Aircon DC bus capacitor: not the one that gets swapped

The word covers two components, and the small repair is only one of them. On a fixed-speed outdoor unit, a capacitor sits in the motor circuit next to the compressor and the fan. Its job is to help those motors turn. It is a stocked item, it is checked with a meter, and changing it counts as ordinary work.

An inverter outdoor unit carries a second one that does no such job. It sits in the power path on the control board. On one side of it is the stage that makes DC out of the mains supply. On the other is the stage that drives the compressor. Nothing about it touches a motor winding.

A homeowner told the capacitor has gone will picture the first one. That is the one the trade talks about, and the one that turns up in most repair stories. The gap shows itself only at the point of approval. One answer describes a component swap. The other describes work on the board that runs the whole outdoor unit.

The confusion is structural, not careless. Almost every capacitor a technician changes in a year is the motor one. It rides in the van, it costs little, and it fixes a common fault. The word attaches itself to that part through sheer repetition. A second component wearing the same name gets no separate vocabulary in ordinary conversation.

An inverter aircon never feeds mains power straight to the compressor. Daikin's service manual for the FTXS-L series sets the conversion out in two phases. "The supplied AC power source is converted into the DC power source for the present." Then, "The DC power source is reconverted into the three phase AC power source with variable frequency." Moving that frequency is how the machine changes compressor speed.

The capacitor in question sits between those two phases. DC made from a mains supply rises and falls with the supply behind it. The drive stage needs a level it can work against. Holding that level is the entire job, which is why the manuals name the part after it. Daikin calls it the built-in smoothing electrolytic capacitor. Mitsubishi Electric writes smoothing capacitor. Toshiba writes smoothing condenser.

Downstream of that, the drive does something less obvious. It builds three separate feeds out of that single DC supply, and the compressor motor turns because those three arrive in sequence. How fast the sequence runs is how fast the motor runs. That is the whole trick. It only works while the supply underneath stays where the drive expects it.

A fixed-speed outdoor unit has no such stage at all. Mains power reaches the motor through a contactor, and the motor runs at one speed or none. There is no conversion, so there is nothing to hold steady. That is worth saying plainly. On a non-inverter machine, a capacitor named in a repair is the motor one, and it should be.

Machines watch the level closely and act on it. One inverter service manual sets an upper and a lower limit on the DC measured across the large capacitor on the outdoor board. Outside that band it stops the compressor and puts a protection code on the display. Nothing in that sequence reads to a homeowner as a capacitor problem. It reads as a unit that keeps stopping.

Capacitors of this type get chosen for how much they hold in a small space. Nippon Chemi-Con's technical note is blunt about it. Aluminium electrolytic capacitors "offer larger CV product per case size and lower cost than the others". Holding a supply level steady between peaks takes real storage. So the component ends up physically big.

Size is a fair proxy for what sits inside it. The motor capacitor is a small cylinder that fits in one hand. The one on an inverter board is usually the tallest thing on it. Makers size that part for the current the drive pulls, not for what a motor needs at startup.

Stored charge does not leave when the power does. Daikin's manual tells the technician to discharge the capacitor completely before repair work. A charged one may cause an electrical shock, it says. It also asks for confirmation that the charge on the smoothing capacitor has fallen before service work starts. Mitsubishi Electric's troubleshooting flow interrupts itself with one line: "Be careful of the residual voltage of smoothing capacitor."

One manual is sharper still, and the detail is worth sitting with. A multi-system outdoor unit manual, reference SM700738, sets two different waiting periods before anything is touched. A control board in normal condition clears the charge in minutes. A board thought to be in trouble gets a wait many times longer. The manual offers a blown board fuse as its example. The fault under investigation can be the reason the charge is still sitting there.

Two makers document a further behaviour that catches people out. Nippon Chemi-Con describes a recovery voltage: a capacitor charged, then fully discharged, then left with its terminals open will build a voltage again on its own over the following days. Nichicon says the same about a part already fitted and powered, and calls it a restriking voltage. Neither maker treats a discharged capacitor as a settled matter. None of this belongs to a homeowner with a screwdriver. It is why work on this component is priced as licensed work on a live power stage.

The end comes as a slow loss of the liquid inside, not as a bang. Nippon Chemi-Con describes it directly. The electrolyte "has gradually evaporated and diffused out of the capacitors through the rubber seal materials with time". That drags the stored value down and pushes the internal loss up. Those two numbers moving are what the makers define as the end of useful life.

Heat sets the pace of that loss. The same note names ambient temperature and the heating caused by ripple current as the crucial factors for lifetime in a normal filtering circuit. Ripple current is the constant charging and discharging the part does while it holds the level. Every bit of it turns to heat inside the can.

The wear then feeds itself. In the maker's own temperature-rise calculation, the heating climbs as internal loss climbs and eases as stored value climbs. Ageing pushes both of those the wrong way. An older part therefore runs hotter on the same current than it did when new. An outdoor unit on a west-facing ledge in Singapore is a hard place for that to be happening.

A normal-looking case proves nothing either way. Nichicon's guidelines list what an inspection looks for on the outside: "Remarkable abnormality such as vent operation, leaking electrolyte etc." Those findings are real, and they settle the question when they are present. The wear-out the same makers describe changes measured values and leaves the outside intact. Nothing to see settles nothing.

From the room, none of this looks like a capacitor. The unit cools, then stops. It may restart on its own, or it may sit there with a code showing. Heat makes it worse, so the afternoon is when it shows. That pattern fits a long list of faults, which is exactly why a reading matters more than the story.

Ageing here also fails to produce the symptoms people attach to the word. A weak motor capacitor shows itself in the fan and the compressor. A slow spin-up, a hum that goes nowhere, a machine that struggles at startup and then gives up. Those signs belong to the motor circuit. A tired capacitor on the inverter board produces none of them, because it never had any part in getting a motor moving.

There is a limit worth stating too. The value of this part gets judged with the machine out of service and the power stage safe to approach. It cannot be sampled from the room. It cannot be inferred from how the unit behaves on a hot afternoon. A quote naming it should state what was measured, and when.

On most home inverter units the honest answer is the board. One service manual describes the high-capacity electrolytic capacitors as sitting on the outdoor mainboard. Daikin's fault flows for an inverter problem end at replacing the outdoor unit PCB. Mitsubishi Electric's end at replacing the inverter board or the outdoor electronic control board. The component is documented in detail, and the repair is still written at board level.

Bigger equipment gets treated differently, which is where some of the confusion comes from. Toshiba's manual for a commercial modular system lists the smoothing condenser as its own check with its own rating, separate from the boards around it. Scale changes what is worth stocking and what is worth replacing whole. A wall unit in an HDB flat falls well short of that scale.

There is a request that comes up here and usually gets declined. Owners ask whether the capacitor alone can be changed on the board, instead of the whole board. On home equipment the answer is usually no, and that tracks what the makers instruct. The manuals send the technician to the board because the board is what they qualify, stock and warrant. A repairer saying so is quoting the manual, not padding the bill.

One question separates the two components without any technical argument. Ask where the part sits. Then ask what reading was taken, because the two are found by completely different checks. A technician who has done the work can answer both without hesitating.

A quote that survives a second reading records three things. It names the part in the maker's own words, so smoothing capacitor and run capacitor never blur together. It states the reading that condemned it. It says what physically comes out of the unit, because that is what the price is attached to. A quote missing all three is a price with no diagnosis behind it.