Bolt Preload and Torque: Getting the K-Factor Right

You do not torque a bolt to hold two parts together — you torque it to stretch it. Tightening turns the fastener into a very stiff spring, and the tension locked into that spring (the preload) is what actually clamps the joint, resists fatigue, and keeps the nut from backing off. The trouble is that you cannot see tension with a wrench. You can only feel torque, and the bridge between the two is a single, slippery number: the nut factor, K. Get K wrong and your carefully specified 75%-of-proof preload can land anywhere from loose to snapped.

The Torque–Tension Relationship

The workhorse equation every mechanical engineer should have memorized is short:

It looks trivial, and that is exactly the danger. Roughly 90% of the applied torque is lost to friction — about 50% under the turning nut face and 40% in the threads — leaving only 10% to actually stretch the bolt. The nut factor K bundles all of that friction, the thread geometry, and the bearing-face geometry into one empirical coefficient. Because friction dominates, K is fundamentally a friction number wearing a geometry costume.

Choosing K: The Number That Decides Everything

K is not a material property you can look up to three decimals. It is a range that depends on plating, lubricant, surface finish, and even how many times the fastener has been used. Typical starting values for steel fasteners:

Condition	Typical K
Plain (as-received, light oil film)	0.20
Zinc-plated, dry	0.22 – 0.25
Cadmium-plated	0.16
Molybdenum-disulfide / wax lubricant	0.10 – 0.13
PTFE-based anti-seize	0.08 – 0.12

Notice the spread. A lubricated bolt at K = 0.12 produces nearly twice the preload of a dry one at K = 0.22 for the same wrench setting. This is why blindly applying a “handbook” torque to a lubricated fastener is a classic way to yield or snap it — the torque table almost certainly assumed a dry K. If you change the lubricant, you must change the torque.

Worked Example: A 1/2-13 Grade 8 Bolt

Take a 1/2″-13 UNC Grade 8 bolt. Its tensile stress area is about 0.1419 in², and the proof strength of Grade 8 is 120,000 psi. A common target is 75% of proof load:

• Proof load = 0.1419 in² × 120,000 psi = 17,000 lbf
• Target preload F = 0.75 × 17,000 = 12,750 lbf

With a dry zinc-plated bolt, K ≈ 0.22 and D = 0.5 in:

T = 0.22 × 0.5 in × 12,750 lbf = 1,403 lbf·in ≈ 117 ft·lbf

Now lubricate that same bolt with anti-seize (K ≈ 0.10) and hold the wrench setting at 117 ft·lbf. Solving for F gives about 28,000 lbf — well past the 17,000 lbf proof load. The bolt yields, and you never touched the torque wrench setting. Same number on the dial, completely different outcome.

Preload Scatter and Why the Method Matters

Even with the right K, torque control is the least precise way to set preload. Because you are inferring tension through a friction curtain, torque-only tightening carries roughly ±25–30% preload scatter. Better methods narrow that band dramatically:

• Torque control: ±25–30% — cheap, fast, friction-dependent.
• Torque-angle (turn-of-nut): ±15% — snug, then rotate a specified angle into the plastic region.
• Bolt stretch / ultrasonic: ±3–5% — measures elongation directly, friction-independent, used on critical joints.

For most machine design a calibrated click-type torque wrench and a controlled K are entirely adequate — just respect the scatter and never design a joint that only survives at the top of the preload band. If you are chasing repeatability, a decent click-type torque wrench with a documented calibration is the single best value upgrade a shop can make; the sloppy adjustable wrench in the toolbox is not a preload instrument.

Don’t Forget the Joint

Preload only helps if the clamped members are stiff relative to the bolt. In a properly designed joint, most of an external working load is absorbed by the reduction in clamp force, not by added bolt tension — which is exactly why preloaded joints resist fatigue so well. A soft gasket or a spongy flange undermines that mechanism, so treat joint stiffness as part of the fastener design, not an afterthought.

Let the Calculator Handle the Friction Math

The torque–tension equation is simple, but the lookups — stress area, proof strength by grade, K by coating, unit conversions — are where mistakes creep in. Our bolt torque calculator takes your bolt size, grade, target preload percentage, and nut factor and returns the tightening torque directly, so you can see instantly how switching from dry to lubricated changes the setting. For the surrounding structural work, the base plate calculator and column buckling calculator round out the connection design.