A bolt that simply "holds two parts together" is doing far more than that. When you tighten it, you stretch it like a spring, and that stretch squeezes the joint members together with a force that often dwarfs the load the joint will ever see in service. A loose-feeling bolt and a properly preloaded one can look identical from the outside, yet behave like completely different machine elements. One rattles itself apart; the other survives millions of load cycles.
This article explains what bolt preload is, why it governs joint reliability, and how to convert a tightening torque into a clamp force with a worked example you can repeat at your bench.
Why this calculation matters
Preload is the quiet hero of every bolted joint. A correctly preloaded bolt resists loosening because the friction generated by the clamp force holds the nut in place. It resists fatigue because most of an external cyclic load is absorbed by the relaxing joint members rather than by the bolt itself. And it keeps gasketed joints sealed because the clamp force never drops below the seating pressure.
Under-tighten, and the joint can gap, leak, or shake loose. Over-tighten, and you yield the bolt during assembly or strip the threads, leaving nothing in reserve. The target preload therefore sits in a deliberate window — high enough to clamp firmly, low enough to stay safely below the bolt's proof strength. Getting that number wrong is one of the most common root causes of mechanical failure, and it is entirely avoidable with a short calculation.
The core formula
Start from the bolt's capacity. The proof load is the largest axial force the bolt can carry without measurable permanent set:
F_p = S_p * A_s
Here S_p is the proof strength of the property class and A_s is the tensile stress area — an effective area based on the mean of the pitch and minor thread diameters, not the nominal shank diameter.
Designers rarely tighten all the way to proof load. A common target preload is a fixed fraction of it, often around 75 % for reused, non-permanent joints:
F_i = 0.75 * F_p
The hard part is hitting that preload on the shop floor, because you cannot measure bolt stretch with a wrench — you measure torque. The two are linked by the torque-tension relation:
T = K * F_i * d
T is the tightening torque, d is the nominal bolt diameter, and K is the nut factor: a dimensionless lump that bundles together thread friction, under-head friction, and thread geometry. For plain, lightly oiled steel fasteners K is often taken near 0.2, but it can range from roughly 0.1 for well-lubricated assemblies to 0.3 for dry or slightly corroded ones.
That single coefficient is why torque control is imprecise. A large share of the applied torque is spent overcoming friction, not stretching the bolt. If friction varies by 25 %, your preload varies by about the same amount, even with a perfectly calibrated wrench.
A worked example
Take an M12 bolt of property class 8.8. From standard tables: nominal diameter d = 12 mm, tensile stress area A_s = 84.3 mm^2, and proof strength S_p = 580 MPa.
Step 1 — proof load.
F_p = S_p * A_s = 580e6 * 84.3e-6 = 48.9 kN
Step 2 — target preload at 75 % of proof load.
F_i = 0.75 * 48.9 = 36.7 kN
Step 3 — tightening torque, using a nut factor K = 0.2.
T = K * F_i * d = 0.2 * 36700 * 0.012 = 88 N.m
So roughly 88 N.m of tightening torque produces about 37 kN of clamp force in this M12 8.8 bolt. That clamp force is the number that actually matters — it is what holds the joint together, resists loosening, and shields the bolt from fatigue. The 88 N.m is just the indirect, friction-dependent means of getting there.
It is worth pausing on the scale. A 37 kN clamp force is equivalent to hanging nearly four tonnes from a single finger-sized bolt. That is the reservoir of force a good joint design quietly stores.
Common mistakes
Using the nominal diameter for stress area. A_s is smaller than the area of a circle at diameter d, because the threads remove material. Sizing the bolt on the shank area overestimates its strength and leads to over-tightening.
Treating the nut factor as a constant of nature. K depends on lubrication, plating, surface finish, and even how many times the joint has been assembled. Quoting torque to three significant figures while guessing K is false precision.
Forgetting that external load and preload are not additive. A preloaded bolt in a stiff joint sees only a fraction of any applied external force, because the joint members unload as the bolt stretches further. Adding the full external load to the preload badly overestimates bolt stress.
Ignoring preload loss after assembly. Embedding of surface asperities and gasket creep can shed a noticeable slice of the initial preload in the first hours or days. If the joint must stay tight, design with that relaxation in mind or re-torque.
Tightening to a torque without knowing the grade. A class 12.9 bolt and a class 8.8 bolt of the same size need very different torques. Mixing them up either yields the weaker bolt or leaves the stronger one barely snug.
Try the interactive NovaSolver calculator
Running these numbers once by hand builds intuition, but real design means sweeping grades, diameters, and friction values quickly. The Bolt Preload & Fatigue Design Calculator on NovaSolver lets you pick the property class and bolt size, set the nut factor and target preload fraction, then add a grip length and external load — and it returns the tightening torque, preload force, bolt stress, the joint load ratio, and both fatigue and overload safety factors in real time.
Related calculators
- Bolted joint calculator — to see how stiffness splits an external load between the bolt and the clamped members.
- Bolt fatigue calculator — for cyclic loading, where preload directly controls the alternating stress the bolt sees.
- Bolted flange calculator — for pressure-vessel and pipe flanges, where preload must keep a gasket sealed.
The full set of fastener and machine-element tools lives in the mechanical engineering tools hub.
Closing note
A bolt is a spring, and preload is how much energy you store in it. Size the bolt on its tensile stress area, choose a target preload as a sensible fraction of proof load, and remember that the torque-tension relation hangs on a friction coefficient you can only estimate. Treat the resulting torque value as a control method, not a guarantee — the clamp force is the real design quantity. Get the preload right and most loosening, leakage, and fatigue problems never appear in the first place.
Top comments (0)