Skip to content
K Knidox Search…
Engineering · Fasteners

Bolt Torque

Tightening torque and clamp force for metric bolts by property class.

Bolt sizeCoarse metric thread.
Property class
Thread conditionThe nut factor K depends almost entirely on friction.
% of proof
65–75% is the usual range for a reusable joint.
Tightening torque
52.2 N·m

To reach 26.1 kN of clamp force in a M10 class 8.8 bolt.

Preload · bolt stress
26.1 kN · 450 MPa

75% of the 34.8 kN proof load. Preload is the whole point of tightening — a bolted joint carries load through friction between the clamped faces, not through the bolt in shear.

The nut factor K is an empirical fudge covering thread and face friction, and it varies by ±25% even between supposedly identical bolts. Torque control therefore delivers preload only to about that accuracy — which is why critical joints use angle control, bolt elongation, or load-indicating washers instead. Never reuse a bolt tightened past yield.

Torque is only a means to an end — the goal is preload. T = K·D·F, so an M10 class 8.8 bolt tightened dry to 70% of proof needs about 49 N·m to reach 24 kN of clamp force.

A bolt is a spring, not a pin

The common mental model — that a bolt holds two parts together by resisting shear across its shank — is wrong for almost every properly designed joint. Tightening stretches the bolt elastically, and that stretch creates clamp force squeezing the joint faces together. The load is then carried by friction between those faces, with the bolt acting as a very stiff spring holding them shut.

This explains why preload matters so much. An under-tightened joint lets the faces slip and the bolt starts seeing shear and bending it was never designed for, which is a classic route to fatigue failure. A properly preloaded bolt in a static joint barely sees the external load change at all.

Torque is a poor proxy for preload

The awkward truth about torque wrenches is that most of the torque never becomes preload. Roughly 90% is spent overcoming friction — about half under the bolt head or nut face, and about half in the threads. Only the remaining tenth actually stretches the bolt.

That makes the result extremely sensitive to friction. The nut factor K is an empirical lump covering all of it, and it varies by ±25% even between nominally identical bolts from the same box. Torque control therefore delivers preload to only about that accuracy, which is fine for general assembly and inadequate for critical joints.

T = K × D × F Proof load = proof stress × tensile stress area Target F = 65–75% of proof load

D is the nominal diameter in metres and F the target preload in newtons. K is the nut factor: about 0.20 dry, 0.15 lightly oiled, 0.12 waxed, and 0.28 for dry zinc plating.

Worked example: M10 class 8.8, dry

Work from the bolt's capacity down to the torque setting:

  1. 1
    Find the tensile stress area. M10 coarse thread has 58.0 mm² — less than the 78.5 mm² of the plain shank, because the threads reduce the section.
  2. 2
    Get the proof load. Class 8.8 has a proof stress of 600 MPa, so 600 × 58.0 = 34.8 kN.
  3. 3
    Choose a preload target. At 70% of proof that is 24.4 kN — inside the elastic range, so the bolt is reusable.
  4. 4
    Pick the nut factor. Dry, as-received threads give K ≈ 0.20.
  5. 5
    Apply T = K·D·F. 0.20 × 0.010 m × 24,360 N ≈ 48.7 N·m.
  6. 6
    Note the sensitivity. Lubricating to K = 0.15 would need only 36.5 N·m for the same clamp force — using the dry figure on oiled threads over-tightens by a third.

Property classes

The class marking encodes the material: the first number is tensile strength ÷ 100 MPa, and the two together give yield as a fraction of tensile.

ClassTensile strengthProof stressTypical use
4.8400 MPa310 MPaGeneral low-duty fixings
8.8800 MPa600 MPaThe structural and automotive default
10.91000 MPa830 MPaHighly loaded joints
12.91200 MPa970 MPaSocket-head cap screws, high performance

Why published tables disagree with each other

Torque tables from different sources rarely match, and the reason is that they bake in different assumptions. One may assume 65% of proof and another 75%; one may assume dry threads and another lightly oiled. Both are internally consistent and they can differ by 30% or more. This calculator exposes those two choices rather than hiding them, which is why its default sits slightly above some published figures.

For joints where preload genuinely matters, torque is not the tool. Angle control — snug, then turn a specified number of degrees — largely bypasses friction variation because it measures stretch geometrically. Bolt elongation measurement and load-indicating washers are more accurate still. Production engine and structural work uses these routinely.

Two safety notes. A bolt tightened past yield has permanently stretched and must not be reused; many automotive fasteners are deliberately torque-to-yield and are single-use by design. And lubricating a joint whose torque figure assumes dry threads is a genuine hazard — the same torque produces far more preload, which can snap the bolt during assembly or leave it with no fatigue margin.

Why does the calculator disagree with a published torque table?
Tables assume a particular preload percentage and thread condition without always saying so. One assuming 65% and oiled threads will read far lower than one assuming 75% and dry. This tool exposes both choices so you can match your source.
What is the nut factor K?
An empirical constant covering all the friction in the joint — under the head and in the threads. It is roughly 0.20 dry, 0.15 oiled, and 0.28 for dry zinc plating, and it varies ±25% between identical bolts.
How much of the torque becomes clamp force?
About 10%. Roughly half the torque is lost to friction under the head or nut face and half in the threads, leaving only a tenth to actually stretch the bolt.
Should I lubricate the threads?
Only if your torque figure assumes it. Lubricating a joint specified dry produces far more preload for the same torque, which can yield or snap the bolt during assembly.
What preload percentage should I target?
65–75% of proof load is the usual range for a reusable joint. That keeps the bolt elastic with margin for the external load, while still providing enough clamp force to stop the faces slipping.
Can I reuse a bolt?
Only if it stayed elastic. A bolt tightened past yield has permanently stretched and must be replaced — and many automotive fasteners are deliberately torque-to-yield, making them single-use by design.
Why does a bolted joint carry load by friction?
Preload clamps the faces together hard enough that friction between them resists the applied load. If the joint slips, the bolt starts carrying shear and bending it was not designed for, which invites fatigue failure.