Friction & assembly

Friction Coefficient 0.13–0.18 and Why Assembly Torque Drifts

Technical note · 2026-10-27 · reading time about 3 minutes

Zinc plated fasteners — Appearance reference only — not a test report.

Assembly torque is not a measure of clamping force. It is a measure of the total resistance the thread and bearing face offer. Change the friction, and the same torque produces a different clamp load — which is why a coating change that passes every corrosion test can still cause a line stop.

The relationship

The short form used on the line:

T = K × F × d

where T is tightening torque, F is clamp load, d is nominal diameter and K is the dimensionless nut factor — a single number that folds together thread friction, bearing friction and geometry. A typical K for as-received zinc-plated fasteners is around 0.20; for wax-lubricated systems, 0.10–0.15.

The arithmetic that catches people out. Going from K = 0.15 to K = 0.20 at the same torque reduces clamp load by about 25 %. If the joint was designed at the edge of its preload window, that is enough to cause loosening or fatigue failure — with no visible defect on the part.

Where friction scatter comes from

Lubricant distribution. Uneven wax or PTFE in the film means some threads are lubricated and some are not.
Film weight variation. Thicker films fill thread flanks and change the effective geometry.
Coating thickness on threads. Zinc-nickel with a thick film can reduce the thread clearance enough to change friction.
Surface roughness of the bearing face. A different stamping die or a worn washer changes K measurably.
Test method. K measured on a lubricated bolt in a laboratory differs from K on a dry, dusty assembly line.
Time and humidity. Some lubricants migrate or absorb moisture in storage.

What a controlled range looks like

Friction coefficient ranges published for coated fasteners are commonly 0.13–0.18 for wax-containing sealer systems, and 0.08–0.12 for systems with a high lubricant load. Those ranges describe the scatter within a validated process, not a target to hit on any single part.

What matters commercially is the width of the range. A process with K = 0.15 ± 0.01 is controllable. A process with K = 0.13–0.18 is only acceptable if the joint design accounts for the full window.

Measuring it

Use ISO 16047 or the customer's method; record torque, clamp load, and compute K per sample.
Test at least 10 assemblies; report mean and standard deviation, not just a range.
Test parts taken from production, not laboratory-prepared coupons.
Test after the full process: plating, passivation, sealing, baking, and any storage period the customer specifies.
Repeat after any change in sealer concentration, film weight or lubricant supplier.

When a customer says "torque is drifting", ask three questions: did the sealer batch change, did the film weight change, and were the parts baked? Those three account for most of the variation, and all three are visible in production records.