. It is what any surface under the line does to air pushed across it.

Water leaving that coil has a designed route. That route is the difference between a working system and a wet wall. Droplets run down the fin faces into a tray. The tray falls to a pipe that carries them out of the building. Every cold surface in the system either sits inside a route like that one, or else is wrapped so that room air never reaches it.

The thick copper line running back to the outdoor unit is the second cold surface, and there is no tray anywhere near it. Refrigerant keeps that pipe well under the line along its entire length, through bedrooms, above ceilings and out across a ledge. Left bare, it would run wet end to end. The foam sleeve on it exists for that reason and no other.

Insulation on a refrigerant line is not there to save energy, whatever it resembles. Its job is to hold the outside of the sleeve above the dew point, so that room air meets something warm and stays clear of it. The cold inside is unchanged. What moved is the temperature of the surface the air is allowed to touch, and that is the only variable condensation cares about.

Read that way, the defect list gets shorter and more specific. Take a sleeve crushed flat at a tight bend, a joint left unwrapped, or a gap where a run passes through a wall. None of those stop a room getting cold. All of them expose metal that sits under the line. How thick the foam should be is settled in the piping specification. A run buried where nobody can inspect it is the subject of the guide on concealed routes.

Most water complaints on an aircon are one event happening at different addresses. A surface was under the dew point, room air reached it, and what came out of that air had to go somewhere. What varies between complaints is which surface and which way the water ran. The mechanism underneath is identical every time.

A sweating pipe is the version everybody recognises. The wet stretch is where the sleeve stopped covering cold metal, and it tends to end sharply at both ends. That boundary marks where the wrap resumes and the metal is warm again. Which of several faults left it in that state is another question entirely. The problem page on pipe condensation is where those get told apart.

The damp bloom that shows on a ceiling beside a bulkhead is the same event, out of sight. A cold run above the boards wets its own length inside the void. That water drips onto board, soaks through it, and shows on the room side wherever the board finally gives way. Where it appears has more to do with the board than with the pipe. The visible mark is a poor guide to the distance back to the cause.

Beads on a supply grille or on louvre blades report on the room rather than the machine. Those plastic faces sit close to the temperature of the air leaving the unit. That is cold, without usually being cold enough to condense the room's own air onto them. When they fog, the room's line has risen to meet them. An open door onto a humid corridor will do it. So will laundry drying nearby, or a squall pushing wet air through the flat.

This also explains an installation that behaved for two years, sweats for one week, and stops again with nobody having touched it. The surface did not move. The line did. A run whose foam had been marginal all along sat over the dew point for most of the year and under it through the wettest stretch. The household reads that as a fault developing. What developed was the weather, and what it exposed was a margin that was always thin.

Two quick dismissals are worth resisting when this comes up. The first calls any sweating normal because the climate is humid, which is fair for a drain outlet and wrong for a wrapped line indoors. The second treats every wet surface as something escaping from inside the system, and quotes accordingly. Water pulled out of room air and water that left a pipe look identical on plaster. Where the wet area starts and stops is what tells them apart.

Condensation announces itself in patterns, and the patterns are readable without a meter. Three things are worth noting: where the wetness stops, when it turns up, and what it does as conditions change. Those answers narrow the problem further than any single figure taken off a wall.

Where the wet area ends is the strongest evidence available to anyone. Condensation stops hard at the point where a surface climbs back over the line, so a bare length of pipe wets as far as the bare metal reaches and not a centimetre further. Water that has run in from somewhere else carries no such boundary. It follows gravity, spreads along a seam, and settles wherever the surface is lowest.

Timing separates the room from the machine. Wetness that shows up only while the unit runs, and clears soon after it stops, is tracking a cold surface. Wetness that peaks in the early morning, or after rain, or on a day the flat stood open, is tracking the air instead. On a marginal run both are true at once, and saying which pattern dominates is genuinely useful to whoever has to find it.

Ask what was built recently. Condensation that began after a renovation, a new false ceiling, a wardrobe pushed against a wall or a box stored against a run is usually about air movement stopping rather than a component failing. Still air against a cold surface reaches the line locally and stays there. The identical surface in moving air stays dry. Nothing has to be broken for that to happen.

Two responses are worth resisting. Wiping and repainting a stained ceiling restores the finish without touching the surface that made the stain. The mark returns during the next wet spell. Buying a dehumidifier to stop a pipe sweating treats an entire flat for a fault living on one length of foam. That appliance runs for years without once reaching the cause.

None of this settles which fault let the surface get that cold. A dew point explains why water appears in a given place. It does not explain what dropped the surface there, and those are different questions with different repairs behind them. What careful observation does is make the second question answerable before anything is opened up.