Engine Misfire and Rough Idle: Read It by Condition

An engine is a set of identical machines bolted to a common crankshaft, taking turns. Each one fills, compresses, burns and empties in its own time, and the whole arrangement runs smoothly because the contributions arrive evenly spaced and roughly equal in size. Take one of them away and the engine does not stop. It develops a limp.

That limp is the most legible mechanical fault available to somebody looking at a used car, and also the easiest to walk past. It is legible because it is tied to conditions rather than to luck: put the engine in the state that exposes it and it appears, reliably, for a stranger. It is easy to walk past because almost nobody puts the engine in that state during a viewing.

The figure above is the reason this page is organised the way it is. When a misfire shows up is not incidental detail to be mentioned after the diagnosis — it is a large part of the diagnosis, because the conditions that reveal one are the same conditions that separate its causes from each other. A fault that only appears at idle and a fault that only appears under a hard pull are being told apart by physics rather than by opinion, and the observation costs you nothing but the discipline to make it deliberately. The sections below take each cause in turn, at more length than a figure can hold.

What a misfire actually is

Strip away the diagnostic vocabulary and a misfire is a cylinder that did not contribute. Something was drawn in, the piston came up, and either nothing burned at all or what burned did so too slowly, too late, or too weakly to push back properly.

The crankshaft is where this becomes measurable. Every firing event gives the crank a shove, and a healthy engine produces a train of shoves of comparable size at even intervals. When one is missing, the crankshaft momentarily stops accelerating the way the engine management system expects, and it decelerates fractionally before the next cylinder picks it up. Modern engines watch for exactly that. A toothed wheel on the crankshaft passes a sensor thousands of times a minute, and the module measures the time between teeth precisely enough to notice a single weak contribution among hundreds of strong ones.

Two consequences follow, and they matter more than they sound.

The first is that the car frequently knows before you do. Misfire counting runs continuously, and a rate low enough to be entirely undetectable from the driver’s seat still accumulates in the module’s memory. That is why a scan tool at a viewing can hand you a fault that the car has never once demonstrated on a test drive, and why the argument for reading stored and pending faults rather than only the ones illuminating a lamp is stronger here than almost anywhere.

The second is that the fuel does not vanish. A cylinder that failed to burn its charge still opens its exhaust valve on schedule and pushes that charge out, unburnt, into the exhaust system. Everything downstream then has to deal with it. That single fact is behind the flashing light, behind the smell, behind the sudden appearance of an expensive part in a bill that started out as a cheap one, and it is dealt with properly further down this page.

Worth separating from all of this: an engine that runs unevenly is not automatically an engine with a misfire. Some of the roughest idles on the road are mechanically fine and badly isolated, and a few are simply how that engine was built. There is a section on that too, because getting it wrong in either direction costs money.

Three ingredients, and what removes each one

Combustion in a petrol engine needs three things to arrive together: a mixture of air and fuel in roughly the right proportion, enough compression to raise it to a temperature and density where it will burn quickly, and a spark at the correct instant. Remove any one and the cylinder goes quiet. The whole diagnostic exercise is working out which one went missing, because the answer decides everything about what happens next.

There is a fourth item that does not fit the list because it is not an ingredient so much as a conductor: the control and timing side. A crankshaft or camshaft position sensor reading badly, a chain that has jumped a tooth, or a variable valve timing actuator stuck out of position can produce misfire behaviour with the coils, the injectors and the compression all perfectly serviceable. It belongs in your thinking because it is the category that most often survives a round of parts replacement untouched.

Notice what the table does and does not tell you. It maps a cause to a symptom, but the mapping is not one-to-one, which is why nobody can diagnose a misfire from a description of it over the telephone. What the mapping is good for is ordering the investigation, and ordering matters enormously when the cheapest candidate costs a fraction of the most expensive one.

The candidates, cheapest and likeliest first

Two warnings about the sequence below. It is descriptive, not statistical: no public archive counts misfire causes by frequency, so a page quoting you percentages for them has invented the percentages. And it is a working order rather than a ranking — the sequence a competent diagnostician moves through, which is a weaker claim than a measurement and a considerably more honest one.

Ignition coils

On almost anything built in the last twenty-five years, each cylinder has its own coil sitting directly on top of its plug. That layout removed a great deal of high-voltage wiring from engines and, in doing so, made the individual coil the single most common cause of a single-cylinder misfire.

A coil is a transformer that lives in a hot, vibrating place and is asked to produce a very high voltage many times a second, for years. Its failure mode is usually gradual and heat-related: it works when cold, it works at light load, and then it breaks down internally when it is hot and being asked for its highest output. This produces one of the most characteristic patterns in the whole subject — a car that is fine around town and stumbles badly on a motorway climb, or fine for the first ten minutes and rough thereafter.

The diagnostic move that makes coils so identifiable is that they are interchangeable. Move the suspect coil to a different cylinder, clear the memory, and see whether the fault follows it. If the misfire moves with the coil, the coil is the fault. If it stays where it was, the coil is exonerated and you have learned something for free. This is the reason a competent workshop can settle a coil question in minutes and why a quotation that skips it deserves a question.

Spark plugs

A plug stops a cylinder firing in one of two ways, and they are worth keeping apart because they behave differently on the road. The first is erosion. Every spark takes a little metal off the electrodes, so the gap slowly opens, and a wider gap demands a higher voltage to bridge — which loads the coil behind it harder every year until one of the pair concedes. That kind of failure shows itself under load first, for the same reason a weak coil does. The second is fouling: a film of oil, fuel or deposit across the insulator nose offers the spark an easier route to earth than the jump it is supposed to make. Fouling is at its worst cold and at low load, so it announces itself the other way round entirely, as a misfire on start-up that improves once the engine is hot and working.

Two related subjects sit deliberately outside this page. What plugs cost to change, how long they are meant to last, and what the appearance of a removed one reveals about the cylinder it came from all belong to the page on plug replacement. What belongs here is narrower and more sceptical: when a seller attributes a rough idle to plugs, that is a diagnosis with a convenient price attached, and it is fair to ask what evidence stands behind it.

Fuel injectors

An injector is a solenoid valve opening for a period measured in thousandths of a second. It fails in three broad ways: it partially blocks, so the cylinder is fed less than the module intended; it leaks, so the cylinder is over-fuelled and the mixture goes too rich to light cleanly; or the coil inside it fails electrically and it stops opening at all.

Direct injection has made this category more prominent than it was. Direct injectors work at far higher pressures, sit in the combustion chamber where they are exposed to heat and deposits, and are considerably more expensive than the port injectors they replaced. On direct-injection engines there is also a second, related problem: because fuel no longer washes over the back of the intake valves, carbon accumulates there over years and disturbs airflow into the cylinder. That produces a rough idle and a misfire that no injector, coil or plug will fix, and it is a well-known characteristic of certain engine families rather than a fault in a particular car.

A rich-running cylinder from a leaking injector often has a visible signature at the tailpipe, and reading that signature is a useful cross-check; our guide to what the colour of exhaust smoke tells you covers what black, blue and white each imply and how to tell steam from something worse.

Vacuum leaks and unmetered air

The engine management system meters the air entering the engine and doses fuel to match it. Air that gets in without being measured breaks that arrangement, and the engine ends up running leaner than the module believes it is.

Sources are numerous and mostly cheap: a perished intake hose, a cracked plastic elbow, a hardened intake manifold gasket, a positive crankcase ventilation hose that has gone brittle, a split vacuum line to the brake booster, an inlet manifold with a warped face. What makes them distinctive is not where they are but when they matter. At idle the manifold is at its strongest vacuum and the volume of air being metered is at its smallest, so an unmetered leak is proportionally enormous. Open the throttle and the metered flow rises by an order of magnitude while the leak stays roughly the same size, so the same fault becomes proportionally trivial.

That is the whole reason a vacuum leak is a classic rough-idle complaint from a car that drives perfectly well. It is also why chasing one on a test drive at speed is a waste of a drive.

Compression, which is the expensive end

Everything above is a part that bolts on. Compression problems are the engine itself, and they are the reason a misfire should never be assumed to be cheap before it has been investigated.

A valve that no longer seals — burnt, bent, or held open by a failed component in the valvetrain — lets the charge escape as the piston rises. A head gasket that has failed between a cylinder and a coolant passage does the same and adds coolant into the combustion chamber, which is its own set of symptoms. Worn rings and bores allow the charge past the piston. A timing chain that has stretched enough to move the camshaft relative to the crankshaft opens and closes the valves at the wrong moments across the whole engine, which tends to present as a rough, weak engine rather than as a tidy single-cylinder fault.

The tell, more often than not, is that a compression-related misfire is boringly consistent. It does not care whether the engine is hot or cold, loaded or idling, wet or dry. It is present from the first firing and it stays present. Ignition and fuel faults are far likelier to come with a condition attached, and that difference is the observation this whole page is built around.

Under load but not at idle, and the other way round

If you take one diagnostic habit away from this page, make it this one, because it does more work than any other single observation and it requires no tools at all.

A misfire that appears under load and disappears at idle is pointing at the ignition side, and at the parts that have to work hardest when the cylinder is full. Here is why. Under a hard pull the cylinder is packed with a dense charge, and dense charge is harder to ionise, so the spark needs a substantially higher voltage to jump the same gap than it does at a light-throttle idle. A coil that is failing internally can produce enough voltage for the easy job and not enough for the hard one. A plug with an eroded gap does the same thing from the other direction. The result is a car that idles like anything else in the car park and stumbles the moment you ask it to climb a hill in a high gear.

A misfire that appears at idle and vanishes under load points the other way, towards mixture and towards air. At idle the throttle is nearly shut, the manifold vacuum is high, exhaust gas recirculation is often active, and the total air being metered is tiny. Anything that disturbs a small quantity of air — a leak, a valve not closing, a carboned intake port — is proportionally large. Open the throttle and the disturbance is swamped.

A misfire at a steady light throttle, and nowhere else, is the third condition, and it is the one drivers most often fail to notice at all because it is subtle and it happens during the least demanding kind of driving. Cruising at constant speed is where the engine runs its leanest mixtures and where exhaust gas recirculation is at its most aggressive, both of which make combustion less certain. A cylinder that is marginal for any reason will show its marginality here first, as a faint hesitation or a periodic tug that a driver will happily attribute to road surface for months.

None of these three is a proof. A coil can fail badly enough to misfire in all conditions, and a large vacuum leak will make itself known everywhere. They are a way of narrowing the field before anybody spends anything, and their value on a used-car forecourt is that they cost only the discipline to drive the car in three different ways instead of one.

A steady misfire and a random one are different problems

This distinction is the one that separates a bounded repair from an open-ended one, and it is available to anybody with a basic scan tool.

When the module identifies a misfire it also identifies where it happened, because it knows the firing order and it knows which cylinder should have been contributing at the moment the crankshaft slowed. It records a fault against that specific cylinder. Where the misfires are spread across several cylinders, or where the module cannot attribute them cleanly, it records a general misfire fault instead.

A misfire attributed to one cylinder is good news, relatively speaking. It says the problem is local. Something belonging to that cylinder alone — its coil, its plug, its injector, its valves, its compression — is at fault, and the investigation has a small, defined scope. The coil-swap test settles a large fraction of these in one move.

A misfire spread across several cylinders, or recorded as random, says the problem is shared. Something common to the whole engine is marginal: fuel pressure that has dropped, a large unmetered air leak, timing that has slipped, an exhaust that is restricted, a mass airflow sensor under-reading, low compression across the board on a tired engine. These are broader investigations and they include some of the most expensive answers available.

The pattern that should worry a buyer most is faults recorded against several cylinders that are neighbours in the block rather than neighbours in the firing order. Adjacent cylinders share a piece of head gasket, and a gasket that has failed between two of them will misfire both. That is not a coil job.

One further nuance worth knowing before you read too much into a scan. Some modules record a general misfire fault at the same time as a cylinder-specific one, simply because the rate crossed both thresholds. Seeing both is not evidence of two faults. What matters is whether the cylinder-specific entries cluster on one cylinder or scatter.

When a rough idle is not a misfire at all

Plenty of rough idles have nothing wrong with combustion. Diagnosing them as misfires is how people end up with a new set of coils and a car that shakes exactly as much as it did before.

Idle control and throttle body

At idle the engine is being held at a speed by a control system, not left to its own devices. On modern cars the throttle itself is opened a fraction by a motor under the module’s command; on older ones a separate idle air control valve bypasses the closed throttle. Either arrangement collects carbon on the throttle bore and on the plate edge over the years, which changes the relationship between commanded position and actual airflow. The result is an idle that hunts up and down, or sits low and lumpy, or dips towards stalling when the air conditioning compressor engages. Nothing has misfired. The engine is simply not being fed a stable quantity of air.

Motor mounts, which make everything feel worse

An engine shakes by design and the mounts exist to keep that shake out of the cabin. When a mount collapses, the same engine vibration arrives in the seat and the steering wheel at several times the amplitude, and a driver reasonably concludes the engine has developed a fault. It has not. The isolation has.

Getting this wrong is expensive in both directions, so it is worth doing properly rather than by feel. The full separation of mount problems from combustion problems, and the checks that tell them apart at a standstill, belong to our page on why a car shakes and what the shaking is telling you, which also covers the vibrations that have nothing to do with the engine at all. What is worth saying here is only this: a genuine misfire usually brings supporting evidence — a stored fault, a changed exhaust note, a change in how the car pulls — while a mount problem brings none of that and simply amplifies something normal.

Exhaust gas recirculation stuck open

The recirculation valve deliberately admits inert exhaust gas back into the intake to lower combustion temperature. It is supposed to be shut at idle. When carbon holds it open, the engine idles on a diluted charge that burns badly, and the result is a lumpy, unstable idle that clears as soon as you drive, because at higher load the recirculated fraction is meant to be there. This one sits on the boundary: it genuinely can cause a misfire, and it also causes rough idles that never trip a misfire fault at all.

Ordinary character

Some engines are simply rougher than others at idle and always were. Engines with an odd number of cylinders have inherently uneven firing intervals. Engines with cylinder deactivation shut down half of themselves at light load by design. Diesels idle with a clatter that would be alarming in a petrol car and is entirely normal in a diesel. And an air conditioning compressor cycling in and out puts a periodic load on the engine that produces a regular dip in idle quality on a perfectly healthy car.

Before using idle quality as an argument about price, it is worth finding out what that engine is supposed to feel like. Sitting in another example of the same model for two minutes is a more reliable calibration than any amount of theory.

The flashing light, and what to do in the next minute

Everything above is diagnosis at leisure. This section is not.

The reason sits inside the exhaust. A catalytic converter is a ceramic honeycomb carrying a wash of precious metals, and it works by oxidising what reaches it. It is designed to do that with the small residue an engine in good order produces, at temperatures it can survive indefinitely. A misfiring cylinder does not send it a residue. It sends the entire contents of a cylinder, unburnt, several times a second, and the converter oxidises that too — releasing all of the energy that should have been released inside the engine, inside a ceramic structure with no cooling and no room to expand.

What happens next is thermal. The substrate softens, the coating sinters and loses its surface area, and in a bad case the honeycomb slumps and partially closes. A converter that has been through this is not merely worn out; it is a restriction in the exhaust system. That produces a second symptom set entirely — an engine that will not rev cleanly, feels choked under load and gets hot — which is routinely diagnosed as something else, because by then the misfire that caused it has been fixed and forgotten.

Two practical consequences a buyer should hold onto.

The first is that a misfire driven on becomes two repairs rather than one, and the second one is usually the larger. Fixing the coil does not un-melt the converter. It is entirely normal for somebody to spend a modest sum on the misfire, feel the car improve, and then discover months later that they also own a catalyst problem and, on many cars, a failed emissions test.

The second is more subtle and it is the reason people keep driving. The flashing frequently stops when you lift off. Misfire rate is not constant; it is worst under load, so easing the throttle can drop it below the threshold that triggers the flash, and the lamp settles to a steady glow. The car feels like it has recovered. It has not. It has returned to a rate that is destroying the converter more slowly, which is a distinction with no practical value to you. Treat the first flash as the message, not the current state of the lamp.

Our page on what a check engine light actually means covers the lamp as a system — what a stored code identifies, why it names a circuit rather than a part, and the readiness-monitor check that reveals whether a seller cleared it this morning. This page is about the one behaviour of that lamp which is an emergency.

One qualification, in fairness to the lamp. A few manufacturers flash it for other severe conditions, so the honest reading is that a flashing lamp means stop first and identify afterwards. That does not weaken the rule; it broadens it.

Putting the causes on a mileage axis

The constraint has to be stated before any figure is, because a reader has no way of auditing a number on a page like this one.

Owner reports reaching the federal safety regulator get indexed against whichever vehicle system was involved, never against what the driver noticed. No misfire heading exists. No rough-idle heading exists. Nowhere on the intake form is there a place to record how an engine behaved at a red light. So nothing below is a count of misfires, and anyone offering you such a count did not get it from here. What the file will support is narrower: it shows where, on the odometer, each of the systems capable of producing a misfire reached the point at which somebody sat down and wrote to a regulator about it — a later, rarer and much higher-threshold event than a fault simply existing.

Hold the three ingredients in mind while reading down that column and a shape appears. Everything on the ignition and control side of a misfire lives under two headings, and both of them come early: electrical at 49,000 and engine at 58,000, both of them below the 60,000 that characterises the pool as a whole. Fuel arrives a very long way afterwards, at 89,000, and cooling later still at 98,000. The ordering is the only thing here worth trusting — every heading came out of one pull and is distorted in identical ways, which is precisely what makes them comparable with each other and with nothing else.

The usable version of that, for somebody with a car in front of them, is a matter of expectations. Ordinary used cars change hands in exactly the mileage band where these owners were already filing about electrical and engine faults. So a car that will not hold a clean idle at that sort of reading does not need an unusual explanation; it is behaving the way the archive says vehicles of that age behave. The reverse inference is the one to refuse. A lower reading has not exempted anything, none of these figures forecasts a particular car, and not one of them is a service interval.

The recall record, and the absence at the centre of it

Complaints are testimony from an owner. A campaign is an admission from the other side of the transaction: the manufacturer accepts, in writing and at its own cost, that a build carried a defect out of the factory gate. The two archives sit at different evidential weights and should never be pooled.

What this page takes from that archive is something it does not contain. Eighty distinct component paths sit beneath those campaigns, and there is no ignition coil among them, no spark plug, and no ignition system heading of any kind. Search the campaign text and the word misfire does not occur once. The families that do appear are about containment and structure rather than combustion. Gasoline fuel system is the widest-spread family in the whole archive at sixteen campaigns from ten separate manufacturers, but read inside it and eleven concern the fuel pump, four concern hoses, lines, piping and fittings, and exactly one is filed against the fuel injection system. Engine itself is a blunt little family of five campaigns from three manufacturers with no part named inside any of them. Next door, engine and engine cooling holds eight campaigns from four manufacturers, and two of those name the exhaust gas recirculation valve — the only part named anywhere in the archive that this page has already described as a cause of a lumpy idle.

None of that is missing data. It follows directly from what a campaign is for, which is removing an unreasonable risk to safety — a fire, a stall in traffic, a loss of control. Coils that break down after a few years, electrodes that erode, injectors that gum up, throttle bores that carbon over: these are consumption and service matters, not defects. The manufacturer’s response to them lives in technical bulletins, quietly revised part numbers, extended coverage and goodwill claims, and none of that machinery is visible to anybody typing a VIN into a public lookup.

Which produces a conclusion that runs against instinct and is worth having in your pocket: a spotless recall record predicts nothing at all about whether a car misfires. Do the search regardless — an open campaign is corrected for nothing by a franchised dealer no matter how many owners the car has been through, so it is the one genuinely free repair available to a used-car buyer. Just do not allow it to substitute for the drive or the inspection. The recorded past of one specific car is a different enquiry again, and two minutes spent to look up what the VIN record holds is time spent before a price is discussed rather than after.

Feeling for it on a test drive

This is the section the whole page has been building towards, and it is addressed to a reader who is not repairing anything: someone with fifteen minutes, a stranger’s car and a decision to make about it.

Few faults on a car are as easy to detect as a misfire, and few are as easy to drive straight past. The entire difference lies in how those fifteen minutes are spent. Exposing one requires a settled idle, a hard pull in a tall gear, and a long stretch of constant speed on barely any throttle. A slow circuit of the neighbourhood and back contains none of the three. Most test drives are a slow circuit of the neighbourhood and back.

Before the engine starts

• Switch the ignition on without cranking and watch the cluster light up. Every warning symbol should appear and then extinguish once the engine runs. A check engine symbol that never appears at all is not a car without faults; it is a bulb somebody has removed, taped over or unplugged.
• Feel near the tailpipe and across the bonnet before you greet anybody. A warm engine has already taken away your best evidence, for reasons the next section sets out.
• Ask to turn the key yourself. You want to watch the dashboard rather than the seller, and you want to hear the first thirty seconds from the driver’s seat.

The idle, which is where most of it lives

Let the engine settle and then give it two full minutes of your attention with everything switched off — radio, fan, air conditioning. Two minutes feels absurdly long standing next to a stranger. Do it anyway.

What you are hunting for is a period — a rhythm. A healthy idle is a continuous texture with no beat in it. A misfiring one has a pulse, because the absent contribution comes round on a fixed cycle and the engine dips every time it does. Rest a flat hand on the inner wing or the top of the engine and the periodicity is often clearer through your palm than through your ears. Watch the mirror, the shut line of the bonnet, a paper cup left on the roof. If the car is an automatic, hold the brake hard and engage a gear so the engine is carrying a small load, then listen again; a few marginal misfires only surface once something is being asked of the engine.

Then, still stationary, raise the engine speed a little and hold it steady for several seconds. A rough idle that smooths out completely when held above idle and returns when released is telling you the fault is idle-specific, which points at the air side and away from the ignition side. That is a free piece of diagnosis, performed in a parking space, before you have driven anywhere.

On the road, three conditions on purpose

• Full load in a high gear. From a low speed in the highest gear the car will accept, open the throttle wide and let it pull. This is the single most revealing thing you can do to an ignition system, because it demands the highest cylinder pressure and therefore the highest spark voltage of any normal driving. On an automatic you may need to hold a gear manually to stop it downshifting and making the job easy.
• Steady light throttle, held. Find a stretch where you can hold a constant speed for a couple of minutes with barely any throttle. This is the lean-cruise condition, and the misfire it exposes is a faint periodic tug rather than a stumble. It is easy to miss with a passenger talking, which is a reason to ask for quiet.
• Repeat the idle check hot. Come back, let it idle again with the engine fully warm, and compare it with what you felt at the start. Coil breakdown is heat-related, so an engine that is rougher after twenty minutes than it was at the beginning has told you something specific.

Two habits raise the value of all three. Keep the stereo and the blower switched off from start to finish, since a good deal of what this page describes reaches the ear some time before it reaches the fingertips. And write the result of each condition down before you climb out, because memory edits detail of this kind within minutes, and it edits in the direction of whatever you were hoping for.

What to do with a positive finding

Feeling a misfire is not grounds for abandoning the car. It is grounds for refusing to guess any further. The distance between the cheapest item on this page and the dearest one is enormous, the sensation reaching the driver is close to identical across that whole span, and no seller alive — honest or otherwise — can tell you which end of it you are standing at.

Turn it into a written diagnosis, then, and do it before any money moves. A workshop with the car raised can interrogate every module for stored and pending faults, watch live misfire counters cylinder by cylinder with the engine running, move a coil to see whether the fault travels with it, and where that still leaves the question open, measure compression directly. Naming the symptom at the point of booking converts a general look-over into a specific question with a specific answer. How to brief a workshop, and what a full multi-module scan reaches that a free read at a parts counter never will, is set out in our guide to the inspection carried out before you buy.

Why a warm engine on arrival matters here more than anywhere

The general argument for insisting on a cold start — cranking speed, time to fire, smoke on the first firing, cold knocks that quieten as oil pressure builds — is made at length in our page on a car that refuses to start, and there is no point restating it. This section takes a narrower slice: idle quality by itself, where the effect of a pre-warmed engine is larger than almost anyone expects.

A cold engine is running an enriched mixture, a raised idle speed, and a set of compensations designed to get the catalyst up to temperature quickly. Almost every rough idle is at its worst during that phase and improves as the engine reaches operating temperature and the module hands control back to its normal strategy. So a car started twenty minutes before you arrived is not merely warm. It has been driven past the window in which its worst behaviour occurs, and it will present to you in the condition its owner has learned it presents best in.

None of which is necessarily calculated. People move cars out of garages, collect them from a relative’s driveway, and run them up out of plain courtesy to a visitor. The effect on your evidence is the same whatever drove it, though, so the useful response is to name the situation without turning it into an allegation. Say that you had wanted to hear it started cold and ask whether a second visit is possible. Then listen to the answer rather than assessing the excuse. Innocent explanations arrive fast and require no work. The reply worth noting is the one arguing that a cold start does not really tell you anything, because anybody who has lived with a lumpy idle understands perfectly well that it does.

Where a return trip truly cannot happen, run the visit backwards: paperwork, service record, tyres and bodywork first, with the engine switched off from the moment you arrive and left off for as long as you are on site. An hour of standing is not a cold start. It is a very great deal better than firing up a hot engine two minutes after shutting it down, and anything that reappears after that hour has earned your attention. While the car sits there, remember that its history and its mechanical condition are two separate enquiries with two separate sources — a car can idle like a sewing machine and still be entirely the wrong purchase — so pulling a history report against the VIN answers a set of questions that no amount of listening to an engine ever will.

What a misfire should do to the price

Two situations produce the same symptom and deserve opposite responses.

The first is a bounded, quotable repair: a coil, a set of plugs, an injector, a cracked intake hose, a throttle body clean. These are ordinary running costs on a used car, they can be priced in advance by anybody with the car in front of them, and the sensible move is to put a number on them and put that number into the offer. A seller who has already had the work done and can show you the invoice has given you something better than a discount, which is evidence.

The second is unbounded: compression that is down on a cylinder, a head gasket that has started to go, timing that has moved, a converter that has already been cooked by a misfire somebody drove on. These do not have a ceiling that can be established from the driver’s seat, and on an older car several of them exceed what the car is worth. The symptom you can feel does not distinguish between the two situations. Only a diagnosis does.

There is one more question worth asking whenever a seller volunteers that a misfire has already been fixed, and it is the question this page exists to make you capable of asking. How long was it driven before it was fixed, and did the lamp ever flash? A misfire repaired the same week is a repaired misfire. A misfire repaired after a month of commuting on a flashing lamp is a repaired misfire and an exhaust system that spent a month being cooked, and the second half of that has usually not been looked at. It is a fair question, asked pleasantly, and the answer is informative whichever way it goes. Keeping ahead of exactly this kind of compounding is most of what ordinary used car maintenance is for.

Common questions

What does an engine misfire feel like?

At idle it is a periodic stumble or shudder rather than a continuous vibration — a beat you can often feel through a hand on the wing more clearly than you can hear it. Under acceleration it is a hesitation, a jerk, or a sense that the engine has gone flat and is not delivering what your foot is asking for. At a steady cruise it can be so subtle that drivers put it down to road surface for months. The exhaust note usually changes too, taking on an irregular quality, which is easier to notice with the windows down and the radio off.

Why does my car idle rough but drive fine?

Because idle is the condition in which small disturbances are proportionally largest. The throttle is nearly closed, the quantity of air being metered is tiny, and manifold vacuum is at its highest, so an unmetered air leak, a recirculation valve that will not shut, or a carboned throttle body all matter enormously. Open the throttle and the same disturbance is swamped by the real airflow, so the car drives normally. It can also be a mount problem, in which case nothing is misfiring at all and the engine vibration is simply reaching the cabin instead of being absorbed.

Can I drive with a misfire?

With a steady check engine light and no obvious loss of power, briefly and gently, while you arrange to have it diagnosed — but the clock is running, because unburnt fuel is reaching the catalytic converter the whole time. With a flashing light, no. A flashing lamp says the misfire is currently severe enough to be wrecking the converter, and the answer is to lift off the throttle, get somewhere safe, switch off, and arrange for the car to be carried to a workshop instead of driven there.

What is the most common cause of a misfire?

On modern engines, an ignition coil, followed by spark plugs and then injectors. That order is descriptive rather than measured, and the reason it holds is layout: almost every car built in the last twenty-five years puts an individual coil on each cylinder, in a hot and vibrating place, working hard for years. The reason it matters is that the coil is also the easiest cause to confirm, because coils are interchangeable and moving one to another cylinder shows within minutes whether the fault travels with it.

Does a misfire always turn on the check engine light?

No, and this catches people out. The module counts misfires and compares the rate against a threshold, so a low-level misfire can run for a long time without illuminating anything while still being recorded internally. A fault can also be stored as pending — seen once, waiting for confirmation — without lighting the lamp at all. It is exactly why a scan reaching stored and pending faults is worth more at a viewing than an inspection of the dashboard, and why a dark warning lamp opens the enquiry rather than closing it.

Why does the misfire only happen when the engine is hot?

Because heat is what finishes off a failing ignition coil. The insulation inside it breaks down as it warms, so the coil delivers adequate voltage cold and fails to deliver enough once it is up to temperature and being asked for its hardest work. That produces the classic pattern of a car that is fine for the first ten minutes and rough thereafter, or fine around town and stumbling on a long climb. The mirror image — rough when cold and better once warm — is more often mixture, sensor or compression related, since a cold engine is running enriched and compensating heavily.

Should I buy a used car that misfires?

Only with a written diagnosis in hand, never on the seller’s estimate of what it needs. The same stumble covers everything from a coil to a failed head gasket, the difference between those two outcomes is the difference between a small bill and more than the car is worth, and nothing available to you from the driver’s seat separates them. Have it inspected with the symptom named in the booking, ask specifically whether the lamp ever flashed and for how long it was driven, and make the report rather than the drive the basis of the conversation about price.