One torque setting on one batch can still leave preload spread over thirty percent, and the cause is friction scatter, not the wrench. Torque is a means, clamp force is the goal, and between them sits a conversion governed entirely by surface condition.
The friction coefficient μ describes the physical surfaces; during tightening two of them exist, the thread pair and the bearing face, and standards report their combined value. The torque factor K is an engineering shortcut that folds thread geometry and both friction faces into one number. K is not a material constant; it moves with geometry and surface state, as GB/T 16823 and VDI 2230 both note.
ISO 16047 and GB/T 16823.2 use a dedicated machine that records torque and clamp force together and separates thread friction from bearing friction. Keep the speed at the specified rate, test parts in the supplied surface condition including coating and lubricant, and sample widely when scatter looks large.
Lubrication ranks first, since grease can cut K by twenty to thirty percent. Coating type and thickness come next, then bearing face conditions such as washer hardness, surface treatment and repeated reassembly, with roughness and cleanliness also involved. Any change moves the K distribution.
Build a K database from incoming and coating lot sampling instead of copying handbook values; set the torque window from measured scatter and check limits against yield; use torque plus angle on critical joints to reduce sensitivity; and write coating and lubrication into drawings and contracts, requiring retest after any supplier change.
Treating K as a fixed 0.2, oiling parts without adjusting the setting, reusing pre plating data after plating, and keeping flat washer K where spring washers sit under the head all recur in failure cases. Zhuhai Jiali Hardware (FASTKING) supplies fasteners with specified coating and lubrication plus friction test data. See the screw calculator directory or contact us.