In our article on bolt preload, we already mentioned the friction coefficient as one of the most uncertain parameters in the calculation. This article dives deeper into this specific topic: which values to use, why they vary so much, and how to reduce this uncertainty when possible.
Two coefficients, not one
The "friction coefficient" isn't a single value: the ECSS methodology distinguishes two, physically different and not to be confused.
- µ th (thread friction coefficient) — governs friction between the bolt's thread flanks and the tapped hole (or nut) during tightening.
- µ uh (under-head friction coefficient) — governs friction between the underside of the bolt head (or washer) and the clamped surface.
These two coefficients enter separately into the torque-to-preload conversion formula, and nothing requires them to be identical — a bolt lubricated in the thread but seated on a dry washer will have two quite different values.
Typical values by surface treatment
The following ranges are commonly used engineering orders of magnitude, provided for guidance — they don't replace an actual characterization test when the application's criticality justifies it:
- Bare, dry steel — 0.15 to 0.25, sometimes higher depending on surface condition (rust, rough machining).
- Lightly oiled — 0.10 to 0.16, the range most commonly used by default absent specific data.
- Zinc plated / phosphated — 0.12 to 0.20 depending on coating thickness and quality.
- Dry-film MoS2 or PTFE coating — 0.08 to 0.13, among the lowest and most reproducible.
- Stainless on stainless (no lubricant) — 0.15 to 0.30, with a galling risk to consider separately.
These ranges are deliberately wide: they reflect real variability, not measurement imprecision. This is precisely why the Fv,min/Fv,max bracket exists rather than a single preload value.
Why so much scatter?
Friction at a bolt interface depends on many simultaneous factors, difficult to individually control in production:
- exact surface condition (roughness, presence of oxidation),
- the nature and amount of lubricant actually applied,
- tightening speed,
- the number of times the bolt has been reused (friction evolves over successive tightening/loosening cycles),
- ambient humidity and temperature during tightening.
The same batch of bolts, with theoretically identical surface treatment, can therefore produce very different measured preloads from one part to another for a strictly identical applied torque.
Impact on the preload bracket
The wider the uncertainty on µ, the wider the calculated Fv,min/Fv,max bracket — which directly penalizes safety margins in both directions: a higher Fv,max brings the bolt closer to yielding during tightening, and a lower Fv,min brings the joint closer to slipping or separating in service.
In practice, narrowing the µ range (via better-controlled surface treatment, or specific characterization) is often the most effective lever to improve a marginal safety margin — sometimes more effective than changing the bolt diameter.
How to reduce uncertainty
- Specific characterization — measure the torque actually needed to reach a target preload on a representative sample, rather than using a generic catalogue value.
- Surface treatment control — a dry-film coating applied at the factory (Dacromet-type, MoS2) is significantly more reproducible than oil applied manually during assembly.
- Tightening method less sensitive to friction — angle control or direct elongation measurement largely bypass the uncertainty on µ, unlike torque alone.
- Batch traceability — documenting the surface treatment batch used allows narrowing the retained range if characterization data exists for that specific batch.
Common mistakes
- Using the same value for µ th and µ uh without justification — sometimes correct by coincidence, but rarely rigorous.
- Reusing a general literature value without verifying it matches the surface treatment actually used on the project.
- Ignoring the effect of the number of tightening cycles on a reused bolt (testing, requalification) — friction tends to decrease after the first few cycles.
Automating this with BoltCore
BoltCore offers pre-configured friction coefficient ranges by surface treatment, with the option to enter a value directly from a specific characterization test if you have one.
BoltCore automatically propagates this uncertainty into the preload bracket and all 11 safety margins, with no additional manual calculation.
Try BoltCore for free →Frequently asked questions
Should I always use the lowest value in the range to be conservative?
No — a lower µ favors some margins (less torque lost to friction, higher preload for a given torque) but penalizes others (risk of yielding during tightening). The full Fv,min/Fv,max bracket must be checked at both ends.
Does the friction coefficient change with temperature?
Potentially, especially for organic coatings (oils, waxes) whose viscosity varies with temperature. For cryogenic or high-temperature applications, characterization specific to the service temperature is recommended rather than a room-temperature value.
Does a bolt that's been tightened and loosened have the same µ on reuse?
Generally not — friction tends to decrease after a first tightening cycle (running-in of contact surfaces), which is why many standards require inspection or replacement of reused critical bolts rather than simple reuse without control.