What a bolt torque setting really achieves TORQUE
About a tenth of the torque becomes tension. The rest is friction, and the same setting through a dry thread and a waxed one differs by two thirds.
The bolt
The friction, which is the whole calculation
The nut factor is a typical value rather than a specification. It is the single biggest uncertainty here, and a spread of a quarter in it is a spread of a quarter in what is holding the joint together.
What the setting achieves
The same setting, six conditions
The working
How it works
A torque figure is not a specification of anything the bolt cares about
The bolt cares about the tension in it: the clamp force holding the joint together. Torque is only a way of getting there, and it is an indirect and lossy one. The relationship is
T = K × D × F
where D is the nominal diameter, F is the clamp force you actually want, and K is a nut factor standing in for every scrap of friction in the thread and under the head.
Here is the part that should change how you think about a torque wrench: roughly ninety per cent of the torque you apply goes into friction, and about ten per cent becomes tension. You are not measuring the bolt. You are measuring the friction, and inferring the bolt from it.
Which makes K the whole calculation
K is typically about 0.20 dry and about 0.12 waxed. Put the same torque through those two and the clamp force differs by roughly two thirds, from the same wrench, at the same setting, on the same bolt.
So “torque it to 80 Nm” is only a complete instruction if the condition of the thread is part of it. Dry, lightly oiled, anti-seize, zinc-plated, waxed and re-used are all different numbers, and anti-seize in particular is the one that catches people: it is a lubricant, it drops K sharply, and a bolt torqued to the dry figure with anti-seize on it can be taken past yield without anything feeling wrong.
This shows the clamp force each condition produces from the torque you enter, side by side, which is the comparison that makes the point. If the spread between them matters for your joint (and on anything structural it does), then the specification has to name the condition, and the person doing the work has to follow it.
What to do when it matters
Follow the manufacturer's figure and their stated condition, not a table from somewhere else. A published torque is only meaningful with the assumptions it was derived under, and those assumptions are theirs.
Use a calibrated wrench and know when it was last checked. A clicker that has been dropped, or left wound up for a year, reads what it reads and gives no indication that it is wrong.
Treat re-used bolts with suspicion. The friction has changed, the surface has changed, and on a critical joint a bolt that has been to yield once is a bolt to replace rather than retighten.
Where the joint really matters, torque is the wrong instrument and everyone in the trade knows it: angle tightening, or measuring bolt stretch, or a tensioner, all measure something much closer to the thing you actually want. Torque persists because it is quick, and knowing that it is a proxy with a 3:1 spread built into it is the difference between using it sensibly and trusting it.
Nothing you type here is uploaded; the arithmetic runs in the tab. Nearby: SLING works out what each leg of a lift is really carrying, AXLE and BALANCE check a loaded vehicle, and FIGURES handles the rounding so a tolerance does not quietly disappear.
The guide: Why a four-leg sling is rated on three legs
Questions people ask about TORQUE
Why is the stress area a formula rather than a table?
Because it is one: a quarter pi times the square of the nominal diameter less 0.9382 pitches, which is the mean of the pitch and minor diameters. It gives 84.3 mm² for an M12 and 157 for an M16, which is what the table says, and it also works for a thread the table does not have.
What is the nut factor and why does it matter so much?
It is the whole of the friction in the thread and under the head, rolled into one number. Roughly ninety per cent of the torque applied to a bolt is spent overcoming it and about ten per cent becomes tension, so the figure on the wrench is mostly a statement about friction. It is typically 0.20 dry and 0.12 waxed, and the same torque through those two differs by two thirds in what is actually clamping the joint.
So is a torque wrench useless?
No, but it is blunt. A properly specified torque with a controlled thread condition gets preload within about a quarter, which is fine for most joints and hopeless for some. Where it matters the answer is tightening by angle past snug, or measuring the bolt’s stretch, both of which look at the bolt rather than at the friction.
Why does the page show a range rather than a number?
Because a single number implies a precision the method does not have. The range is what a realistic spread of friction produces from one setting, and seeing that a dry figure on a lubricated thread goes past the proof load is the thing worth knowing.
Where do the proof stresses come from?
From ISO 898-1, and they are editable on the page. They depend on the property class and, for 8.8, on the diameter. An invented figure is a bolt that snaps or a joint that opens, so check them against your own copy.
Doing this often? The OBSCURA Engineering Suite designs the parts the bolts hold together, with sections, interference checks and mass properties. On its own, US$2 a month; with everything else, US$5 a month or US$48 a year.