Car Care

Car Shaking When Driving: When It Happens Tells You What It Is

A close-up of an alloy wheel and tyre on a parked car, the spokes and sidewall filling the frame

The short version

  • Three questions settle most vibrations before anybody opens a bonnet: at what speed does it happen, what is the car doing at the time — braking, accelerating, idling, holding a steady cruise — and whereabouts in the car do you feel it.
  • Where you feel it separates front from rear. A shaking steering wheel is being fed by something at the front of the car. The identical fault at the back reaches you through the seat base and the floor instead, because a rear wheel has no mechanical path into the steering column.
  • What the car is doing when it happens narrows the rest. Only while braking is the discs. Only at a standstill is the engine or what holds it. Only under power is the driveline. Always, and getting worse the faster you go, is usually a bearing.
  • Sudden onset is its own piece of evidence. A vibration that appeared the morning after a pothole or a kerb strike means something changed shape overnight, and the part that changed shape is not always the part you can see.
  • A vibration that stopped on its own is only good news when you can name the thing that left the car. A few of them genuinely resolve. Others go quiet because the fault has moved on to its next stage.
  • In our archive of NHTSA owner complaints, the median odometer reading at the moment of reporting is 53,000 miles for suspension, 64,000 for steering and 60,000 for hydraulic service brakes. Those are reporting mileages drawn from a censored sample rather than failure rates, and the ordering between them is the part worth reasoning from.

Most faults on a used car are shy. They wait for a cold morning, for a motorway run long enough to get everything hot, for the third week of ownership. A vibration is the exception. It is bolted to a condition, it returns every time that condition returns, and it will perform for a stranger on a short test drive as reliably as it performs for the person selling the car.

That makes it the most legible fault available to a buyer, and the legibility comes from timing rather than from mechanical knowledge. Almost every vibration is a rotating part putting a force into the body of the car once per revolution, or a system reacting to one, so the moment it appears and the route it takes to your hands or your back are already telling you what turns at that rate and what does not.

Reading a vibration by when and where it happensA two-column figure pairing each of six vibration symptoms with the group of faults that symptom points at.WHEN YOU FEEL ITWHAT IT NARROWS TOSteering wheel, one speedbandFront wheel balance. It appears in a windowand eases above and below it.Felt in the seat, not thewheelSame fault at the other end of the car. Rearwheels rather than front.Only under braking, pulsingDiscs, not tyres. Uneven thickness ordeposits, felt through the pedal as well asthe wheel.At idle, gone once movingEngine mounts or a misfire. The car is shakingbecause the engine is.Under acceleration onlyDriveline. Joints, shafts, or mounts lettingthe engine move under load.Constant, rising withspeed, droningA wheel bearing. It gets louder on the turnthat loads it.
A vibration is the most legible fault on a test drive, because when it happens tells you most of what it is. Establish the answers to these before anyone puts the car on a lift, and the diagnosis is largely done.

Read the figure as three separate strands of evidence collapsing into one answer: speed, condition, location. Any one of them on its own leaves a wide field. Two of them together usually leave two or three candidates. All three together are frequently enough to name the part, which is why it is worth being deliberate about noticing all three rather than reporting “it shakes” to a workshop and paying somebody to find out what you could have observed for free on the way there.

The rest of this page takes those pairings apart one at a time, at more length than a two-column figure can hold, and then turns the whole thing around for somebody standing next to a car with the keys in their hand and a decision to make.

The three questions

Ask them in this order, and write the answers down before you get out of the car. Memory rewrites this kind of detail within minutes, usually in the direction of whatever you already suspected.

At what speed?

The useful distinction is between a vibration that lives in a band and one that never stops. A band means you can drive into it, feel it build, and then drive out the other side and have the car go smooth again. That behaviour is the signature of an out-of-balance rotating assembly: the force it produces grows steeply with rotational speed, and the vibration peaks where that forcing frequency meets a resonance somewhere in the steering or the suspension. Above the band the two fall out of step again and the car settles.

A vibration with no band — one that starts early, grows steadily, and never improves — is a different family. So is one that only exists at a standstill. So is one that only exists in the two seconds a gear is engaged.

Under what condition?

This is the highest-value question of the three, because it is the one that separates parts that would otherwise feel identical.

Hold a steady speed with light throttle. Then lift off entirely and coast through the same speed. Then apply the brakes gently through it. A vibration that is present in all three is tied to road speed, which points at wheels, tyres, hubs and the parts that turn with them. One that appears only while you are on the throttle is load-dependent, which points at the driveline and at whatever is supposed to stop the engine and gearbox moving. One that appears only under braking is the brakes, and effectively nothing else.

There is a second version of this test for anything you suspect of being engine-related. At a standstill, with the parking brake on, raise the engine speed gently and hold it. If the vibration tracks engine speed rather than road speed, the wheels are not involved at all.

Where do you feel it?

The steering wheel, the seat, the floor, the brake pedal, the gear lever, or the whole car at once. This is the question people answer least precisely and it carries more information than either of the others.

Take your hands off the wheel for a moment where it is safe to do so, and notice whether the shake continues in your hands or only in the seat. Then put one hand flat on the centre console and one on the door card. A fault at the front of the car has a short, stiff mechanical path into the steering rack and up the column to your hands. A fault at the back has no such path, so it arrives through the body shell and reaches you through the seat, the floor and your back.

Steering wheel, seat, or pedal

Because this single observation does so much work, it is worth being clear about what it does and does not prove.

A steering wheel that shakes is almost always reporting on the front of the car. The wheels, tyres, hubs, bearings, brake discs and steering joints at the front axle all sit on the same short load path into your hands. That is why “my steering wheel shakes” and “my car shakes” are so often the same fault described by two different drivers — the first has the fault at the front, the second has it at the back.

A seat and floor vibration with a still steering wheel points at the rear axle. Rear wheel balance, a rear tyre with a fault in its structure, a rear bearing, rear brake discs. None of them can reach the steering column, so they announce themselves as a buzz through the seat base that many drivers put down to road surface for months before it gets bad enough to investigate.

A pulse felt specifically in the brake pedal, with the pedal moving under your foot, is coming from the brakes rather than from anything else, and its front-to-rear position follows the same rule: front discs push it up the steering column as well as the pedal, rear discs tend to arrive as a pedal pulse and a shudder through the body with the wheel comparatively calm.

Two things that muddy the reading, and how to allow for them. First, a big enough vibration anywhere shakes the whole car, so the location test works best on a fault that is still mild. If everything is shaking, get the speed and condition answers instead and come back to location later. Second, front-wheel-drive cars put the engine, gearbox, driveshafts and steering at the same end, so the front-of-car finding still leaves several families in play. The condition question is what separates them.

The speed band that comes and goes

Steady cruise, no braking, no particular throttle demand, and a vibration that fades in and back out again over a stretch of the speedometer. This is wheel balance, and it is far and away the most common answer to the whole question.

A wheel and tyre assembly is a heavy object spun at speed, and it is never perfectly even. The correction is a set of small weights fitted where the assembly is light — clipped to the rim flange, or stuck to the inside of the barrel where you cannot see them from outside. They are held on by a clip or by adhesive, and they come off. A pothole, a kerb, a jet wash aimed into the wheel, corrosion under the adhesive, or a tyre fitter who did not clean the rim properly are all it takes.

There are two kinds of imbalance and they feel different. A heavy spot in one plane makes the wheel hop vertically, which arrives as a vertical shudder. A heavy spot offset across the width of the wheel makes the assembly try to steer itself from side to side as it turns, and that one feeds straight into the steering as a side-to-side shimmy. The second kind is the classic shaking steering wheel at speed, and it is the reason a wheel has to be balanced dynamically rather than just hung on a spindle to see which way up it settles.

The buyer-relevant part is how cheap this fault is and how expensive its lookalikes are. Balancing four wheels is one of the smallest bills in motoring. A vibration that behaves exactly like imbalance but survives a proper rebalance is telling you the problem is the shape of something rather than its weight distribution, and that leads to a bent rim, a tyre that is not truly round, or a hub that is not running true. So the useful sequence is to have the wheels balanced first, precisely because it eliminates the cheap explanation quickly.

Two field observations worth making before you drive anywhere. Look along the inner rim flange for a bright, clean rectangle of bare metal with a faint adhesive edge — that is where a weight used to be. And note whether the weights on all four wheels appear to have been fitted at the same time and by the same hand, because a single wheel carrying a cluster of weights when the other three carry one each is a wheel that has been fighting to balance.

Rule out the loose wheel first, every time

A wheel that is not properly clamped to its hub produces a shake that grows with speed and gets dramatically worse under braking, and it is the one cause on this page that can turn into a crash within minutes rather than months. If a vibration appears shortly after any work that involved removing a wheel — a tyre change, a puncture repair, a brake job, a rotation — stop somewhere safe and check that every nut or bolt is present and tight before you go any further. This costs nothing and takes a minute. Everything else on this page can wait until you are somewhere with a lift.

When the tyre itself is the problem

Balance is about weight distribution. The other tyre faults that cause vibration are about shape and structure, and they cannot be balanced away.

Out of round. A tyre can be perfectly balanced and still not be a true circle. The result is a rise and fall in the force the tyre transmits once per revolution — the same rhythm as imbalance, so it feels similar, but it does not respond to weights. Some tyres leave the factory this way within tolerance and become noticeable only on a car sensitive to it; others develop it after damage. A tyre shop with a machine that measures road force rather than just weight distribution can separate the two, and it is worth asking for by name when a car has been balanced twice and still shakes.

A separated belt. Inside the tread there is a package of steel belts bonded to the rubber. When that bond fails, usually after an impact or a badly executed puncture repair, a section of the tread starts to move independently of the carcass. This one has a distinctive feel: a rhythmic wobble or thump that is present at low speed as well as high, often with a visible bulge or ripple in the tread or the sidewall when you look along the tyre. It is the one tyre fault on this page that should stop you driving the car.

Irregular wear. Feathering from a toe misalignment, scalloping from tired dampers, or a shoulder worn away by a camber problem all make the contact patch uneven, and the result is a coarse roughness through the car that most people read as bad road surface. It rarely produces a clean speed band. What it produces is a car that never quite feels smooth, and a set of tyres being destroyed while the owner assumes the noise is the road. Reading a wear pattern properly, and dating a tyre from the code moulded into its sidewall, are the other half of this conversation; both are covered in our page on how long tyres last, and neither is repeated here.

Flat spots. A car that has stood for weeks, particularly in cold weather, can take a set in the contact patch. That produces a vibration for the first few miles which then eases as the tyres warm and the rubber recovers. Genuinely temporary, genuinely harmless, and a common false alarm on the first drive of a car that has been sitting on a forecourt.

Mismatched tyres deserve a note of their own here. Four different brands, or two new tyres on one axle and two worn ones on the other, will not by itself cause a vibration, but it changes how the car responds to any of the faults above, and it is a small statement about how the previous owner made decisions.

The judder that only exists under braking

Nothing at a steady speed, nothing under acceleration, and then a rhythmic pulse through the pedal and often the steering wheel as you slow, fading away as the car does. This is the easiest condition on the list to identify, because the trigger is unambiguous.

The mechanism is a variation in how hard the brake grips as the disc turns. Two things produce it. The first is a disc that is not the same thickness all the way round, so the pads are squeezed apart and released once per revolution. The second is friction material transferred unevenly onto the disc face, which leaves patches with different grip and produces the same once-per-revolution pulse without the disc having changed shape at all. Both get described as a warped disc. Neither usually involves a disc bent like a record left in the sun, which is what the word suggests and what almost never happens.

The practical difference matters. Deposit-driven judder can sometimes be reduced by a period of firm, progressive braking that scrubs the disc face clean, and it can also be caused in the first place by holding the car stationary on a hot brake — sitting at a long light with your foot down after a hard stop bakes pad material onto one spot. Thickness variation does not improve on its own and does not respond to new pads. A car that judders now and juddered when the seller had the pads changed is a car that needs discs.

Where you feel it points at the axle in the usual way. Judder through the steering wheel is the front discs; judder felt mainly as a pedal pulse and a shudder through the seat is the rear. Rear judder is under-reported, because most people assume brake judder is always a front-wheel event.

Two things that are not faults. A rapid buzzing or grinding pulse under your foot during a genuine emergency stop or on a slippery surface is anti-lock braking doing its job, and it should be there. A single scrape or a light shudder on the first stop after the car has stood overnight in the wet is surface rust being wiped off the discs, and it disappears in a few applications. Pad and disc wear, what can be seen through a wheel without removing it, and the gap in cost between fitting pads alone and fitting both are dealt with separately in our page on how long brake pads last.

Shaking at a standstill that vanishes as you pull away

The car shudders at a red light, the mirror blurs, a cup of coffee ripples, and the moment you are moving it is gone. Nothing here is about wheels.

Two families cover almost all of it. The first is what holds the engine. An engine is a machine that shakes as a matter of design, and the mounts between it and the body exist to absorb that. Many modern mounts are fluid-filled, which lets them be soft enough to isolate an idle and stiff enough to control the engine under load. When one collapses or leaks, its isolation goes and the idle arrives in the cabin.

The second is a misfire, where one cylinder is contributing less than the others and the engine is running unevenly rather than being poorly isolated. A misfire usually brings company: an engine management light, sometimes flashing, a change in the exhaust note, and a shake that is present under load as well as at rest even if it is loudest when idling.

Three checks separate them, and none needs a tool. Put an automatic in drive with your foot firmly on the brake and see whether the shake gets markedly worse — loading the mounts against the transmission is exactly what exposes a collapsed one. Watch the engine as somebody starts and stops it: a large lurch as it shuts down is a mount finding to note. And look at the dashboard, because a misfire that is bad enough to feel through the seat is usually bad enough to have set a code, whether or not the lamp is currently lit.

Two ordinary explanations deserve to be ruled in rather than out. An air conditioning compressor cycling on and off loads the engine and produces a regular pulse at idle that is entirely normal. And engines with an odd number of cylinders, or with cylinder deactivation systems that shut half the engine down at light load, are simply rougher at idle than a smooth six ever was. Neither is a fault, and both are worth knowing before you use idle quality to argue about price.

Only under power

Smooth on a trailing throttle, smooth while braking, and a shudder or vibration that appears the moment you ask for acceleration. Load is doing something that speed alone does not, and the candidates are the parts that only carry torque when torque is being made.

Inner joints on a front-wheel-drive car. Each front driveshaft has a joint at each end that lets it change length and angle as the suspension moves. The inner joint is the one that plunges, and when it wears it produces a shudder under load — classically when accelerating from low speed, and worse on a slight turn or up a slope. Outer joints tend to announce themselves differently, with a rhythmic clicking on full lock, so a car that clicks on lock and shudders under power may well be telling you about both ends of the same shaft.

The driveshaft on a rear-wheel-drive or four-wheel-drive vehicle. The shaft that runs the length of the car under the floor turns fast, has to be balanced like a wheel, carries universal joints at its ends, and on many vehicles hangs from a rubber-mounted support bearing part way along. An imbalance in the shaft, a dry or worn joint, or a perished centre support each produce a vibration that builds under power and eases the moment you lift off. It arrives through the floor and the seat rather than the steering wheel, which is a helpful signature in itself, and it comes up more often on pickups and larger SUVs for the plain reason that a longer shaft is harder to keep true.

Mounts again, in their other role. Worn engine and gearbox mounts let the whole powertrain rock when torque is applied and released. That shows up as a shudder on take-up, a clunk when you lift off or change gear, and a general sense that the car is loose in a way that is hard to point at. On a front-wheel-drive car it can also change the driveshaft angles enough to make the joints complain, which is why a joint that has been replaced and is still shuddering sometimes turns out to be a mount problem wearing a driveshaft disguise.

Transmission shudder. A light, cyclic shudder at a steady light throttle, often at cruising speed and often mistaken for a coarse road surface, can come from the lock-up clutch inside an automatic torque converter chattering rather than gripping cleanly. It is worth naming because it is the one item in this section where the diagnosis leads somewhere expensive, and because a fluid condition and a scan tool will get you a long way towards confirming or dismissing it before anybody quotes for anything.

Constant, rising with speed, and audible

A wheel bearing rarely starts as a vibration. It starts as a noise: a drone, a hum or a growl that comes in at moderate speed, rises in pitch with road speed, and is there whether you are on the throttle or coasting. Vibration comes later, once there is enough play in the bearing for the wheel to move on its axis.

The test costs nothing and takes seconds on an empty road. Sway gently from one side of your lane to the other. Weight transfer loads one side of the car and unloads the other, and a worn bearing usually gets louder when it is being loaded — so a noise that grows in a gentle left-hand curve and quietens in a right-hand one is pointing at the right-hand side of the car. Repeat it a few times before you trust it.

Two further details are useful. Bearing noise is speed-related and not gear-related, so a drone that changes when you change gear at the same road speed is not a bearing. And on many modern cars the wheel speed sensor for the anti-lock system reads a magnetic ring built into the bearing itself, so a failing bearing can produce an anti-lock or stability control warning with no braking fault anywhere. If a car shows a stability light and a drone from one corner, those are probably the same finding rather than two.

For a buyer, a bearing is a bounded, well-understood job on most cars, and it is one of the few faults on this page that gets worse in a predictable direction. What matters is catching it, because bearing noise is easy to miss with the windows up and the radio on — which is exactly how most test drives are conducted.

It started after a pothole

Sudden onset is the most informative history a vibration can have, and it is the one a seller is most likely to know and least likely to volunteer.

An impact hard enough to be memorable can do several things at once. It can knock a balance weight off, which is trivial. It can bend a rim, usually on the inner flange where you cannot see it with the wheel on the car — alloy wheels bend and steel wheels dent, and either will produce a vibration that no amount of balancing will remove. It can damage the tyre internally without leaving a mark you would notice, which is one of the routes to a separated belt. It can knock the alignment out, which shows up as a pull and as tyre wear rather than as a vibration on its own. And on a hard enough hit it can bend a steering or suspension arm, which changes geometry in a way no alignment adjustment can bring back.

The tell is the sequence. A vibration that was not there on Monday and was there on Tuesday has a cause with a date, and that date is worth asking about directly. On a car you are viewing, phrase it as a plain question rather than an accusation: has it been kerbed, has it hit anything, when was a wheel last taken off it.

The reason to take that seriously beyond the cost of a wheel is that impacts that reach the suspension often have a history behind them that reaches further. A kerb strike is one thing. An unreported collision is another, and it can leave a car that tracks straight and still vibrates because something in the structure is no longer where it was. That is a question the paperwork answers better than the driveway does, and it is worth taking two minutes to see what the VIN record says about accident and damage history before you spend money diagnosing the symptom.

The vibration that went away

This is the part of the subject most likely to cost somebody money, because a fault that stops announcing itself feels like a fault that has been resolved, and often it is the opposite.

A small number of vibrations genuinely do end on their own, and they are worth knowing so you can recognise them:

  • Flat spots in cold tyres, which disappear within a few miles as the tyres warm up and stay gone until the car stands again.
  • Ice, snow or mud packed inside the wheel rim, which unbalances the assembly beautifully and then falls out or melts.
  • A stone wedged in the tread, or one lodged between a disc and its dust shield, both of which can free themselves.
  • Deposit-driven brake judder that has been scrubbed off the disc face by a period of firmer braking.
  • Surface rust on the discs after standing, which is gone before you reach the end of the road.

Every one of those has an identifiable thing that left the car. If you cannot name what left, the more likely explanations are less comfortable. A wheel weight that was marginal has finally departed and the imbalance has moved into a speed band you rarely drive in. A bearing that was rumbling has worn to the point where the noise has changed character rather than stopped. A mount that was allowing movement has settled into a new position. A tyre with a structural fault has redistributed the damage. In each case the underlying fault is still there, still progressing, and now less well signposted than it was.

There is a version of this that matters specifically to buyers. A car that was juddering last week and is smooth today may have had the brakes bedded in on the way to meet you. A car that shook at speed and now does not may have been round a tyre shop the day before the viewing. Neither is dishonest by itself and both change what the test drive can show you, which is a good reason to ask what work has been done recently and to look at whether the answer matches what you can see on the car.

What the federal records show, and what they cannot

Two archives on this site touch this subject from different angles, and they support quite different kinds of claim. It is worth being precise about both, because the temptation to turn either into a prediction about a specific car is strong and neither will bear it.

Owner complaints, and the mileage on the odometer at the time

The first is our archive of complaints filed by owners with the National Highway Traffic Safety Administration. Only a minority of them state an odometer reading in the intake text, and only those are usable here: 1,324 readings out of 9,096 complaints read, a recovery rate of 14.6%. They span 13 models and 15 component categories, and the vehicles behind them come from three model years — 2015, 2016 and 2018. Broken out by category, the ones that bear on vibration look like this.

Median odometer reading at the time an owner filed a complaint, by component category
Component categoryReports with a mileageMedian odometer
Power train12449,000 miles
Suspension2753,000 miles
Service brakes, hydraulic38760,000 miles
Steering8464,000 miles
All categories in the file1,32460,000 miles

What these numbers are not

Each figure marks where an owner’s odometer stood on the day they decided a problem was worth reporting to a federal regulator. That is not a failure rate, not a service interval, and not a forecast for any particular vehicle. A complaint is a voluntary account of one car by one person, and nothing in the intake process verifies it. Censoring sits on top of that: these vehicles have only been watched up to now, so nobody in the file has had the opportunity to report anything at a mileage they have not yet reached, and the sparse readings at the top of the range measure the age of the sample rather than the durability of anything in it. Nothing about that far end can be relied upon. Comparison between the categories is what survives, since all of them are drawn from one pool of vehicles over one span of years.

Read carefully, the ordering says something a buyer can act on. The four categories that produce vibrations sit in the middle of a used car’s life rather than at the end of it, with only steering landing above the median for the file as a whole and then not by much. That is not a claim that your suspension will need attention at a particular reading. It is a reason to treat a car in that mileage territory as squarely inside the window where owners of these models were already reporting steering, suspension, brake and driveline problems, and to test-drive it accordingly rather than assuming that a car with plenty of life left in it has yet to develop any of these faults. The broader version of that argument, built on the same file, is the subject of its own page: how many miles is too many.

Recalls, where the manufacturer conceded the defect

The second archive is different in kind. A recall is a defect the manufacturer has admitted and agreed to fix at its own expense, so it carries a weight no complaint does. Our pull covers the models this site holds records on: 158 rows returned, 13 duplicates removed where a campaign covered several model years, leaving 145 distinct campaigns across 24 component families.

Recall campaigns in the archive, families relevant to vibration
Component familyDistinct campaignsManufacturers affected
Service brakes, hydraulic126
Steering94
Suspension43
Tires11

This is not a census of US recalls. Every query behind it was aimed at one specific vehicle from the set this site covers, so what the counts really measure is the shape of that query list and not the shape of the American recall record. Nothing here identifies the most recalled component, manufacturer or model, and nothing here is meant to. The modest claim it does support is that certain systems keep reappearing among manufacturers that have nothing else in common.

The detail inside those families is more interesting than the counts. The four suspension campaigns name, between them, a lower control arm, an upper ball joint, a front wheel bearing and a rear suspension item — which is close to a list of the parts this page has been describing. Three of the nine steering campaigns concern the electric power assist system rather than anything mechanical. And two of the power train campaigns name the driveshaft specifically, which is the part behind the under-power vibration described above.

The practical point for a buyer is procedural rather than statistical. Outstanding recall work is performed at no cost by a franchised dealer, and neither the current owner, nor the way the car changed hands, nor the presence or absence of a warranty alters that. The search takes a couple of minutes and needs nothing but the VIN. Should a car you are considering have both a vibration and an unfinished campaign against a part that could plausibly produce one, you have found the cheapest repair available anywhere. How to run the search, and what a clean result does and does not prove, is set out in our guide to running a recall check on a VIN.

Provoking it deliberately

A test drive that stays in traffic at low speed will not find any of this. The point of the drive is to visit each of the conditions on purpose, once, and to note what happened before moving to the next one.

Before you turn a wheel

  • Look at all four wheels from the outside and then from behind. Kerb rash on the rim edge tells you the car has been parked hard. Flat sections or ripples on the inner flange are what you are really looking for, and a torch helps.
  • Find the balance weights. Note how many each wheel carries and look for the clean patch where one has come off.
  • Run a hand around each tyre where you can reach it, feeling the sidewall for a bulge and the tread for a step or a ripple. Do the inner shoulders too, because that is where the evidence usually is and where nobody looks.
  • Check that the tyres match across each axle, and note anything obviously newer than the rest.
  • Start the engine and watch it settle. Stand at the front, look at the engine rather than listening to it, and see how much it moves as it starts and as it is switched off.

On the drive

  • Idle, in neutral and in gear. On an automatic, hold the brake firmly and select drive, then reverse, noting whether the shake changes.
  • Accelerate firmly once in a straight line, from low speed through the mid range, hands relaxed on the wheel.
  • Hold a steady cruise and sweep slowly through the speed range, giving each part of it a few seconds. This is where a balance band appears and it is easy to drive straight past.
  • Lift off and coast through the same speeds, then repeat with the lightest possible throttle. The difference between those two is the load-dependence answer.
  • Brake progressively from a moderate speed twice, somewhere with nothing behind you, once gently and once firmly. Note whether the pulse is in the pedal, the wheel, or both.
  • Sway gently within your lane on an empty road to load each side in turn, listening rather than feeling.
  • Turn the radio and the fan off for the whole drive. Half of what is described on this page is audible before it is palpable.

Do all of that and you will have answers to the three questions on a car you have driven for a quarter of an hour, which is more than most owners can supply about a car they have had for years. It is also the raw material for a sensible conversation with the seller and, if it comes to it, with a workshop.

What a vibration should do to the deal

Finding a vibration is not, on its own, grounds for walking away from a car. It is grounds for establishing which of two very different situations you have in front of you.

The first is a maintenance item with a known price and a defined end: wheel balancing, a tyre, a bearing, a pair of discs. Those are ordinary costs on a used car, they are quotable in advance, and the correct response is to price them into the offer rather than to abandon the purchase.

The second is a symptom of something structural or historical: a bent wheel that came from an impact that also reached a control arm, a driveline vibration on a vehicle that has been worked hard, a shake with no identifiable source on a car whose paperwork does not match its condition. Those are unbounded, and the reason to care about the distinction is that the symptom is identical from the driver’s seat.

What separates them is an inspection by somebody who can put the car on a lift, and this is one of the clearest cases for it on the whole site, because most of what causes a vibration is invisible with the wheels on the ground. A proper pre-purchase inspection will check wheel and tyre runout, bearing play, joint and bush condition, disc thickness and mounts — and if you name the symptom when you book it, a general look becomes a specific question with a specific answer. Take the notes you made on the drive with you.

Two habits are worth carrying into the negotiation. Ask the seller to reproduce it, because a vibration is one of the few faults that can be demonstrated on demand, and watch how they respond to being asked. And keep the mechanical picture and the documentary picture separate in your head: a car can be mechanically sound and still be the wrong car, which is why it makes sense to check the history record alongside the test drive rather than treating the drive as the whole assessment. And once the car is yours, tyre pressures, rotation and an alignment check when the wear calls for one remain the cheapest way to keep most of this page from ever applying to it.

Two of those causes are worth their own guides. Worn dampers produce float and wallow rather than a vibration proper, and a misfire is what you are usually feeling when the shake is there at idle and gone once moving.

Common questions

Why does my steering wheel shake at higher speeds?

A shake that arrives in a speed band during a steady cruise, with no braking and no particular throttle demand, is wheel balance at the front of the car until proven otherwise. The steering wheel is the giveaway: the front wheels have a short mechanical path into the steering column, so a front imbalance reaches your hands directly while a rear one arrives through the seat. Have the wheels balanced first, because it is the cheapest explanation and eliminating it quickly is what makes the remaining possibilities worth investigating.

Why does my car shake when I brake?

A pulse that appears only under braking and fades as the car slows is coming from the discs. Either the disc is not the same thickness all the way round, or friction material has transferred onto its face unevenly, and both make the brake grip harder and softer once per wheel revolution. Judder through the steering wheel points at the front discs, judder felt mainly in the pedal and the seat points at the rear. New pads on their own will not fix it.

Why does my car shake at idle but drive smoothly?

Nothing about that pattern involves the wheels. Either the engine mounts are no longer isolating a normal amount of engine vibration from the body, or one cylinder is contributing less than the others and the engine is genuinely running unevenly. Putting an automatic in drive with your foot on the brake usually makes a mount problem markedly worse, while a misfire tends to bring an engine management light and a change in the exhaust note with it.

Is it safe to drive a car that vibrates?

It depends entirely on what is causing it, and two of the causes on this page are urgent. A wheel that is not properly clamped to its hub and a tyre with a separated belt can both fail suddenly, and neither should be driven on. Anything that appeared immediately after an impact deserves the same caution until somebody has looked at the wheel and the tyre. A long-standing balance vibration or a slowly worsening bearing drone is a different order of risk, but the honest position is that a new vibration you cannot explain should be inspected before it is ignored.

Will a wheel alignment fix a vibration?

Usually not, and this is one of the most common misunderstandings on the subject. Alignment sets the angles at which the wheels point; balance corrects the weight distribution of a wheel and tyre spinning at speed. They are separate operations that treat separate faults. An alignment problem shows up as a car that pulls, a steering wheel that sits off-centre, and tyres wearing unevenly — and that uneven wear can eventually cause a roughness, which is the indirect route by which alignment gets blamed for vibration.

Should I buy a used car that vibrates on the test drive?

Only once you know which kind of vibration it is. Balancing, a tyre, a bearing or a pair of discs are bounded, quotable costs that belong in the negotiation rather than in the walk-away decision. A vibration nobody can account for, or one that arrived with an impact, is a different proposition, because the visible fault may be the smallest part of what that impact reached. Get it on a lift, name the symptom when you book the inspection, and make the decision on the report rather than on the drive.

My car was shaking and now it has stopped. Is that good news?

Only if you can name what left the car. Cold flat spots, packed snow inside a rim, a stone in the tread, surface rust on the discs and deposit-driven judder all genuinely resolve, and all of them have an obvious explanation attached. Where there is no explanation, the more likely story is that the fault has moved on: a weight has finally come off and shifted the imbalance into a speed band you rarely use, or a bearing has worn past the point where it announces itself the same way. Silence is not repair.

Sources and further reading

Recall, complaint and safety-rating figures on this page were retrieved from the federal databases above on August 19, 2026. Federal data changes — re-check any VIN before you rely on it.

Baron Auto Editorial Team We research used cars against federal data — NHTSA recall campaigns, owner complaints and EPA fuel-economy records — and publish what we find. We do not sell cars, loans, or insurance, and no manufacturer or dealer pays for coverage here.

Last updated August 28, 2026. Found something out of date or wrong? Tell us and we will correct it.