Bolt Torque Calculator

A bolt torque calculator converts the tightening torque you apply with a wrench into the preload (the axial clamp force) it creates in the bolt — and back again. That preload is the whole point of a bolted joint: it is the tension locked into the shank that squeezes the connected parts together, keeps the joint from separating or slipping, and stops the bolt from working loose under vibration and fatigue. Torque itself is only the means of installing that preload; it is never the design objective.

The relationship both directions rest on is the short-form torque equation (also called the Motosh or nut-factor equation):

T = K · F · d

• T — tightening torque applied to the nut or bolt head (N·m or lbf·ft);
• K — the nut factor (a.k.a. torque coefficient), a dimensionless number that lumps together thread friction, under-head friction and the thread lead. It is dominated by surface finish and lubrication, typically 0.10–0.25;
• F — the target preload (bolt tension, in N or lbf);
• d — the nominal bolt diameter (m or in).

This calculator does the full chain for you. Pick the diameter (metric M or imperial UNC), the strength class, the surface condition, and the fraction of proof load you want to reach; it computes the tensile stress area, the proof and yield loads, the target preload, and the torque — all pre-solved the moment the page loads, with the answer shown in both kN and lbf for preload and both N·m and lbf·ft for torque. Unlike a generic mechanical torque app, it also carries the real proof strengths of every common bolt class and the structural minimum pretension Tb that steel design codes require for slip-critical and pretensioned connections.

The full torque balance of a threaded fastener has three parts — the torque that stretches the bolt through the thread helix, the torque lost to friction on the thread flanks, and the torque lost to friction under the turning nut face:

T = F · [ (P / 2π) + (μt · rt / cosα) + μn · rn ]

where P is the thread pitch, μt and μn are the thread and under-head friction coefficients, rt and rn the effective thread and bearing radii, and α the thread half-angle. Because all of those geometric and friction terms scale with the diameter, the whole bracket collapses to a single dimensionless nut factor K times the diameter d, which is why the short form T = K·F·d works so well in practice. The catch is that K is not a friction coefficient — a "K = 0.20" joint does not have μ = 0.20. K is an empirical property of the whole fastener system: bolt, nut, washer, plating and lubricant together.

Tensile stress area Aₛ. The bolt does not resist tension on its nominal area — the threads reduce it. Design uses the tensile stress area, the area of a hypothetical round bar with a diameter halfway between the pitch and minor thread diameters:

metric:    Aₛ = 0.7854 · (d − 0.9382·P)²          (mm², P = pitch)
imperial:  Aₛ = 0.7854 · (d − 0.9743/n)²          (in², n = threads per inch)

For an M20 coarse bolt (P = 2.5 mm) this gives Aₛ = 244.8 mm² — matching the 245 mm² published in every fastener table. Every strength quantity below is Aₛ times a stress.

From stress to preload. Each class has a proof stress Sp (the stress the bolt can carry with no measurable permanent set). The proof load is Fp = Aₛ·Sp, and the target preload is a chosen fraction of it — commonly 75 % for reusable joints and up to 90 % for permanent ones (closer to the yield load Aₛ·Re). Enter that fraction and the calculator returns F, then T = K·F·d.

The number stamped on a bolt head is not decoration — it encodes the strength. This calculator carries the real, standardised values, not one generic curve:

ISO 898-1 property classes (metric). The two-part mark "x.y" means ultimate tensile Rm ≈ x·100 MPa and yield ≈ x·y·10 MPa:

ASTM F3125 (the standard that absorbed the old A325 and A490) — the workhorses of North-American steel construction:

(For imperial A325 the tensile strength drops from 120 ksi to 105 ksi above 1-in diameter; the calculator applies that reduction automatically.)

SAE J429 (imperial, common in mechanical and automotive work): Grade 2 (Sp 379 MPa / 55 ksi), Grade 5 (585 MPa / 85 ksi ≈ ISO 8.8), Grade 8 (827 MPa / 120 ksi ≈ ISO 10.9). Grade markings are the radial lines on the head — no lines for Gr 2, three lines for Gr 5, six for Gr 8.

Switch class in the tool and every downstream number — proof load, preload, torque and the whole torque chart — updates instantly.

If torque control gets a bad reputation, K is the reason. Because T = K·F·d, any error in K passes straight through to the preload: two joints torqued identically but with K differing by 25 % end up with preloads differing by 25 %. And K really does vary that much — it is set almost entirely by friction and lubrication, not by how hard you pull.

Typical published nut factors (Bickford, Introduction to the Design and Behavior of Bolted Joints; Fastenal / Machinery's Handbook):

Two practical consequences:

• Lubrication cuts the torque for the same preload roughly in half. A waxed galvanized bolt (K ≈ 0.12) needs about half the torque of the same bolt dry (K ≈ 0.25) to reach the same tension. Torque a lubricated bolt to the dry value and you will over-tension it — a common way to snap high-strength bolts.
• Never torque a galvanized structural bolt to a plain-finish chart. Hot-dip galvanizing raises K dramatically until the nut is lubricated; that is exactly why A325/A490 assemblies ship with a waxed nut.

For anything critical, calibrate K on your own fastener–lubricant combination (a bolt-tension calibrator or a load cell in a Skidmore-Wilhelm device). The table gives a defensible starting point, not a certified value — which is also why steel codes prefer preload-verifying methods over torque, as the next section explains.

This is where a structural bolt torque calculator parts ways with a mechanical one. In machine design you pick a preload as a fraction of proof load. In steel construction, slip-critical and pretensioned connections have a code-mandated minimum bolt pretension Tb that must be installed regardless — it is what develops the friction (faying-surface slip resistance) that the joint is designed on.

Both NBR 8800 (Tabela 20) and the AISC 360 / RCSC specification set the same value: 70 % of the minimum tensile strength,

Tb = 0.70 · Fu · Aₛ

The calculator computes Tb for the chosen diameter for both structural grades and shows the wrench torque K·Tb·d needed to reach it. A few benchmark values it reproduces exactly:

These match NBR 8800 Tabela 20 and AISC/RCSC Table 7.1 to the kilonewton.

Important: for structural work, torque is a last-resort installation method. RCSC ranks four: turn-of-the-nut, direct-tension-indicator (DTI) washers, tension-control (twist-off) bolts, and calibrated wrench. The calibrated-wrench method is the only one that uses torque, and it requires daily calibration on a bolt-tension device — with the torque set 5 % above K·Tb·d to cover scatter. Use the torque here to estimate and to sanity-check; verify the actual tension with turn-of-nut or DTIs on the real joint.

1. Choose the thread system — metric (M coarse) or imperial (UNC) — then pick the diameter from the list. The bolt sketch redraws with the diameter dimensioned AutoCAD-style.
2. Select the grade/class. ISO 8.8/10.9/12.9, ASTM A325/A490 or SAE Gr 2/5/8 — the proof, yield and tensile strengths load automatically.
3. Pick the surface / lubrication to set the nut factor K. This is the single biggest lever on the torque — dry vs waxed can halve it.
4. Set the preload target as a percentage of proof load (75 % reusable, up to 90 % permanent).
5. Read the results — they update with no "calculate" button. The KPI strip shows the tightening torque (N·m and lbf·ft), the preload (kN and lbf), the tensile stress area and the proof/yield loads. The T = K·F·d panel shows every intermediate number.
6. Check the structural panel for the NBR 8800 / AISC minimum pretension Tb of A325 and A490 at your diameter, and the bolt torque chart for the whole size series at your chosen class and surface — click any row to jump to that size.
7. Read the ±25 % scatter band on the torque KPI and in the derivation — it shows both the torque range for the same preload and the preload actually installed at the same torque, so you never treat the single number as exact. Tick fine thread (metric) to see the larger Aₛ and higher preload.
8. Open the connection-engine capacity card for the bolt's shear and tension design strength (Fnv/Fnt, φRn) pulled live from the CalcSteel connection engine, with the threads-in/out-of-plane toggle.
9. Export, share or hand off. Download the whole chart + result strip as CSV, print → PDF the sketch and results (no watermark), copy the permalink (every input travels in the URL, so a screenshot and the link reproduce your exact bolt), or open this bolt in the 3D editor as a prefilled base-plate connection.
10. Toggle SI ⇄ imperial at the top to flip the primary units; both systems are always shown so nothing is lost.

• Torque control is inherently scattered. Even with a good K, the torque method typically holds preload only to ±25–30 %. If you need tighter control, use angle (turn-of-nut), stretch measurement, or a load-indicating method.
• K is empirical. The table values are representative, not certified. Reused bolts, galling, dirty or damaged threads, hard vs soft washers, and temperature all move K. Calibrate for critical joints.
• Nominal diameter, coarse thread. The tool uses the nominal d and the standard coarse pitch for Aₛ. Fine threads have a slightly larger Aₛ (and thus preload for the same stress) — check a fine-thread table if that applies.
• Elastic behaviour assumed. Preload above ~90 % of proof load starts to yield the bolt; some methods (turn-of-nut on structural bolts) deliberately do this, but the T = K·F·d preload here is meant to stay elastic.
• Not a joint design. This sizes the installation of one bolt. It does not check bearing, tear-out, shear, prying, fatigue or the number of bolts — that is the connection design, which you can run in the CalcSteel editor and profile pages under NBR 8800 / AISC 360.
• Lubricant on the correct surfaces. K assumes lubrication where the standard intends it (under the turned element and on the threads). Random oil on only part of the joint invalidates the value.