“Wheel bearing has slight play”

The characteristic sound is a hum or drone that rises and falls with road speed, not with engine speed. That distinction alone eliminates a great many other faults: anything that changes when you press the accelerator in the same gear is not a bearing.

The load test identifies the side. On a quiet road at a steady speed, swing the car gently left and right. Weight transfers to the outside wheels, so a bearing on the left gets quieter when you steer left and louder when you steer right — because turning right loads the left-hand wheels. Most people find this counterintuitive, and it is the single most useful diagnostic anyone can do without tools.

Front or rear is harder by ear, and this is where the ramp earns its money. A mechanic spins each wheel by hand and listens or feels for roughness, then rocks the wheel top-to-bottom to check for play. Both take seconds per corner.

The impostor worth ruling out is tyre noise. Uneven or feathered tread produces a drone that sounds mechanical and also rises with speed, and it is far cheaper to diagnose than to replace a bearing that was fine. If you also have a tyre advisory, check the tread pattern before agreeing to a bearing.

Legally, until the play becomes excessive. The car passed.

Practically, months — and the bearing tells you when. This is one of the few components that gives genuinely useful escalating warning. Slight play is silent or nearly so. Then a faint hum appears at motorway speed. Then the hum is audible around town. Then it becomes a growl you can hear with the windows up. That progression usually takes months, occasionally longer.

What you should not do is run it to the growl stage. A badly worn bearing generates heat, and heat damages the hub it sits in, which turns a bearing job into a hub-and-bearing job. In extreme neglect a bearing can seize, and a seized front bearing at speed is a genuine loss of control.

The sensible trigger is when the noise becomes noticeable in normal driving, not when it becomes annoying. There is enough margin in that to book the work at your convenience.

The test distinguishes slight play, an advisory, from excessive play, a major defect. A bearing that is rough or noisy enough to be judged defective also fails, as does one that is obviously seizing.

No measurement here either. The tester grips the wheel at the top and bottom and rocks it, watching for movement at the hub. Detectable movement that is small is slight; obvious movement, or a knock as the wheel rocks, is excessive. It is the same lever-and-judge approach used for suspension arms, with the same year-to-year variability between test stations.

One useful distinction a good tester makes: play at the top and bottom of the wheel can come from the bearing or from a suspension ball joint, and the two are separated by watching where the movement actually occurs while someone rocks the wheel. If your certificate lists both a bearing and a joint, ask whether they were separately confirmed — occasionally one fault gets recorded twice.

Roughness matters as much as play. A bearing can have almost no measurable movement and still be failing, because the surfaces have pitted. Spinning the wheel by hand and feeling for a gritty or notchy rotation catches that, and it is why the hands-on check is worth more than the rock test alone.

This page exists partly to say this, because it runs against everything else in the same area of the car.

Springs, dampers and brake pads are replaced in pairs because they affect how the car behaves side to side. A new damper beside a tired one makes the car handle asymmetrically. A new spring beside a settled one makes it sit unevenly. Those are real reasons.

A wheel bearing does none of that. It either spins freely or it does not. A new bearing on one side and an original on the other produces exactly no difference in handling, braking, tyre wear or anything else. The bearings do not interact, and they are not a matched pair in any meaningful sense.

So “we may as well do both while we are in there” is not a technical argument for bearings, and the labour is not shared either — each corner is a separate job requiring the same work again. If both bearings genuinely have play, do both. If one does, do one.

This is the sort of thing our symptom checker is built around: the difference between advice that has a reason behind it and advice that is simply the default.

£120–£300 per corner fitted at an independent garage. The spread comes almost entirely from how the bearing is mounted: many modern cars use a complete hub unit that unbolts and replaces as one piece, while others require the old bearing to be pressed out of the hub carrier and the new one pressed in, which needs a press and more time.

Rear bearings are often cheaper than fronts on front-wheel-drive cars, because there is no driveshaft to disturb. On rear-wheel-drive and four-wheel-drive cars that advantage disappears.

ABS sensors are worth a question. On many cars the sensor ring is integrated into the bearing, and on some the sensor itself sits close enough to be disturbed. Replacing a bearing and triggering an ABS warning light afterwards is a known annoyance, usually caused by a damaged sensor or a misaligned ring — ask whether the sensor comes out and goes back.

What should not be on the quote: the other side, unless it also had play. And an alignment, because a bearing does not affect wheel geometry — unlike a track rod end, where alignment is mandatory.