AC Compressor Replacement Cost: It Never Fails Alone

People arrive at this question expecting to price a part. There is a compressor, it has failed, a replacement exists, and the reasonable assumption is that the bill is the compressor plus the hours it takes to unbolt one and bolt on another. That assumption is why the quote, when it comes, is so often met with the suspicion that somebody is inventing work.

They usually are not. The reason a compressor job costs what it does has almost nothing to do with the compressor being an expensive lump of aluminium, and almost everything to do with where it sits. It is a machine with moving metal parts running inside a sealed circuit that it shares with every other component in the system. When it wears out gently, at a seal, it leaks and that is the end of the story. When it comes apart internally, it distributes itself around the loop. There is no filter in there to catch that, no sump for it to settle into and no drain to let it out. Whatever the compressor sheds stays in the car until somebody removes it deliberately.

What that figure lays out is the real extent of the work, and it is worth absorbing before you read a single estimate. Every item on it earns its place the same way: it either holds contamination now or it is the thing contamination will destroy next. Fit a compressor on its own and the debris waiting in the condenser and the lines gets drawn into the new pump within days. The trade has a name for the result, and a shop that has watched it happen twice needs no persuading to do the job properly the first time. What follows is why each element of that scope is there, which failures are genuinely the big one and which merely look like it, and what any of it means if you are standing next to a car you are considering buying.

What the compressor is, and why its position is the problem

Everything in an air conditioning system exists to move heat out of the cabin and dump it at the front of the car, and the compressor is the thing that makes that possible. It takes low-pressure vapour returning from the evaporator behind the dashboard and squeezes it into hot, high-pressure gas, which is the state it has to be in before the condenser can take heat out of it and turn it back to liquid. Nothing else in the circuit does any work. The compressor is the pump, the pressure boundary and the only component with parts that move.

It is also driven directly by the engine on the great majority of ordinary cars, hanging off the front of the block and turned by the accessory belt. That mechanical connection is worth keeping in mind because it runs in both directions: the belt turns the compressor, and a compressor that seizes can stop the belt. Our guide to the belt that drives the accessories deals with what else is on that loop, and on most cars the answer includes the alternator and often the water pump, which is why a compressor failure occasionally arrives disguised as a charging fault or a temperature gauge climbing.

The one piece of background you need from elsewhere is that the circuit is sealed and self-contained. Refrigerant is not consumed by anything, so a system that is short of it has leaked, and our page on why air conditioning stops blowing cold works through what that means for diagnosis. Take the point in one line and move on, because for a compressor it has a very specific consequence. The sealed loop that keeps refrigerant in is equally good at keeping everything else in.

Consider what that means physically. Inside the compressor, depending on the design, there are pistons driven by a swash plate, or a pair of interleaved scrolls, or vanes running in a rotor. All of it is close-tolerance metal, and all of it is turning at engine speed whenever the system is switched on, lubricated by oil that arrives dissolved in the gas it is pumping. There is no oil filter. There is no strainer of any consequence between the compressor and the rest of the circuit. When those internal parts start to gall and shed material, the discharge port sends the fragments straight out into the high-pressure line and off around the system, along with the fine black abrasive slurry that carbonised oil and aluminium oxide make together.

That is the whole argument, and every remaining cost on the page descends from it.

Three ways a compressor gets switched on, and why it changes what you can check

Before anything else, it is worth knowing which kind of compressor a car has, because a good deal of the advice in circulation applies to only one of the three and quietly misleads owners of the other two.

The familiar arrangement is a fixed-displacement compressor with an electromagnetic clutch on the front. The pulley spins constantly with the belt, riding on its own bearing, and does nothing at all until the car energises a coil behind it. The coil pulls a steel plate into contact with the pulley face, the plate is splined to the compressor shaft, and the compressor starts pumping. Cooling is regulated by switching that clutch on and off, so a healthy system on a mild day will engage, run, disengage and repeat. This is the design behind every piece of advice that begins watch the front of the compressor.

The second arrangement is a variable-displacement compressor, and on many of them there is no clutch to watch. The pulley is coupled to the shaft permanently, so the compressor turns whenever the engine does, and output is regulated internally instead by tilting the swash plate to change how far the pistons travel. On the older versions that tilt is set by internal pressure. On the newer ones it is commanded electrically by a control valve in the back of the compressor, which the car varies continuously. The practical consequence is important: on such a car there is nothing to see engaging, no click to hear, and a failure of that little control valve produces a car that blows warm with a perfectly serviceable compressor and a perfectly correct refrigerant charge.

The third is the electric compressor, standard on hybrids and electric cars and now appearing on some conventional ones. It has no belt connection at all. A high-voltage motor is built into the compressor body and it runs whether or not an engine is turning, which is exactly why hybrids can keep the cabin cold with the petrol engine shut down. These carry a specific trap that has cost people a great deal of money: the lubricating oil in an electric compressor has to be non-conductive, so they use a polyol ester oil rather than the polyalkylene glycol oil found in belt-driven systems. Put the ordinary oil into one of these and the insulation resistance inside the compressor collapses. The car will usually detect it, and the repair is a compressor.

None of this needs memorising. It matters for one reason, which is that the commonest piece of used-car advice about air conditioning tells you to watch the clutch engage, and on a growing share of the fleet there is no clutch. Knowing which car you are looking at stops you drawing a confident conclusion from an observation that does not apply.

The clutch is not the compressor, and the difference is most of the money

On a clutched system, the two most commonly confused failures sit a few centimetres apart and differ in cost by a wide margin.

Start with the clutch. It is an electromagnet, an armature plate mounted on leaf springs, and a running clearance between them. Every engagement slams that plate against the pulley face, and over hundreds of thousands of engagements the friction surfaces wear and the gap grows. Magnetic force falls off sharply with distance, so there comes a point where the coil can no longer close the gap it is being asked to close. What the owner experiences is an air conditioning system that works in the morning and gives up in the afternoon, because a warm coil has slightly higher resistance, draws slightly less current and produces slightly less pull. It is a classic, and it is routinely reported as the compressor dying.

The coil itself fails too, usually open-circuit, and then nothing happens at all. So do the things upstream of it. There is a relay, sometimes a dedicated fuse, a pressure switch or transducer that inhibits engagement when the system is outside its safe range, an ambient temperature input that stops the compressor running in cold weather, and on modern cars a control module that has the final say. Any one of them can leave you with a compressor that never turns while the compressor itself is entirely healthy.

There is a third item in the same area worth separating out, because it is the one that gets misread most expensively. The pulley runs on its own bearing, and that bearing turns constantly whenever the engine is running, whether or not you have ever pressed the air conditioning button. When it dries out it rumbles or shrieks, and the giveaway is that the noise is completely indifferent to whether cooling is switched on. That is a bearing, not a pump. On some compressors it can be replaced on its own for very little. On others the pulley assembly is not serviced separately, and a failed bearing takes the entire compressor with it, which is a genuinely irritating way to spend money and one of the clearest examples of why the part that failed and the part you must buy are not always the same part.

Set against all of that, internal compressor failure has its own signature. The noise appears only when the system is engaged, changes with engine speed, and tends to be a grinding or knocking rather than a bearing rumble. The system may still produce some cold air while it is doing it. And in the worst version the shaft locks solid, at which point the clutch either slips against the seized shaft and burns, or the belt loses the argument.

What happens inside the loop when a compressor lets go

This is the section that explains the number, so it is worth being concrete about the mechanism rather than gesturing at contamination in the abstract.

When the internals of a compressor begin to fail, they do it by rubbing. Aluminium against steel, at engine speed, with insufficient oil and rising temperature. The material that comes off is not a few visible chips. It is a mixture of coarse fragments and an extremely fine suspended powder, and it leaves through the discharge port because that is the only way out. The first place it goes is the high-pressure line to the condenser.

Follow it from there and you can see the whole repair scope assemble itself without anybody having to argue for it.

The condenser is a heat exchanger sitting at the nose of the car, and on any reasonably modern vehicle it is a parallel-flow design: the refrigerant is split across a large number of very small extruded passages to get as much surface area against the airflow as possible. That construction is excellent at rejecting heat and close to hopeless at giving up debris. Solvent pushed through it takes the path of least resistance and leaves the blocked passages blocked, which is exactly where the debris is. This is why a shop that flushes a serpentine-tube condenser on an older car will often insist on replacing a parallel-flow one, and it is one of the larger single items on the invoice. It is also, awkwardly, one of the items an owner is most likely to challenge, because the condenser was not what broke.

Next in line is the metering device, either an expansion valve or an orifice tube depending on the system, and it is the narrowest passage in the entire circuit by design. Its whole function is to restrict flow. Every particle circulating in the high side arrives at it, and orifice tubes carry a small screen precisely because that is where things collect. A blocked metering device produces poor cooling on a system with a correct charge, so leaving the old one in place after a compressor failure produces a car that has just had an expensive repair and still does not blow cold.

Then the accumulator or receiver-drier, the canister of desiccant whose job is to hold the trace of moisture that finds its way in. It has to be replaced anyway whenever the loop has been opened to the atmosphere, because desiccant that has breathed room air is spent. After a compressor failure there is a second reason: it has been catching debris, and putting a used one back means feeding that debris to the new compressor.

Finally the hoses and hard lines. Most can be flushed. Some cannot, because a number of discharge and suction hoses contain internal mufflers or baffles fitted to damp the pulsing of the pump, and those are ideal traps for material that will never come back out. A shop that knows a particular car will know which of its lines are flushable and which are on the replace list.

Put that chain together and the point makes itself. The compressor is one item on a bill made of consequences.

What a proper repair does beyond fitting the part

Two operations on the invoice are not parts at all, and both are worth understanding because they are where an owner is most tempted to ask whether the corners can be cut.

The first is the oil. There is no dipstick on an air conditioning system and no reservoir. The compressor is lubricated by oil that circulates dissolved in the refrigerant, which means that when a repair removes the condenser, the drier and a couple of lines, it also removes a quantity of oil that was distributed through them. New compressors arrive either dry or shipped with a full system quantity that has to be drained out and metered back in. Getting this wrong in either direction causes trouble: too little and the new compressor is starved from its first revolution, too much and the surplus coats the inside of the evaporator, insulates it and the car never cools properly despite everything being new. This is careful, unglamorous work, and it is a real part of the labour figure.

The second is evacuation. Once the circuit has been open to the air it contains moisture, and moisture inside a refrigerant loop is not inert. Combined with refrigerant and heat it forms acids that attack aluminium from the inside, which is one of the quieter reasons systems that have been opened carelessly fail again later. The remedy is to pull the whole loop down under deep vacuum for long enough to boil the water out of it before any refrigerant goes back in. It takes time, that time is chargeable, and rushing it is invisible to the customer on the day and expensive to them a year later.

Both of those steps also explain a smaller thing people find odd, which is why the shop wants the car for most of a day for what looks like a bolt-off, bolt-on job.

What actually sets the figure

So here is what to interrogate in a quote, in rough order of how much it moves the total.

Whether the failure was internal or external. This is the largest fork in the road by a distance. A compressor leaking at its shaft seal, caught before the system was run dry, may need nothing more than the compressor, the drier and a recharge. A compressor that has come apart inside takes the condenser, the metering device and possibly lines with it. Same part, same labour to swap it, entirely different job.

Whether the condenser can be flushed or must be replaced. Decided by how the condenser is built rather than by anybody’s preference, and it is usually the second largest part on the list.

How the compressor is packaged. Whether the clutch, the coil, the pulley bearing and the control valve are available as separate items or only as part of a complete compressor is a decision the manufacturer made years ago, and it determines whether a small fault stays a small fault.

What has to come off to reach it. Compressors are mounted wherever there was room, which on a transverse engine in a small car can mean underneath everything else, reached from below with the undertray and sometimes more out of the way. On other cars it is at the top of the block and visible the moment you open the bonnet. The part is identical and the hours are not.

Which refrigerant the car takes. Covered below, and knowable in advance from a label.

New, remanufactured or aftermarket. A genuine part, a remanufactured unit from a specialist and a budget aftermarket compressor are three different prices attached to three different risk profiles, and the warranty conditions attached to each are worth reading rather than assuming.

Whether anything unknown has been put in the system. Sealants and unidentified refrigerant mixtures complicate recovery, and some shops will decline the work rather than risk their equipment. A car with cans in its history can be materially harder to get repaired at all.

The refrigerant question, briefly

A compressor repair opens the loop, so the entire charge has to come out and a fresh one has to go back in. That makes the refrigerant a real line on this invoice rather than an afterthought, and which one the car takes is fixed by the car.

Three generations are in service simultaneously. The EPA sets the order out plainly enough: CFC-12 at the start, HFC-134a through the long middle of the fleet, and a newer generation on recent cars, of which HFO-1234yf is the one most buyers will actually encounter. For a buyer the practical content of that history is short. The three are priced nothing like each other, the newest is dearest by a wide margin, and any repair that empties and refills a circuit feels the difference at once. Two cars of similar age and value can therefore carry quite different air conditioning economics for no reason but what is in the pipes, and the label under the bonnet tells you which in a matter of seconds.

The legal frame matters here as well, though only in outline — our guide to diagnosing air conditioning faults works through the rules properly. Two provisions reach into your quote. Certification under section 609 is required of anyone servicing one of these systems for consideration of any sort, along with approved handling equipment, and it applies whichever refrigerant the car uses. Separately, section 608 forbids deliberately releasing refrigerant during maintenance, repair or disposal, so the charge has to be captured and then either recycled or reclaimed before it can go back — even when it is going back into the vehicle it left. Under the Significant New Alternatives Policy programme, established at section 612, every approved refrigerant carries its own fittings so that two of them cannot be combined by accident.

What this means for your quote is straightforward. A compressor job includes recovery, evacuation and a metered recharge performed on certified equipment by a certified technician, and those steps are not optional extras that a cheaper shop has cleverly found a way around. If a price looks impossibly low, one of the reasons may be that it does not include them.

Why running it low is what actually kills compressors

Very few compressors die of old age. Most are killed, and the mechanism is consistent enough that it is worth spelling out, because it is also the one part of this whole subject an owner has direct control over.

The compressor has no independent lubrication. Its oil is carried around the circuit dissolved in the refrigerant, returns to it through the suction line, and is only present in the quantity that the circulating refrigerant delivers. Reduce the amount of refrigerant in the loop and you reduce oil return in the same proportion. At the same time, a system running short of charge runs hotter at the discharge side, because there is less mass flowing to carry heat away, and thinner oil films at higher temperature are precisely the conditions under which aluminium and steel start to gall.

There is a second mechanism on clutched systems that compounds it. As the charge falls, the low-pressure switch begins cutting the clutch out and cutting it back in, faster and faster, until the compressor is being engaged and disengaged every few seconds. Each of those engagements is a shock load on the clutch and a hammer blow to the drive. Short-cycling is audible from outside the car as a rapid, insistent clicking, and it is a system telling you that it is destroying itself.

Which produces one instruction worth more than everything else on this page. The moment the vents stop going cold, turn the system off and keep it off until somebody has put gauges on it. Persevering preserves nothing. Every mile driven with the button lit on a starved circuit is spending a component that costs several times what closing the leak would have.

The seasonal shape of this subject follows directly. Nobody notices an air conditioning fault in December, and nobody thinks about it. What happens instead is that the system leaks quietly through the winter with nothing to reveal it, the first genuinely hot day arrives, the button gets pressed, and the compressor spends the following weeks working at maximum load on a charge that will not support it. The failure is discovered in summer because summer is when the load is applied, and by then the damage that decides the size of the bill has already been done. A car bought in winter with a system that was never tested is a car whose air conditioning history you have simply not read. Keeping an eye on this the way our list of maintenance a used car actually needs suggests costs nothing and prevents the expensive version.

Why there is no federal failure data here

This site normally reaches for the federal archives before it reaches for an opinion, so it is worth stating plainly that on this subject there is nothing there to reach for.

The archived pull of owner complaints that sits behind our other repair pages is filed by component category, and air conditioning is not one of them. The categories run through brakes, steering, engine, fuel system, electrical, suspension, airbags and the rest, and no air conditioning entry exists. The safety recall archive is the same: the component families in it cover everything from latches to stability control, and none of them is air conditioning or climate control.

The gap is jurisdictional rather than a verdict on how often these things break. Climate systems answer to EPA, which treats them as a refrigerant-handling and environmental question under the Clean Air Act — mostly at section 609, with sections 608 and 612 doing smaller jobs. They are not, in the usual case, a crash-safety question, and the archive of campaigns and defect filings is assembled by the safety regulator for safety purposes. So the data is missing because the agency that would have gathered it was never handed the subject. Stretching some neighbouring category over the hole would be a good deal worse than pointing at it.

Something adjacent does appear in the records, and it bears directly on compressors. The condenser sits ahead of the radiator at the extreme front of the vehicle, which makes it among the first components disturbed in even a modest shunt, and any car repaired after one has had the entire cooling pack out and back in. How carefully that was done leaves no visible trace under the bonnet. Collision damage, unlike a worn compressor, does tend to get written down somewhere, so two minutes spent to see whether front-end damage has ever been recorded against the VIN is time well spent before you get as far as arguing about vents.

Checking the compressor on a car you are looking at

Everything above concerns a car somebody already owns. This section is for the person standing next to one they do not.

Climate control occupies an odd position during a viewing. It is among the dearer comfort repairs on an ordinary vehicle, it is trivially easy to leave undemonstrated, and through the colder months the weather itself supplies a seller with an excuse that sounds entirely sincere. Most buyers never check it. A fair number give it no thought at all until the following June, by which point the discussion about who pays finished months ago.

Testing it when it is cold outside, and one caveat almost nobody mentions

You can certainly assess one of these systems in winter. Cooling is not a seasonal function: the machinery pulls heat out of the cabin and dries the air passing through it, and both effects are demonstrable in February as readily as in August. Warm the engine properly first, choose the coldest setting with cooling selected and recirculation off, and set the blower somewhere in the middle rather than at maximum.

Here is the caveat that matters, and it cuts against the usual advice. Many cars deliberately inhibit the compressor when the ambient temperature is close to or below freezing, because running it in those conditions risks liquid refrigerant reaching a compressor that cannot pump it and moisture freezing where it should not. That inhibition is designed in and is entirely normal. So on a genuinely cold morning a car that refuses to engage the compressor has not necessarily told you anything at all, and treating that refusal as proof of a fault is how buyers talk themselves out of good cars.

What still works in the cold is the dehumidification test. Select the windscreen demist setting, which on most cars commands the compressor on regardless of what the temperature dial is doing, and watch how quickly a misted screen clears. A system that clears the glass fast is a system that is removing water from the air, and it cannot do that without a working refrigeration circuit. It is a proper functional test dressed up as an ordinary thing to do on a cold day.

Listen to it, then listen to it under load

Raise the bonnet, leave the engine ticking over, and have somebody cycle the cooling on and off while you stand at the front. On a clutched car you want to see the centre plate of the pulley snap into contact, and to hear a soft clunk with a small dip in idle as the engine picks up the load. On one without a clutch there is nothing to see at all, which is information rather than a fault.

Then attend to noises, and separate them the way the previous section did. A rumble or a whine present with cooling switched off, unchanged when it is switched on, is the pulley bearing. Grinding or knocking that appears only once the compressor is engaged, and rises with engine speed, is coming from inside the pump. And a rapid click-click-click of the clutch engaging and dropping out every few seconds is short-cycling, which means the charge is low and the system has probably been running that way for some time. That last one is the most useful sound in this whole article, because it tells you both that there is a leak and that the compressor has been working under the conditions that ruin compressors.

Then leave it running. Switch it on when you arrive and then forget about it. Go through the history folder, walk round the tyres, lift the boot floor, take the drive, all with cooling selected, and return to the vents at the end with a spell of stationary idling somewhere in the middle. Marginal systems are convincing for a few minutes and much less so after forty. Cooling that gives up specifically at a standstill points at airflow across the condenser rather than at the pump, which brings in the radiator, the fans and everything else in the pack that our page on why a temperature gauge climbs deals with.

The trap: a compressor that never engages

This is the specific thing to be careful about, and it deserves its own heading because it is the failure most likely to be priced wrongly by both parties.

A car whose system has simply leaked out entirely will not engage its compressor. That is not a fault, it is a protection: the low-pressure switch is refusing to run a pump that has no refrigerant and therefore no oil reaching it. The car presents as air conditioning that does nothing whatsoever, which sounds dramatic and may be nothing more than a perished seal and a recharge away from working.

A car with a seized or electrically dead compressor presents in precisely the same way. So does a car with a failed clutch coil, a blown fuse, a bad relay, a faulty pressure transducer or a control module that has stopped asking. From where you are standing in the road, all of these look identical: press the button, nothing happens.

The spread between the cheapest and the dearest of those outcomes is enormous, and no amount of looking at it will tell you which one you have. What settles it is fifteen minutes on a set of gauges, which is why this belongs inside a proper pre-purchase inspection rather than in a negotiation. Beware in particular of the seller who has already decided which it is. “It just needs gassing” is a claim about a diagnosis nobody has performed, and it happens to be the cheapest of the available explanations.

The belt that goes round the compressor, or does not

One check takes ten seconds and is almost never performed. Look under the bonnet and follow the accessory belt round with your eyes. On any car with a belt-driven compressor, that belt should wrap the compressor pulley along with everything else on the drive.

Occasionally it does not, and there is a shorter belt fitted that routes around the compressor entirely. This is a real repair that real garages perform, and in fairness it has a legitimate use: when a compressor seizes, bypassing it gets a car home or keeps it usable while a decision is made. What it also does is make the fault invisible. Nothing on the dashboard complains, the car drives normally, and the only sign is a belt that takes a different route from the one in the manual. If you find one, you have found a car with a known dead compressor and a seller who either knows and has not said, or does not know and has been told the car is fine.

While you are looking, glance at the condition of the belt itself and at the compressor pulley face. A belt with glazing or shredding on a car that also has poor cooling is a coherent story rather than two separate ones.

What a very recent recharge tells you when the compressor is the question

A recharge shortly before a sale is not automatically sinister, but it is never neutral, and on this page it carries a sharper meaning than it does elsewhere.

A system only needs recharging because it was low, and a system is only low because it leaked. That much is arithmetic. The extra step that belongs here is what the low charge did to the compressor while it was low. Nobody recharges a system the week it starts leaking; they recharge it after a season of lukewarm air, which is a season of the compressor running with reduced oil return. So the receipt documents two things at once: an unrepaired leak, and a period of exactly the operating condition that wears compressors out.

Ask plainly where the leak was found and what was fitted to cure it. Good answers exist and come with a part on an invoice. Ask too whether any product has gone in from an aerosol at any point, because sealants and unidentified mixtures make later servicing considerably harder. Silence on both counts tells you nobody ever looked, and what you would be taking on is not merely the leak but whatever it has already done to the pump.

Putting a figure on it at the kerb

Say the cooling is dead, the rest of the vehicle is right, and you want it anyway. Perfectly reasonable. On an older car a non-functioning climate system is often well down the list of things that ought to worry you.

The error at this point is letting the seller supply the estimate. Usually not out of bad faith — most of them are repeating what somebody said at a parts counter — but because four completely different repairs are hiding behind one symptom, and their costs are nowhere near each other. A relay. A clutch coil. A pump caught before the charge ran out. A pump that has already distributed itself through the condenser and the metering device. Nobody guessing between those four guesses against their own interest.

What changes the conversation is paper: an independent written diagnosis, obtained before you commit, produced as a finding rather than as an opinion. That method is the whole basis of our page on how to negotiate on a used car, and here it does something a haggle never can. It sorts the vehicle into one of two boxes — a modest electrical fault, or a full system repair — and only one of them justifies the deduction you were preparing to ask for. A seller confident it is the first will usually fund the test to prove it.

Then hold on to some perspective. Cooling is comfort, not roadworthiness. A car that is sound underneath with a dead compressor is frequently the better purchase against one with freezing vents and a past nobody has examined, which is the argument for looking up a vehicle’s recorded background well before you spend an afternoon on its air conditioning.

Common questions

Why is replacing an air conditioning compressor so expensive?

Because it is usually not one part. If the compressor failed internally it has shed metal into a sealed circuit with no filter, so a correct repair also replaces the drier or accumulator and the expansion valve or orifice tube, flushes or replaces the condenser, checks the lines for internal baffles that cannot be cleaned, meters the lubricating oil back in, evacuates the system under vacuum and recharges it with recovered or new refrigerant on certified equipment. The compressor is one line on that list. The rest is what stops the new one failing the same way.

Can I replace just the compressor and skip the rest?

You can, and it is the reason some quotes come in far below others. What you are buying is a new pump installed into a circuit that still contains the debris from the old one, which the new pump will draw in through its suction port as soon as it starts turning. It may last a season. Most remanufacturers make replacement of the drier and the metering device, and flushing of the system, a condition of their warranty, so a failure afterwards is likely to be yours to pay for a second time.

Is it the compressor or just the clutch?

They are separate components and the difference is most of the cost. A worn clutch, a failed coil, a tired relay, a blown fuse or a sticking pressure switch all produce a compressor that never engages, while the pump itself is fine. Noise helps: a rumble that is present whether or not cooling is switched on is the pulley bearing, whereas grinding that appears only when the system is engaged comes from inside. The only reliable separation is a technician confirming the clutch is being energised and then reading pressures on both sides of the system.

What does it mean if the compressor clutch clicks on and off rapidly?

That the refrigerant charge is low. The low-pressure switch is cutting the clutch out to protect the compressor, pressure recovers, it cuts back in, and the cycle repeats every few seconds. It is a leak, and it is also actively damaging: each engagement is a shock load, and a compressor running on a low charge is running on reduced lubrication, because the oil is carried around in the refrigerant. Switch the system off and have it looked at rather than driving on it.

Does the refrigerant a car uses change what this costs?

Yes, and before anybody touches the car. Three generations are in circulation. CFC-12 came first, HFC-134a covers the long middle of the fleet, and newer vehicles use refrigerants of lower global warming potential, principally HFO-1234yf. They are priced nothing like one another and the newest is much the dearest. A compressor job empties and refills the entire system, so the charge is a real line on the invoice rather than a rounding error. The label under the bonnet states which one the car takes.

The air conditioning does not work on a car I want to buy. How much should I knock off?

Nothing, until somebody has put gauges on it. An empty system with a good compressor, a failed relay, a dead clutch coil and an internally destroyed compressor that has contaminated the whole loop all look exactly the same from the driver’s seat, and they span an enormous range of cost. Get an independent written diagnosis before you negotiate, and negotiate against the document rather than against the symptom.

Do air conditioning compressors get recalled?

Rarely enough that you should not plan on it, and there is a structural reason. Air conditioning falls to the EPA, as a refrigerant and emissions question, rather than to the safety regulator. Nothing about it is therefore recorded in the campaign and complaint files most repair questions can be tested against. The archives this site uses elsewhere contain no air conditioning component category and no climate control recall family at all.