Torque Spec Tables (Fastener)

Torque spec tables with clamp load by size and grade: ISO 898-1 8.8, 10.9, 12.9 (M3 to M12) and SAE J429 Grade 5, 8 (1/4 to 3/4 in), method stated.

Fastener torque specs are calculated, not measured. A published torque value is the number that should stretch a bolt to a target clamp load, so a good chart always pairs the two. These tables cover both systems manufacturing work uses: metric ISO 898-1 property classes 8.8, 10.9, and 12.9 from M3 to M12, and SAE J429 Grade 5 and Grade 8 from 1/4-20 to 3/4-10 UNC. Every value follows one method, T = K x F x d with clamp load at 75% of proof load, and lists dry (K = 0.20) and lubricated (K = 0.15) torque side by side. Treat them as assembly guidance: the clamp load your joint actually sees moves with lubrication, plating, and friction.

How the numbers are calculated

A torque table starts from proof load. Proof stress per ISO 898-1:2013 is 580 MPa for class 8.8 (sizes to M16), 830 MPa for 10.9, and 970 MPa for 12.9. Multiply by the tensile stress area of the thread and you get proof load, the highest axial load the bolt carries with no permanent set. The table then targets clamp load at 75% of proof load and converts it to torque with T = K x F x d, where d is nominal diameter and K bundles all the friction. The imperial table works the same way: SAE J429 sets proof stress at 85,000 psi for Grade 5 (sizes 1/4 through 1 inch) and 120,000 psi for Grade 8 (sizes 1/4 through 1 1/2 inches), applied to the standard UNC stress areas.

The K factor decides the answer

K is where the honesty belongs. These tables use K = 0.20 for dry plain threads and 0.15 lubricated, and Fastenal’s charts note zinc-plated (waxed) threads varying 0.12 to 0.16. NASA’s Fastener Design Manual (RP-1228) calls 0.2 the commonly assumed value, calls 0.15 a realistic value for steel on steel, and warns that the underlying coefficient of friction can run from 0.04 to 1.10 depending on materials and lubricants. That spread is why the lubricated column sits about 25% below the dry column for the same clamp load.

Reading the metric table

Property-class numbers encode strength: 8.8 means 800 MPa minimum tensile strength with yield at least 640 MPa, 10.9 means 1040 MPa, and 12.9 means 1220 MPa. Class 8.8 sits near Grade 5 (827 MPa) and 10.9 near Grade 8 (1,034 MPa), but the threads do not interchange, so stay in one system per assembly. For thread geometry, pitch, and class of fit, see the thread standards page; this page carries the torque data only.

Worked example: M10, class 10.9

An M10 x 1.5 coarse thread has a tensile stress area of 58.0 mm². Proof stress of 830 MPa gives a proof load of 830 x 58.0 = 48,140 N. The clamp target is 75% of that, 36,105 N. Dry torque is T = 0.20 x 36,105 N x 0.010 m = 72.2 Nm; lubricated, K drops to 0.15 and the torque to 54.2 Nm for the same clamp load.

Reading the imperial table

Grade 5 is a medium-carbon steel bolt proof-stressed to 85,000 psi with 120,000 psi minimum tensile strength; Grade 8 is a medium-carbon alloy bolt at 120,000 psi proof and 150,000 psi tensile. Torque on 1/4-20 and 5/16-18 is listed in pound-inches because the numbers are small; 3/8 and larger read in foot-pounds.

Worked example: 1/2-13 UNC, Grade 8

A 1/2-13 thread has 0.1419 sq in of tensile stress area. Proof load is 0.1419 x 120,000 = 17,028 lbf (the table rounds to 17,000 lbf), the clamp target is 75% of that at 12,771 lbf, and dry torque is 0.20 x 12,771 x 0.5 = 1,277 in-lb, or 106 ft-lb, which is the table value.

Why published torque charts disagree

Look up one bolt in three references and you will get three torques, all defensible. This table targets 75% of proof load. EngineeringToolBox lists M10 class 8.8 at 57 Nm because it assumes lightly oiled threads and 85% of proof load, and VDI 2230-based industry guidelines list M12 class 8.8 at 86 Nm because they target roughly 90% of yield with a mid-range friction coefficient. The clamp target and the friction assumption are the two levers. A 10 to 20% spread between reputable charts is method, not error.

Putting a torque callout on the drawing

A torque value is only complete with its condition. Write the grade, the value, and the thread condition together, for example: M10 x 1.5 - 10.9 hex cap screw, torque 72 Nm dry; or 1/2-13 UNC Grade 8, torque 106 ft-lb dry. If assembly will be lubricated or waxed, carry the lubricated value, or the fastener gets over-tightened by roughly the drop in K. Check tool access too: a recessed fastener no socket can reach is a design problem. Thread depth, tap drill, and hole rules for machined parts sit with design for manufacturing.

Limitations and safety

These values are assembly guidance for carbon and alloy steel fasteners under ISO 898-1 and SAE J429, not working limits for every joint. Real clamp load varies with lubrication, plating, thread condition, and tool calibration, and both Fastenal and NASA stress that torque is an indirect indication of tension, ranking tightening methods from a mechanic’s feel up to strain gauges on the bolt. For a joint whose failure matters, size the clamp load to the loads and gaskets it must hold, confirm values against the current editions of ISO 898-1 and SAE J429, and use a controlled tightening method rather than a chart value alone. Terms such as proof load and clamp load are defined in the manufacturing glossary.

About this data

Methodology: Torque and clamp load for coarse-thread bolts, calculated as T = K x F x d (K factor times clamp force times nominal diameter) with clamp load F set at 75% of proof load. Proof load = proof stress x tensile stress area per ISO 898-1:2013 (classes 8.8, 10.9, 12.9: proof stresses 580, 830, 970 MPa) and per SAE J429 (Grade 5: 85,000 psi through 1 in; Grade 8: 120,000 psi through 1 1/2 in; the table covers 1/4 to 3/4 in). K = 0.20 dry plain threads, 0.15 lubricated; the Fastenal charts note zinc-plated (waxed) threads varying 0.12 to 0.16. Metric M5 to M12 rows reproduce the published Fastenal torque-tension chart; M3 and M4 rows apply the same formula to ISO 898-1 proof loads because that chart starts at M5 (cross-checked against a VDI 2230 guideline; values fall within about 5 to 12%, consistent with the different clamp targets). Imperial rows reproduce the published Fastenal coarse-series chart with proof load computed as stress area x proof stress. Values are assembly guidance for carbon and alloy steel fasteners only.
Sources: Fastenal engineering torque-tension charts, ISO 898-1 and SAE J429 series (values M5 to M12 and all imperial torque and clamp figures).
Proof stresses per ISO 898-1:2013 and SAE J429; K factor discussion per NASA RP-1228 Fastener Design Manual.
How to read this: Find your size and grade row. Proof load is the standard minimum; clamp load is the assembly target (75% of proof); the two torque columns are for dry plain threads and lubricated threads respectively. Metric rows read in newtons (N) and newton-metres (Nm); inch rows read in pound-force (lbf) with in-lb torque on 1/4-20 and 5/16-18 and ft-lb on all larger sizes.

Recommended tightening torque and clamp load by size and grade
fastener	stress area	proof load	clamp load (75% of proof)	torque dry (K = 0.20)	torque lubricated (K = 0.15)
M3 x 0.5, class 8.8	5.03 mm²	2,920 N	2,190 N	1.3 Nm	1.0 Nm
M4 x 0.7, class 8.8	8.78 mm²	5,100 N	3,825 N	3.1 Nm	2.3 Nm
M5 x 0.8, class 8.8	14.2 mm²	8,236 N	6,177 N	6.2 Nm	4.6 Nm
M6 x 1.0, class 8.8	20.1 mm²	11,658 N	8,744 N	10.5 Nm	7.9 Nm
M8 x 1.25, class 8.8	36.6 mm²	21,228 N	15,921 N	25.5 Nm	19.1 Nm
M10 x 1.5, class 8.8	58.0 mm²	33,640 N	25,230 N	50.5 Nm	37.8 Nm
M12 x 1.75, class 8.8	84.3 mm²	48,894 N	36,671 N	88.0 Nm	66.0 Nm
M3 x 0.5, class 10.9	5.03 mm²	4,180 N	3,135 N	1.9 Nm	1.4 Nm
M4 x 0.7, class 10.9	8.78 mm²	7,290 N	5,468 N	4.4 Nm	3.3 Nm
M5 x 0.8, class 10.9	14.2 mm²	11,786 N	8,840 N	8.8 Nm	6.6 Nm
M6 x 1.0, class 10.9	20.1 mm²	16,683 N	12,512 N	15.0 Nm	11.3 Nm
M8 x 1.25, class 10.9	36.6 mm²	30,378 N	22,784 N	36.5 Nm	27.3 Nm
M10 x 1.5, class 10.9	58.0 mm²	48,140 N	36,105 N	72.2 Nm	54.2 Nm
M12 x 1.75, class 10.9	84.3 mm²	69,969 N	52,477 N	126 Nm	94.5 Nm
M3 x 0.5, class 12.9	5.03 mm²	4,880 N	3,660 N	2.2 Nm	1.6 Nm
M4 x 0.7, class 12.9	8.78 mm²	8,520 N	6,390 N	5.1 Nm	3.8 Nm
M5 x 0.8, class 12.9	14.2 mm²	13,774 N	10,331 N	10.3 Nm	7.7 Nm
M6 x 1.0, class 12.9	20.1 mm²	19,497 N	14,623 N	17.5 Nm	13.2 Nm
M8 x 1.25, class 12.9	36.6 mm²	35,502 N	26,627 N	42.6 Nm	32.0 Nm
M10 x 1.5, class 12.9	58.0 mm²	56,260 N	42,195 N	84.4 Nm	63.3 Nm
M12 x 1.75, class 12.9	84.3 mm²	81,771 N	61,328 N	147 Nm	110 Nm
1/4-20 UNC, Grade 5	0.0318 sq in	2,700 lbf	2,029 lbf	101 in-lb	76 in-lb
5/16-18 UNC, Grade 5	0.0524 sq in	4,450 lbf	3,342 lbf	209 in-lb	157 in-lb
3/8-16 UNC, Grade 5	0.0775 sq in	6,590 lbf	4,940 lbf	31 ft-lb	23 ft-lb
7/16-14 UNC, Grade 5	0.1063 sq in	9,040 lbf	6,777 lbf	49 ft-lb	37 ft-lb
1/2-13 UNC, Grade 5	0.1419 sq in	12,100 lbf	9,046 lbf	75 ft-lb	57 ft-lb
9/16-12 UNC, Grade 5	0.1819 sq in	15,500 lbf	11,599 lbf	109 ft-lb	82 ft-lb
5/8-11 UNC, Grade 5	0.2260 sq in	19,200 lbf	14,408 lbf	150 ft-lb	113 ft-lb
3/4-10 UNC, Grade 5	0.3345 sq in	28,400 lbf	21,322 lbf	267 ft-lb	200 ft-lb
1/4-20 UNC, Grade 8	0.0318 sq in	3,820 lbf	2,864 lbf	143 in-lb	107 in-lb
5/16-18 UNC, Grade 8	0.0524 sq in	6,290 lbf	4,719 lbf	295 in-lb	221 in-lb
3/8-16 UNC, Grade 8	0.0775 sq in	9,300 lbf	6,974 lbf	44 ft-lb	33 ft-lb
7/16-14 UNC, Grade 8	0.1063 sq in	12,800 lbf	9,568 lbf	70 ft-lb	52 ft-lb
1/2-13 UNC, Grade 8	0.1419 sq in	17,000 lbf	12,771 lbf	106 ft-lb	80 ft-lb
9/16-12 UNC, Grade 8	0.1819 sq in	21,800 lbf	16,375 lbf	154 ft-lb	115 ft-lb
5/8-11 UNC, Grade 8	0.2260 sq in	27,100 lbf	20,340 lbf	212 ft-lb	159 ft-lb
3/4-10 UNC, Grade 8	0.3345 sq in	40,100 lbf	30,101 lbf	376 ft-lb	282 ft-lb

Frequently asked questions

How much torque for an M10 class 10.9 bolt?

About 72 Nm dry or 54 Nm lubricated, targeting a clamp load of 36 kN, which is 75% of the 48 kN proof load. Use the dry value only if the threads go together plain and dry, because lubricated threads reach the same clamp at the lower torque.

Should I use the dry or the lubricated torque value?

Whichever matches the real assembly condition. The table assumes K = 0.20 for dry plain threads and 0.15 lubricated, so a lubricated fastener reaches target clamp load at about 25% less torque. Applying a dry value to a waxed or oiled bolt over-tightens it.

What is the difference between proof load and clamp load?

Proof load is the highest axial load the bolt carries without permanent set, equal to proof stress times tensile stress area from the standard. Clamp load is the preload you actually target at assembly; this table targets 75% of proof load, which leaves margin below yield.

Why do torque charts give different numbers for the same bolt?

Charts differ in the clamp target and the friction assumption. This table uses 75% of proof load with K = 0.20 dry; some references publish 85% of proof load on lightly oiled threads, and VDI 2230-based tables target about 90% of yield. A 10 to 20% spread between reputable charts is method, not error.

Can I use these values for stainless steel bolts?

No. ISO 898-1 covers bolts made of carbon steel and alloy steel, and SAE J429 covers the same materials, so these values assume those strengths. Stainless fasteners carry their own class numbers (50, 70, 80) with different strength levels, so use data specific to the stainless grade.

What K value should I use if the condition is not specified?

K = 0.20 for plain, dry threads is the commonly assumed default, and NASA RP-1228 calls 0.15 a realistic value for steel on steel. Zinc-plated (waxed) threads vary 0.12 to 0.16 per Fastenal charts. For a critical joint, measure the torque-tension relationship on the actual parts rather than assuming K.

Are these torque values maximums or targets?

They are assembly targets that aim for 75% of proof load, not failure limits. The bolt is proof-tested above these clamp loads, but friction scatter means achieved preload can sit far from target even at the specified torque, so treat the values as guidance and verify critical joints against the governing standard.

Do metric and inch grades cross over?

Approximately: class 8.8 (800 MPa minimum tensile) sits near Grade 5 (827 MPa), and class 10.9 (1040 MPa) near Grade 8 (1034 MPa). The threads still do not interchange, so match hardware within one system and never substitute across systems.

Sources

Last reviewed: 2026-08-16