Industry News

Fastener friction coefficient and surface roughness: taming preload scatter at the source

2026-10-18

Compiled by JIALI Industrial (164.hk) — a Hong Kong fastener supplier specialising in screws, bolts, nuts and selection know-how.

Why one torque wrench gives ten different preloads - the essence of factor K

Only roughly 10-15% of tightening torque actually becomes clamp load; the remaining 85-90% is consumed overcoming friction - about 40-50% in the thread pair and another 40-50% under the bolt head or nut bearing face. Engineering practice lumps this into the total torque factor K in the relation T = K·d·F, with K typically between 0.10 and 0.20. For the same batch of M12 bolts, K moving from 0.12 to 0.18 means one third less preload at identical torque. GB/T 16823.3 (aligned with ISO 16047 and JIS B1082) and the VDI 2230 guideline both treat the friction coefficient as the core variable of the tightening process: torque-controlled tightening typically scatters ±25-30% in preload, and batch-to-batch and piece-to-piece variation of the friction coefficient is the largest single contributor. So buyers and assemblers should ask the same question - what is this batch's friction coefficient, and how stable is it - instead of fixating only on the torque value and strength grade.

How surface roughness, plating and lubrication rewrite K

The friction coefficient is not a material constant; it is a function of surface condition. Thread and bearing-face roughness is usually held at Ra 1.6-3.2 μm: surfaces polished too smooth promote adhesion between micro-asperities and raise friction, while rougher surfaces (Ra above 6.3 μm) or burrs shrink the effective contact area and push local pressure up, widening K scatter rapidly. Plating is the other major variable: electroplated zinc in the dry state runs a friction coefficient of about 0.15-0.20, phosphating about 0.12-0.16, while zinc flake coating (ISO 10683) with a wax topcoat holds a steady 0.10-0.14. Molybdenum disulfide or dedicated tightening lubricants can cut K by 20-30% - per the VDI 2230 coefficient classes, the same bolts can drop from K = 0.16 to 0.11 once lubricated. If torque is not adjusted accordingly, preload overshoots by more than thirty percent, which is exactly why bolts snap and threads strip on the shop floor. Uneven plating thickness, poor passivation and transit-damaged burrs all make K wander within one batch; consistent surface condition matters more than a pretty surface.

Process and acceptance: three lines of defence for friction control

Line one is testing: run a thread-pair torque-clamp-load test per GB/T 16823.3 / ISO 16047 to measure total friction coefficient and factor K, and require a friction test report with each batch for critical products. Line two is process: a given joint must lock in plating and lubrication state - if the drawing assumes K = 0.14 for the torque, nobody should grab a spray lubricant on site; the installation instruction should state the K assumption behind the torque and its tolerance (e.g. K = 0.14 ± 0.02, corresponding to roughly ±15% preload scatter). Line three is acceptance: sample per GB/T 2828.1 with AQL 1.0-2.5, re-verify plating thickness per GB/T 5267.1, and spot-check thread and bearing-face roughness with a profilometer. High-strength bolt assemblies to GB/T 1231 should ship with the mean torque factor and its standard deviation (commonly 0.110-0.150 with standard deviation no more than 0.010). Writing friction coefficient and surface roughness into the purchase contract and drawing notes costs far less than re-torquing every joint afterwards. Need fasteners with stable friction coefficients and batch friction reports? Contact Chiali Industrial (164.hk) with sizes, plating and tightening details for a matching proposal and quotation.

Related Reading

Need a quote or technical advice? Tell us spec and quantity — we quote fast.

Get a Quote

← Back to News