Torque Calculations

How to Calculate Bolt Torque, Preload, and Torque Angle for Bolted Joints

The core math behind a bolted joint: preload from torque, the K-factor, torque-angle rotation, and audit sample size, worked with real units and numbers.

Start with the short-form torque equation, T equals K times D times F. T is torque in newton meters, D is nominal bolt diameter in meters, F is target preload in newtons, and K is the nut factor, a dimensionless lumped term for thread and bearing friction. For an M12 bolt at 0.012 m and a target preload of 30,000 N with K equal to 0.20, T equals 0.20 times 0.012 times 30,000, which is 72 Nm. Work in consistent units. If you prefer inch-pound, D in inches and F in pounds give T in inch-pounds, so an equivalent joint reads roughly 637 in-lb. The Bolt Preload Estimate calculator runs this both ways.

K is the input people get wrong most often. It is not the coefficient of friction, it bundles thread pitch geometry, thread friction, and underhead friction into one number. Typical values run 0.10 to 0.16 for lubricated or waxed fasteners, 0.17 to 0.22 for plain or as-received steel, and 0.20 to 0.30 for zinc plated dry. A shift in K from 0.20 to 0.16 at the same 72 Nm raises preload from 30,000 N to 37,500 N, a 25 percent jump. That single scatter is why torque-only control lands within plus or minus 25 to 30 percent of target clamp load even when the wrench is perfectly calibrated.

Set target preload from the fastener, not guesswork. Compute proof-based clamp as F equals a fraction of proof load, where proof load equals proof strength times the tensile stress area. For an M12 property class 8.8 bolt, tensile stress area As is 84.3 mm squared and proof strength is about 580 MPa, so proof load is roughly 48,900 N. Engineers commonly target 65 to 75 percent of proof for reusable joints, giving 31,800 to 36,700 N. Take the 70 percent case, 34,200 N, feed it into T equals K times D times F at K equal to 0.20, and torque becomes 0.20 times 0.012 times 34,200, or about 82 Nm.

Torque-angle control tightens scatter by using rotation past a snug point. The math splits into two stages. First torque to a snug or threshold value, typically 20 to 30 percent of final, to seat the joint and close gaps. Then rotate a defined angle. The stretch per turn is governed by pitch P, since one full 360 degree turn advances the nut one pitch. For an M12 by 1.75 coarse thread, one degree of rotation past snug advances 1.75 divided by 360, or 0.00486 mm. Rotating 90 degrees past snug advances about 0.437 mm of combined bolt stretch and member compression. The Torque Angle Workload calculator converts a target angle into cycle demand.

Convert that advance into preload through joint stiffness. Preload F equals the elastic advance delta times the combined stiffness, where 1 over combined stiffness equals 1 over bolt stiffness plus 1 over member stiffness. Bolt stiffness Kb equals E times As divided by grip length L. For the M12 8.8 example, E is 205,000 MPa, As is 84.3 mm squared, and a 40 mm grip gives Kb of roughly 432,000 N per mm. If member stiffness is about three times the bolt, combined stiffness is near 324,000 N per mm, so 0.437 mm of net elastic advance yields on the order of 30,000 N of preload once seating loss is subtracted.

Size your torque audit statistically, not by habit. To verify a running process, sample size depends on confidence and the defect rate you must detect. For attribute checks at 95 percent confidence, the classic rule uses n equals ln(1 minus C) divided by ln(1 minus p). To catch a 5 percent nonconformance rate, n equals ln(0.05) divided by ln(0.95), which rounds up to 59 joints with zero failures allowed. Drop the target defect rate to 1 percent and required n climbs past 298. The Torque Audit Sample Size calculator returns these counts so you audit enough joints without over-inspecting.

Check residual torque correctly during an audit. Breakaway or marked-angle methods measure a different quantity than installation torque, so do not compare them one to one. A first-movement breakaway reading typically reads 85 to 110 percent of the original installation torque because static friction differs from dynamic. A cleaner method is the marked-nut re-torque, where you mark the fastener, apply the spec torque, and record the additional angle to first movement. Under 5 to 10 degrees of extra rotation usually means the joint held preload; more than 15 degrees points to relaxation, embedment, or a soft joint that never reached target.

Close the loop with a joint-integrity screen before you trust any of the above. Compute the tension safety margin as preload divided by external service load per bolt. If a 6-bolt flange sees 90,000 N of separating force, each bolt carries 15,000 N, and a 30,000 N preload gives a 2.0 clamp-to-load ratio, above the 1.5 minimum many specs use for non-gasketed joints. The Joint Failure Risk calculator combines preload, external load, and stiffness ratio to flag joints likely to open or fatigue. Run the formulas above first, then let the calculators check your arithmetic against real limits.

Published 2026-07-01.