How the torque numbers are made

Every bolt torque on this site is calculated with T = K × D × F at 75% of proof load, from published proof stresses, stress areas and nut factors. Every vehicle torque is quoted as printed in the manufacturer's document, with its page. Neither is estimated or copied from another chart.

By the Torque Reference editors 15 sources cited Last reviewed 21 Oct 2026 Method

Contents9 sections

Key takeaways

Torque T = K × D × F, the short-form equation in the Fastenal guide and NASA RP-1228 ⁠¹ ⁠².
F is 75% of the bolt's proof load: proof stress × tensile stress area × 0.75 ⁠¹.
Five thread conditions, each with one sourced nut factor: 0.20, 0.18, 0.15, 0.16 and 0.11.
Real clamp load scatters about ±25% around the calculated value with a torque wrench ⁠².
Vehicle values are copied from owner's manuals and OEM bulletins, not converted, with the document and page.

The bolt torque equation

Bolt torque (the SAE, metric and grade charts and the calculators) is calculated from standards; no third-party torque chart is used as an input.

Torque comes from the short-form relationship

T = K × D × F

where T is the tightening torque, K the nut factor, D the nominal bolt diameter and F the target clamp load. Fastenal’s technical reference guide uses this form ⁠¹, and NASA RP-1228 gives it as T = KFd ⁠².

Clamp load basis. F is 75% of the bolt’s proof load. Fastenal gives F as “75% of bolt material proofload for standard bolts” and 90% for special applications such as structural bolts, and its own torque tables use 75% ⁠¹. Proof load is itself held below yield; NASA RP-1228 notes it is usually about 75% of theoretical yield, to prevent inadvertent yielding through torque measurement inaccuracies ⁠².

Other conventions exist, which is partly why published charts disagree. Bolt Science describes tables that assume a direct stress of usually 75% of yield, and a more consistent approach that targets 90% of yield as the combined tensile and torsional stress ⁠³. RCSC sets structural pretension at 70% of minimum tensile strength ⁠⁴. Our charts always state the 75%-of-proof basis.

Proof load = proof stress × tensile stress area. The proof stresses are:

Proof stresses used
Grade	Size range	Proof stress	Source
SAE Grade 2	1/4 to 3/4 in	55,000 psi	⁠¹
SAE Grade 2	7/8 to 1-1/2 in	33,000 psi	⁠¹
SAE Grade 5	1/4 to 1 in	85,000 psi	⁠¹
SAE Grade 5	over 1 to 1-1/2 in	74,000 psi	⁠¹
SAE Grade 8	1/4 to 1-1/2 in	120,000 psi	⁠¹
ISO class 4.6	all	225 MPa	⁠⁵
ISO class 8.8	up to M16	580 MPa	⁠⁵
ISO class 8.8	above M16	600 MPa	⁠⁵
ISO class 10.9	all	830 MPa	⁠⁵
ISO class 12.9	all	970 MPa	⁠⁵

Scroll the table sideways to see every column.

Inch values come from the Fastenal guide’s reproduction of SAE J429 ⁠¹. Metric values are from ISO 898-1:2009 Table 3; the 2013 edition is described in its foreword as a minor revision ⁠⁶.

Tensile stress areas. Inch areas are the tabulated values in the Fastenal guide’s appendix, checked against the ASME B1.1 formula As = 0.7854 (D − 0.9743/n)², where n is threads per inch ⁠⁷. Metric coarse-thread areas are the nominal stress areas in ISO 898-1 Tables 4 and 5 ⁠⁵. Metric fine-thread areas use the Fastenal table where it lists the pitch, and otherwise As = π/4 × ((d2 + d3)/2)², with d2 = d − 0.649519P and d3 = d − 1.226869P. Second fine pitches rest on the ISO 262 table as reproduced by Wikipedia ⁠⁸, a secondary source we flag as such.

Thread conditions and nut factors

K is “not to be confused with the frictional coefficient”; it lumps together thread geometry, thread friction and underhead friction, and published tables vary ⁠¹. We use one value per condition, each from a named document:

Nut factors used on every chart and calculator
Condition	K	Source
Plain steel, dry (as-received)	0.20	Fastenal torque table notes ⁠¹
Zinc-plated, dry	0.18	Fastenal torque table notes ⁠¹
Lubricated (oil or grease)	0.15	Fastenal torque table notes ⁠¹
Copper or metal-free anti-seize	0.16	LOCTITE C5-A TDS ⁠⁹
Moly paste	0.11	LOCTITE LB 8012 TDS ⁠¹⁰

These are indicative: Fastenal’s lab ranges run 0.17 to 0.22 for zinc and 0.12 to 0.16 lubricated ⁠¹, and NASA cautions that K = 0.2 “should not be used blindly” ⁠². Where a maker publishes no K, as for LOCTITE LB 771 nickel anti-seize ⁠¹¹ or LOCTITE 243 ⁠¹², we don’t invent one.

Units and rounding

Conversions use NIST SP 811: 1 lbf·ft = 1.355818 N·m, 1 lbf·in = 0.1129848 N·m, 1 lbf = 4.448222 N, 1 kgf·m = 9.80665 N·m ⁠¹³. Results are rounded only for display:

torque under 10 ft-lb is shown in whole in-lb, otherwise whole ft-lb;
N·m to one decimal under 10 N·m, otherwise whole;
clamp load to the nearest 10 lbf for inch bolts and 0.1 kN for metric.

Check. Fastenal’s worked example, a 1/2-13 zinc-plated Grade 5 cap screw at K 0.20 and 75% of proof, comes to 75 ft-lb ⁠¹. Our 1/2-13 Grade 5 at the same K of 0.20 also gives 75 ft-lb (102 N·m).

Uncertainty

A calculated torque targets a clamp load; it does not guarantee one. The IFI figures reproduced by NASA put torque-wrench control at ±25% of preload ⁠². NASA-STD-5020B’s default, without test data, is 25% for lubricated and 35% for non-lubricated or as-received hardware ⁠¹⁴. RCSC warns that torque-to-pretension variation as high as ±40% must be anticipated for structural bolts ⁠⁴. The 75% basis leaves room for most of that scatter below proof load. More on this in torque vs tension.

What we don’t calculate

Pretensioned structural bolts. RCSC prohibits installation torque taken from tables or equations without verification ⁠⁴. The structural pages give RCSC’s pretension tables and methods instead.
Torque-to-yield and torque-plus-angle bolts. These follow the maker’s procedure; see torque-to-yield bolts.
ASTM A307 bolts by a proof-load basis. A307 Grade A specifies a minimum tensile strength of 60 ksi ⁠¹⁵ but no proof load, so the 75%-of-proof basis has no input.
Any vehicle torque. Wheel, drain plug and spark plug values are quoted, never computed.
Joint design. No checks of shear, fatigue, thread stripping in soft materials, gasket stress or prevailing-torque locknuts.

How vehicle data is sourced and checked

Vehicle values come from the manufacturer’s own documents: owner’s manuals published by the manufacturer, and OEM service bulletins or recall documents hosted by NHTSA. For each value we record the document title, the URL, the page, and the sentence or table row exactly as printed. Values are stored in the units printed, not converted or rounded; where a manual prints a range, the range is shown.

Each value is tied to the model year of its manual. Years between two checked manuals of the same generation are assumed to share the value, as the vehicle pages state. Where a manual’s applicability is ambiguous, the entry is flagged for re-checking against an OEM-hosted copy. When the official file could not be downloaded directly, an archived copy of the exact official URL was used, and the official URL is the one cited.

Spark plug values come from the plug makers’ own charts (NGK, Denso, Champion) and the engine makers’ bulletins, and are shown as each one prints them. Forums, parts-store blogs and other torque websites are used only as cross-checks, never as the source of a number.

Corrections

If a value does not match its source, or a manufacturer has changed a specification, use the “Report an error” link under the sources on any page, or the contact page. We check the report against the cited document, correct the page and change its review date.

Sources

Spotted a value that doesn’t match its source? Report an error