The short version
- The axle nut sets bearing preload. It is not a normal fastener and there is no safe guess.
- Published torque tables disagree with each other and with themselves. We checked two editions of one and found the errors.
- Use your vehicle’s own service data. Treat any chart, including ours, as a sanity check rather than a source.
- Torque with the wheel on the ground, with a calibrated wrench. Never an impact gun.
Most fasteners on a car just need to stay done up. Wheel bearing torque specs are different. On a driven corner the axle nut sets how tightly the bearing’s internal components are held, which makes the number a functional specification rather than a tightness.
Get it wrong in either direction and the bearing fails in weeks. So the number matters, and this is where the subject gets awkward.
Why wheel bearing torque specs charts get it wrong
Search for wheel bearing torque specs and you will find charts. They are convenient, they look authoritative, and a meaningful share of them are wrong.
We can put a number on that, because torque guides get republished and the editions can be compared.
We took two editions of the same manufacturer’s axle nut torque catalog, four years apart, and extracted every specification from both. Then we checked each figure against itself.
Every entry lists the same spec twice, once in pound feet and once in newton meters. Those two numbers have a fixed relationship: one pound foot is 1.3558 newton meters. If a pair does not convert correctly, at least one of the two figures is wrong.
What we found
| Edition | Pairs checked | Internally inconsistent |
|---|---|---|
| Earlier edition | 1,336 | 70, or 5.2 percent |
| Later edition, four years on | 1,799 | 142, or 7.9 percent |
The error rate went up, not down. Some individual entries are badly wrong rather than marginally.
One listing gives a minivan as 118 pound feet and 244 newton meters. But 118 pound feet is 160 newton meters. Follow the metric column and you are at roughly 180 pound feet, half as much again as the same line says in imperial.
Another lists a small hatchback as 144 and 106. Those two numbers are the correct conversion of each other in the wrong order, so the columns have simply been swapped.
If you own a torque wrench that reads in one unit and you copy the figure from the other column, an error like that is invisible. Everything about the process feels correct. The bearing is destroyed anyway.
The specs that genuinely changed
Separately from the arithmetic errors, we looked for entries where the manufacturer revised the figure between editions and both versions were internally consistent. Those are real corrections rather than typos.
There were 39, and several are dramatic.
| Vehicle | Earlier edition | Later edition | Change |
|---|---|---|---|
| Toyota Cressida, 1989 to 1992 | 203 lb-ft | 108 lb-ft | Nearly halved |
| Lexus GS300, 1993 to 1997 | 213 lb-ft | 147 lb-ft | Down 31 percent |
| Infiniti FX35 and FX50 RWD | 133 lb-ft | 92 lb-ft | Down 31 percent |
| Honda S2000, 2000 to 2009 | 181 lb-ft | 242 lb-ft | Up 34 percent |
| Toyota Supra, 1987 to 1992 | 203 lb-ft | 147 lb-ft | Down 28 percent |
Read the first row again. Somebody working from the earlier edition would tighten a Cressida axle nut to nearly double what the same publisher later said was correct.
None of this means the publisher is careless. Compiling several thousand specifications across four decades of vehicles is genuinely hard, and revising them is the system working. It does mean that a chart is a secondary source and should be treated like one.
So where do you get a number you can trust?
In this order.
- The vehicle manufacturer’s own service information. A factory service manual or a subscription service. This is the only primary source, and it is the one that carries the tightening sequence and any angle stage too.
- The instructions in the box. Bearing manufacturers often include a spec sheet with the part, and it is written for that exact application.
- A published chart, cross checked. If you must use one, verify the pair converts correctly and compare against a second independent chart before you rely on it.
- Never a forum post or a figure from a different model year. Torque specs change within a model run more often than people expect.
Multiply the pound foot figure by 1.36. If the answer is not close to the newton meter figure printed next to it, that entry is wrong and you should not use either number.
What the range actually looks like
Useful for sanity checking rather than for setting your wrench, because the spread is enormous.
| Vehicle class | Typical axle nut range |
|---|---|
| Compact and midsize cars | Roughly 130 to 185 lb-ft |
| Larger sedans and crossovers | Roughly 150 to 240 lb-ft |
| Full size trucks and SUVs | Frequently 165 to 310 lb-ft |
| Some European performance models | Over 300 lb-ft |
That is a range of more than four to one across the fleet. It is also why wheel bearing torque specs borrowed from a similar looking vehicle are worthless.
Doing it properly
Three things matter as much as the number itself.
Wheel on the ground. Torque the axle nut with the vehicle’s weight on the wheel. Doing it in the air can turn the hub, and leaning on a long bar with the car on stands is how cars come off stands.
A calibrated click wrench, not an impact gun. An impact wrench cannot hit a figure, and it routinely overshoots, which crushes the rolling elements and cooks the grease. Use it to remove the old nut if you like, never to fit the new one.
Clean, dry threads unless told otherwise. Most axle nut specs assume dry threads. Oiling them changes the clamp load a given torque produces, sometimes by a third, which puts you well past spec while the wrench reads correctly.
It does not just stress the nut. It crushes the races together, removing the internal clearance the bearing needs to run. That generates heat immediately, breaks the grease down, and produces a failure that looks exactly like a defective part.
Staked nuts and angle stages
How a staked axle nut comes off and goes back on has its own page.
Two details that charts often omit and manuals do not.
Many axle nuts are staked, meaning a collar is peened into a groove on the shaft after tightening. Those are single use. Reusing one leaves nothing locking it, and the stake material has already been worked once.
Some specs are a torque plus angle: tighten to a figure, then turn a further specified number of degrees. If your source gives only a single figure for a vehicle that uses an angle stage, your source is incomplete.
Hub mounting bolts, where fitted, have their own separate figure and sequence. They are not the same as the axle nut and should not be guessed from it.
If you have already torqued it wrong
Worth knowing, because the temptation is to hope.
If you over-torqued significantly, the bearing has probably been damaged and it will announce itself within weeks. Loosening and re-torquing does not undo crushed raceways.
If you under-torqued, catching it early may genuinely save the bearing. Play means impact loading, and a short period of that is survivable where sustained running is not.
Either way, the check is to test the corner properly and watch for the hum described in the noise progression. A new bearing that hums from the first drive was damaged during fitting.
Angle-tightened fasteners cannot go back in twice. Why a reused stretch bolt clamps less explains the mechanism.
Finding a socket that fits the nut is covered under axle nut sizes, and a specification written in degrees needs an angle gauge.
Common questions
What torque should a wheel bearing axle nut be?
Wheel bearing torque specs have no universal figure. It ranges from around 130 to over 300 pound feet depending on the vehicle, and it must come from your vehicle’s own service data rather than a general chart.
Can I just tighten it as hard as I can?
No, and this is the most damaging thing you can do. Over-torque crushes the bearing internally and causes failure within weeks.
Are online torque charts reliable?
Partly. In two editions of one professional catalog we found 5.2 and 7.9 percent of entries were internally inconsistent, meaning the imperial and metric figures did not convert to each other. Cross check any chart before using it.
How do I check a chart is right?
Multiply the pound foot number by 1.36 and compare it to the newton meter number beside it. If they do not match, at least one figure is wrong.
Do I torque with the wheel on or off the ground?
On the ground, with the vehicle’s weight on the wheel. It is safer and it stops the hub turning as you apply load.
Can I reuse a staked axle nut?
No. The stake is deformed metal that has already done its job once. New nuts are inexpensive and this is a genuine failure path, as the replacement overview covers.
Does the hub bolt torque matter as much?
It matters, but differently. Hub mounting bolts hold the unit to the knuckle rather than setting preload, so getting them wrong causes looseness rather than internal damage. They still have their own spec and sequence.