Aircon filter drier: why a compressor repair needs one
A quote for a compressor replacement usually carries a second, much smaller line for a filter drier. It reads like padding. Leaving it off is the cheapest way to make the expensive part of that repair fail a second time.
By Team Snowflake | Updated 5 Aug 2026
The canister on the liquid line
A filter drier is a sealed metal canister fitted into the liquid line, usually within reach of the outdoor unit. Refrigerant passes straight through it on every lap it makes. Inside are two things doing two separate jobs: a desiccant that holds water, and a screen that stops solid particles.
The desiccant is the half most owners have never heard of. It is a bed of porous beads, or a moulded core made from the same material, chosen because it grips water molecules and will not hand them back into warm refrigerant. The sachet in a shoebox works on the same principle. Only the scale and the stakes differ.
The screen is ordinary by comparison and no less useful. Fine mesh, sometimes with a felt pad behind it, sitting there to catch grit, swarf and flakes of whatever has broken loose further upstream. Whatever it holds is material that never arrives at the narrowest gap in the system.
Position is not incidental. The canister goes on the stretch running from the outdoor coil towards the room, where refrigerant is at its densest, and it goes there because the next thing downstream is the metering device. Everything caught in the canister is caught before that point.
What a drier is not
A drier cannot be serviced. Nothing about it opens, rinses, dries out or goes back. It is brazed or flare-fitted into the pipe as one sealed unit, and its only maintenance is removal with a fresh one put in its place.
It is also not the filter behind the indoor grille. The two get confused because both are called filters and both catch dirt. The one in the room handles air and dust, gets washed at every service, and never touches refrigerant. The canister handles what is inside the pipe and is invisible from inside the flat.
What water does once it is shut inside a circuit
Refrigerant oil and water do not stay separate once they are shut in together. The two react, what comes out of the reaction is acidic, and it then circulates wherever the oil circulates. That puts it inside the compressor shell, against the motor windings, on every single run.
Damage by that route works from the inside outwards, which is why it takes so long to announce itself. Windings are insulated with a varnish coating rather than bare copper. The acid attacks that coating, the winding carries on working, and the motor fails only when the weakened insulation finally gives way. Nothing about the failure points back at the water that began it.
The second effect arrives sooner and is easier for an owner to notice. Free water travelling with the refrigerant reaches the metering device, where pressure falls sharply and temperature falls with it, and at that point it can freeze. A plug of ice chokes the flow, cooling drops away, then the ice clears and flow resumes. Cooling that fades and returns with nothing else changing is one of the shapes this takes.
Neither effect needs very much water to begin. The circuit does not have to be flooded. Damp air pulled in through an open joint carries enough vapour to start both processes, and a sealed loop offers nothing that arrived inside it any way out. That gap is what the desiccant exists to close.
Where the water gets in
Almost all of it enters while the pipework is open. Singapore air is humid, and copper standing open at a cut end or a flare draws moisture out of the air around it while the work runs. Pulling a vacuum before charging is what draws that back out again, and what a proper vacuum has to reach is settled elsewhere.
A slow leak lets water in too, less obviously. A circuit that is losing refrigerant is by definition not sealed, and a path out under pressure becomes a path in whenever the system stops and pressures even out. A unit that has been quietly shedding gas has usually been quietly taking on damp air alongside it.
A failed compressor spreads itself around the pipework
Mechanical failure inside a compressor does not stay inside the compressor. Bearings, valve plates and windings all sit in the path of circulating oil, so anything that wears or shatters leaves with it. Fine metal ends up distributed through pipe, coils and fittings alike.
That is what makes debris an argument specific to this repair rather than a general one. A choked drain leaves nothing behind. A dead capacitor leaves nothing behind. A compressor that ground itself apart has already seeded every metre of copper, both coils and the metering device with the evidence of how it died.
An electrical burnout produces something worse than filings. Windings overheat, the insulation and the oil break down together, and what comes out is carbon, sludge and acid. It coats surfaces rather than settling on them, and getting it back out afterwards is far harder than catching metal.
Fitting a clean compressor into that circuit without renewing what filters it is exactly the outcome this line item prevents. The replacement is the most expensive item in the repair. It is being asked to pump through pipework that still holds whatever finished its predecessor.
Why the debris does not simply settle
Refrigerant moves whenever the system runs, and it moves fast enough to keep fine particles in suspension. Debris does not sink to a low point and stay put. It travels round with the oil and arrives, sooner or later, at whichever passage is tightest.
That passage is the metering device, and obstructing it produces a fault that reads like something else. Cooling falls off, pressures look wrong, and the investigation tends to open on the refrigerant charge rather than on a partial blockage. A system that recently lost a compressor is the one place where that suspicion belongs at the front of the queue.
A drier is a consumable, not a fixture
Whatever a drier takes up, it keeps, and its capacity for keeping is finite. The desiccant fills. Once it is full there is no overflow anybody can see. It simply stops accepting anything further, and the circuit downstream carries on as though something were still guarding it.
A saturated one can be worse than none at all under the right conditions. Desiccant gives back what it holds when it gets hot enough, so water taken up during one part of a cycle can be released later in another. A full canister that has also collected debris turns into a restriction of its own, producing the same complaint a blocked metering device produces.
The question is therefore never whether the old drier looks damaged. It never looks like anything. There is no window in it, nothing to measure with the unit running, and no reading at the service valves that reports how loaded it has become. Its condition is worked out from history instead: what the circuit has been through, and how often it has been opened.
That history is why opening the circuit and renewing the drier are treated as a single action. Every opening admits fresh damp air, and all of it lands on desiccant that was already carrying a load. A canister with capacity to spare beforehand may have none left afterwards. The part is inexpensive against what it guards, and the moment to fit it is the moment the pipe is already apart.
The case is strongest after a compressor failure
A new compressor arrives clean and goes into a circuit that is not. Recovery removes the refrigerant, but recovery does not lift what has stuck to the inner wall of the pipe, and coils cannot be rinsed out where they sit. Some of what killed the last compressor is still in there when the new one is fitted.
Dealing with that residue is the whole job of the drier. It is the only item on the circuit whose entire purpose is to trap what a previous failure left behind. Omitting it does not make the repair fail on the day it is signed off. It shortens whatever life the new part had ahead of it, and the second failure presents identically to the first.
Reading the drier line on a repair quote
A compressor replacement quoted with no drier anywhere on it is the one worth raising. Treat that as a question rather than a verdict. The part may be folded into a labour or sundries line, the document may be a summary rather than an itemised breakdown, or it may have been overlooked while pricing a large job.
Asking costs nothing and the reply is short. A supplier who intends to fit one says so straight away, because it is a routine step in a sequence they already have in mind. A supplier who does not intend to fit one will say that too, and the reasoning behind it is worth hearing before the work is approved.
What separates a solid answer from a thin one is whether it names the order of work. Recovery, the part swapped, a new drier, evacuation, then a measured charge. That order is not adjustable and it is not decoration. Anyone who does this work carries it in their head anyway, so hearing it back is a fair expectation rather than a trap.
The table sets out the common shapes a quote takes and the question each one leaves open.
| What the quote shows | What that leaves open | The question worth putting |
|---|---|---|
| What the quote showsCompressor replacement with no drier named anywhere | What that leaves openEither folded into another line or absent from the plan | The question worth puttingIs a new drier included, and which line is covering it |
| What the quote showsA drier listed and fitted on the liquid line | What that leaves openNothing. That is the standard scope for work that opens pipework | The question worth puttingNone. This is the line you were looking for |
| What the quote showsThe drier shown as optional, or as an upgrade | What that leaves openThe same repair is being priced two different ways | The question worth puttingWhat changes for the new compressor if it is left out |
| What the quote showsA burnt-out motor, with one drier quoted | What that leaves openThe liquid side is covered and the return leg is not | The question worth puttingWhether a suction line drier is warranted after a burnout |
| What the quote showsA coil or valve change with no drier on it | What that leaves openThe pipework still had to be opened to reach that part | The question worth puttingWhether the existing drier is being reused, and on what grounds |
The suction side variant, and when it appears
A second drier sometimes shows up on a quote after a burnout, fitted on the suction side instead. That is the larger insulated pipe carrying refrigerant back towards the compressor. A canister placed there catches acid and fine debris on the last stretch before the new part draws it in.
It exists because burnout contaminates differently from a mechanical failure. Acid does not confine itself to one side of the system, and the return leg feeds the compressor directly, so the last trap before the inlet is where it earns the most. Seeing one on a quote is a sign the supplier has priced for the failure that actually happened rather than a generic one.
It is usually temporary by design. A suction drier is fitted to clean up a dirty circuit, and once it has done that it becomes a restriction the system no longer benefits from. Ask whether taking it out again, or swapping it, sits inside the quoted price or comes back later as a separate visit.
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.