Engineering Bible

Bolt torque and preload

Preload and tightening torque for ISO metric bolts by the VDI 2230 method, with property class, friction, proof load and preload scatter.

Results are estimates. They do not replace the governing code, class or flag requirements, or a check by a competent person. Report a problem

Introduction

A bolted joint works because of preload, not torque: the clamp force holds the parts together and carries the external load. Torque is only the means of producing that preload, and most of it, typically 85–90 %, is spent overcoming friction in the thread and under the head. This calculator follows the VDI 2230 method used throughout European mechanical engineering: it determines the preload a bolt can take (90 % of yield under the combined tension and thread torsion of tightening, or a fraction of the ISO 898-1 proof load, or a value you give) and converts it to a tightening torque for the actual friction condition and bearing diameter. It reports the thread and under-head shares of the torque, the equivalent nut factor K, the stresses, and the minimum preload after torque-wrench scatter.

Marine and mechanical engineers use it for flange, foundation and coupling bolts when no manufacturer table applies; students use it to see why "T = 0.2 F d" is only a rough rule.

Formulas

Thread geometry (ISO 68-1, coarse pitch P, nominal diameter d):

d₂ = d − 0.6495 P        pitch diameter
d₃ = d − 1.2269 P        minor diameter
d_s = (d₂ + d₃) / 2
A_s = π d_s² / 4          tensile stress area

Proof load (ISO 898-1):

F_p = S_p × A_s

Assembly preload at 90 % yield under combined stress (VDI 2230 eq. R8/1):

F_M = A_s × 0.9 R_p0.2 / √( 1 + 3 [ (3/2)(d₂/d_s)( P/(π d₂) + 1.155 μ_G ) ]² )

Tightening torque (VDI 2230 eq. R8/4):

M_G = F_M (0.16 P + 0.58 d₂ μ_G)          thread torque
M_K = F_M μ_K D_Km / 2                     under-head friction torque
M_A = M_G + M_K
D_Km = (d_w + d_h) / 2                     mean bearing diameter: head bearing OD and hole diameter

Stresses during tightening:

σ = F_M / A_s,   τ = M_G / (π d_s³ / 12),   σ_red = √(σ² + 3 τ²)

Equivalent nut factor and the rule of thumb:

K = M_A / (F_M d)        (≈ 0.20 for μ_G = μ_K = 0.15, Shigley)
T₀.₂ = 0.2 F_M d

Minimum preload after scatter:

F_M,min = F_M / α_A      α_A 1.6–2.0 for a standard torque wrench
  • μ_G, μ_K — thread and under-head friction coefficients; R_p0.2 — yield or 0.2 % proof strength; S_p — ISO 898-1 proof stress

Assumptions and limitations

  • The VDI torque formula uses the standard approximations 0.16 ≈ 1/(2π) and 0.58 ≈ 1/(2 cos 30°) for 60° threads and treats friction as Coulomb friction with a constant coefficient.
  • Friction is the dominant uncertainty. The presets are mid-values of the VDI 2230 friction classes; the real coefficient of a given bolt, nut, coating and lubricant can lie anywhere in the class range, which changes the torque for a given preload by ±30 %.
  • The default bearing diameter is for an ISO 4014/4017 hex head on an ISO 273 medium clearance hole. Washers, flange bolts, socket-head cap screws and nuts have different bearing areas; enter D_Km when it matters.
  • Strength values are ISO 898-1 minimums (8.8 is 660/600 MPa above M16). Stainless A2-70 and A4-80 use R_p0.2 as proof stress and are specified up to M24.
  • The 90 %-yield preload uses the plastic torsion modulus π d_s³/12 as in VDI 2230, which is why it reproduces the VDI tables; a purely elastic von Mises check would give about 8 % less.
  • Torque control gives preload scatter; the minimum preload F_M/α_A is what the joint can rely on. Angle- or yield-controlled tightening is more accurate but outside this calculator.
  • No joint analysis: embedding losses, external load sharing (load factor), thread engagement and stripping, fatigue, gaskets and temperature effects require the full VDI 2230 procedure or an equivalent.
  • Coarse ISO threads only.

Worked example

M12 × 1.75, property class 8.8, dry zinc-plated (μ_G = μ_K = 0.14), hex head on a 13.5 mm hole.

  1. Geometry: d₂ = 12 − 0.6495 × 1.75 = 10.863 mm, d₃ = 9.853 mm, d_s = 10.358 mm, A_s = 84.3 mm².
  2. Strength: R_p0.2 = 640 MPa, S_p = 580 MPa → proof load 580 × 84.3 = 48.9 kN.
  3. Preload at 90 % yield: the bracket term is 1.5 × (10.863/10.358) × (1.75/(π × 10.863) + 1.155 × 0.14) = 0.335; √(1 + 3 × 0.335²) = 1.156; F_M = 84.3 × 576 / 1.156 = 42.0 kN. VDI 2230 Table A1 lists 41.9 kN.
  4. Torque: D_Km = (16.63 + 13.5)/2 = 15.07 mm. M_G = 42.0 × (0.16 × 1.75 + 0.58 × 10.863 × 0.14) = 42.0 × 1.162 = 48.8 N·m; M_K = 42.0 × 0.14 × 15.07/2 = 44.3 N·m; M_A = 93 N·m, the VDI table value.
  5. Nut factor: K = 93 / (42.0 × 12) = 0.185. The 0.2 rule would give 101 N·m, 8 % high.
  6. Scatter: with α_A = 1.7 the minimum preload is 42.0 / 1.7 = 24.7 kN.

With MoS₂ paste (μ = 0.08) the same bolt reaches 44.9 kN preload at only 61 N·m. Using the dry torque on a lubricated bolt would overload it: that is why the friction condition must match the torque table.

FAQ

Why not just use T = 0.2 F d?

Because K ranges from about 0.10 (well lubricated) to 0.30 (dry, rough). For a given torque the preload can vary by a factor of three. K = 0.2 happens to match μ ≈ 0.15.

Should the preload be 90 % of yield? That sounds high.

During tightening, yes: VDI 2230 and most European tables use 90 % of yield including the torsion. The torsional stress largely relaxes after tightening, leaving a safe tensile stress. US practice quotes 75 % of proof load for reusable and 90 % for permanent joints; both options are available.

What is the difference between yield and proof load?

Proof stress S_p (ISO 898-1) is about 90 % of the minimum yield; a bolt loaded to its proof load must show no permanent set. Preloads are usually set relative to proof or yield, never to ultimate strength.

Do I need a washer?

A hardened washer increases D_Km (more under-head torque) and protects the surface. Enter the washer outside diameter and the hole diameter via the "given" bearing option.

How accurate is torque tightening?

±17 % to ±43 % on preload for a torque wrench depending on how well the friction is known (VDI α_A 1.4–2.0); impact wrenches are far worse. If the joint is critical, use angle control or ultrasonic length measurement.

Can I use the result for fine threads or UNC/UNF bolts?

Not directly. Fine threads have a larger stress area and smaller pitch; inch threads have different geometry and SAE grades. The formula structure is the same if you compute d₂, d₃ and A_s for that thread.

Does the bolt loosen if I lubricate it?

Lubrication lowers the torque needed for a given preload; it does not reduce the clamp force. Loosening is governed by preload, joint stiffness and transverse movement, not by lubrication as such.

Related calculators

Embed this calculator on your site

Copy this HTML into your page. The calculator stays free to use and links back here.

By Christos Papandreou. How the numbers are checked.