Aircon control transformer: control power, not drive power

An aircon runs on two voltages at the same time, and only one of them moves anything. Mains voltage turns the compressor and the fans. A much lower voltage runs the deciding parts. That means the sensors reading the room, the board weighing what they say, and the coils that pull the heavy switches in.

The low supply exists because the deciding parts cannot live at mains level. A sensor, a small processor and a display all work at a fraction of it. The product standard that governs these transformers puts it plainly in its own scope. They exist to give equipment a voltage different from the supply, for the functional needs of that equipment.

That circuit does more than feed a board. On ducted plant the low supply also carries the safety chain. Wiring diagrams for split system air conditioners show the high and low pressure switches sitting in the same low-voltage loop as the coil that starts the compressor. Open any one of them and the machine stops, with nothing actually broken.

The drive side is a separate story with separate hardware. An inverter unit turns the mains into DC across a diode stack, steadies it on a smoothing capacitor, then switches it back out at whatever frequency the compressor needs. That path exists to move a motor. The control supply exists to decide when the motor should move at all.

The two fail in ways that look nothing alike. A fault on the drive side usually announces itself, because the deciding part is still awake to see the problem and post an error code. A fault in the control supply takes the deciding part down with it. Nothing gets reported, because nothing is left to do the reporting.

Household split systems do not carry one as a separate part. Open the wiring diagram of a wall-mounted split and the indoor electrical box holds a board, a fuse or two, a varistor, sensors and small motors. Where a transformer appears at all, it appears inside that assembly.

The MSZ-GA wall-mounted range makes the point cleanly in its Mitsubishi Electric service documentation. The indoor wiring diagram names a transformer, coded T111, fed from the mains through a 3.15 A fuse. The parts list at the back of the same manual sells the fuse, the varistor and the electronic control board on their own. It never lists the transformer. The board is the part.

Newer designs often show none at all. Daikin's manual for its R32 split range lists every item in the indoor wiring diagram, and that list runs board, buzzer, fuses, lamps, motors, sensors, switches and terminal strip. Nothing steps voltage down as a named component. The board derives its own low supply, and the indoor unit is fed from the outdoor unit over a four-core cable instead of from the mains directly.

Discrete control transformers belong to plant with contactors in it. Addison's manual for its commercial split condensing units and air handling units names two of them, coded T1 and T2, sitting alongside compressor contactors, motor starter protection and a power distribution block. The cover of that manual reads: not for residential use. That line is doing real work.

Ducted equipment run from a wall thermostat is the other home for one. A Carrier wiring diagram manual for split system air conditioners calls for a low-voltage circuit of at least 40 VA. That rises to 60 VA where a liquid line solenoid valve is fitted. It places the transformer in the indoor section. A few thin wires then carry the thermostat's decisions out to the condensing unit.

That last example is North American equipment, and Singapore homes do not use it. Naming it is still worth the space. Anyone searching this term after a quote has likely been reading material written for a market where the part is standard. In a flat with wall-mounted splits, it is absent.

A control transformer is rated for two loads at once, and the standard requires both on the marking. One figure is the continuous rating. The other is what the part must deliver for an instant. They are printed side by side, separated by a stroke, and the standard gives 100/300 VA as its own example.

The instant is the moment a contactor pulls in. An electromagnet pulling its armature closed wants several times the current it will need once that armature is shut. Rockwell Automation's technical data for control circuit transformers puts the window at 30 to 50 milliseconds, and the multiple at three to ten times normal.

Its tables show what that means in practice. A small contactor listed there wants 192 VA to pull in and 29 VA to stay in. Those numbers belong to one product line, so the ratio is what travels, not the figures. Stack several coils, plus relays and indicator lamps energising in the same breath, and the momentary demand dwarfs the settled one.

So the part gets sized on the pull-in. The selection method in that document works backwards from total inrush, never from the load the circuit settles at. It also asks how steady the incoming supply is, and pushes you to a bigger part where the mains wanders. Sizing on the settled load alone gives a transformer that is comfortable almost all the time and fails at the only moment that counts.

This is why a weakening control supply is such a quiet fault. Voltage that looks correct with nothing drawing from it can collapse the moment a coil demands its share. The standard sets the bar tightly for exactly that reason. Output must sit within five per cent of the marked figure at the continuous rating, and hold nearly all of that while the momentary load is on it.

Not every small transformer is held to that bar. The companion standard covering safety isolating transformers, the ones whose output stays below 50 V, lets the no-load and loaded outputs of a small one differ by half, and by more again at the smallest sizes. A part built to that allowance would be hopeless driving contactors. Control transformers earned their own standard, their own marking and their own tolerance because the job is different.

The same standard lets the part be trimmed to the site it lands on. Where the incoming voltage sits high or low, tappings on the input side shift the ratio. Each tapping has to be marked with the change it makes. A part left on the wrong tapping runs a control circuit that is slightly wrong forever, which is a harder fault to spot than a dead one.

The nameplate declares how the part behaves when it is overloaded, and there are only three answers. A short-circuit-proof transformer survives the fault and works again once it is cleared. A non-short-circuit-proof one does not, and leans on a fuse or breaker outside it. A fail-safe one destroys itself on purpose, opening a winding so the fault cannot continue. One of those three symbols has to appear on the marking.

Short-circuit-proof then splits again. An inherently short-circuit-proof part rides the fault out on its own design, running hot but inside its limits. A non-inherently short-circuit-proof one survives because something within it opens: a cut-out, a fuse, a link, a resistor that chokes the current as it warms. That distinction decides whether a technician is waiting for a part to cool or hunting for a device that has already operated.

External protection is normal on both sides of the part. Hammond Power Solutions ships secondary fuse clips as standard on its moulded control transformers above 100 VA, and sells a primary fuse block as an accessory. Addison's drawing uses both ideas at once. There is a fuse table sizing each device against the transformer's VA rating, and a separate circuit breaker in the low-voltage circuit.

The low side is often earthed on purpose, which surprises people who expect a transformer to keep two circuits apart. This type is used where full separation is not demanded, so part of the output may be bonded to earth. Hammond bonds an output terminal to the chassis as standard practice, and Carrier's diagram assumes the same.

Environment kills more of these than electricity does. Hammond specifies a clean, dry, ventilated position, free of dust, moisture and corrosive vapours, with the surrounding air no hotter than 40°C. Plant rooms and roof enclosures in Singapore run warm, and a closed panel inside one runs warmer still. A part in still, hot air is being asked to work at the edge of its rating every day it runs.

A household split carries the same idea in miniature. There is a fuse on the control board ahead of whatever makes the low supply, and that fuse is usually a stocked part even when the transformer is not. The fuse is what gets ordered. The board behind it is what has to be diagnosed.

None of this is owner-side territory. Hammond's own instruction sheet limits installation and servicing to qualified electrical personnel, and the enclosure the part sits in is live at mains on one side. Work on the supply feeding a machine sits with a licensed electrical worker, which is a separate question from who services the aircon.

From the room, silence carries no cause with it. A unit with no supply reaching it and a unit whose control side has stopped both do nothing at all. No display, no beep, no twitch from the fan. One observation, two very different jobs, and one of them sits upstream of the machine entirely.

The split decides who is called and what gets opened. If the supply never arrived, the fault lives in the breaker, the isolator, or the wiring between them, and the machine is innocent. If the supply arrived and the control side stayed dark, the fault is inside, on the board or on whatever feeds it. Those are different scopes, and on larger installations they are different trades.

The pattern across units is the strongest single clue in a multi-unit flat. One dead unit with three working ones points inside that unit. Every unit dead points at whatever they share, which is the supply and the protective device feeding it. Neither answer needs a meter, and both change what a technician brings on the first visit.

An owner can narrow this before anyone opens a panel, and none of it needs a tool or a cover coming off.

• •Whether the breaker on the aircon circuit has moved, or is still sitting where it always sits
• •Whether anything else fed from that same circuit is still working normally
• •Whether the unit died at a moment, at switch-on, during a storm, after other work in the flat, or faded across weeks
• •Whether one unit is silent while the others run, or every unit went dark together