I watched a mechanic torque an M8 bolt to 25 N·m with a wrench. “That’s standard,” he said. I asked him what preload that gave him. He shrugged. I don’t blame him — torque wrenches don’t tell you preload, and most people don’t know the formula. But in a clamping fixture, if the preload is wrong, the joint slips or the bolt breaks. Here’s how I calculate it.
The relationship between torque and preload
The torque you apply goes into three things: overcoming thread friction, overcoming under-head friction, and stretching the bolt (preload). The simplified relationship is:
T = K · F · d
Where:
- T — torque in N·m
- K — nut factor (dimensionless). For dry steel, K ≈ 0.20. For zinc-plated, K ≈ 0.18. For lubricated, K ≈ 0.12.
- F — preload in Newtons
- d — nominal bolt diameter in meters
This is the rough formula. It’s good enough for fixture design. The nut factor K is the wildcard — it varies with lubrication, plating, and how new the bolt is. That’s why torque wrenches are ±25% on a good day.
The preload I aim for
The standard rule: preload the bolt to 70% of its yield strength. For an 8.8-grade bolt, yield strength is 640 MPa. The stress area of an M8 bolt is 39.7 mm². Yield load = 640 × 39.7 = 25,408 N. 70% of that is 17,786 N. That’s the target preload.
Now back-calculate the torque. For dry steel (K = 0.20), d = 0.008 m:
T = 0.20 × 17,786 × 0.008 = 28.5 N·m.
That matches the “standard” 25-30 N·m for M8 8.8. The mechanic wasn’t wrong — he just didn’t know why.
Reference table I use
For 8.8-grade bolts, dry (K=0.20), 70% yield preload:
| Bolt | Stress area (mm²) | Yield load (N) | Target preload (N) | Torque (N·m) |
|---|---|---|---|---|
| M6 | 20.1 | 12,864 | 9,005 | 10.8 |
| M8 | 39.7 | 25,408 | 17,786 | 28.5 |
| M10 | 58.0 | 37,120 | 25,984 | 52.0 |
| M12 | 84.3 | 53,952 | 37,766 | 90.6 |
| M16 | 157 | 100,480 | 70,336 | 225 |
| M20 | 245 | 156,800 | 109,760 | 439 |
Three things that break this table
1. Lubrication changes everything. If the bolt is oiled or has anti-seize compound, K drops to 0.12. The same torque gives 67% more preload. You’ll yield the bolt. I always torque dry for fixture bolts, or I explicitly use the lubricated K factor. The “standard” torque values assume dry, unplated steel. If you zinc-plate or oil it, halve the torque.
2. Clamped material matters. If you’re clamping aluminum (soft), the bolt head sinks into the material under torque. The preload drops. I use hardened washers under the head and nut for soft materials. Without the washer, you can torque to 30 N·m and get half the preload — the bolt is loose but the torque wrench says it’s right.
3. Repeated torque loosens. After 3-4 cycles of torque and loosen, the bolt’s friction changes (galling on the threads). The same torque gives different preload. For fixture bolts that get removed often, I use new bolts or clean and re-lubricate the threads.
Joint separation check
Once I know the preload, I check that the joint doesn’t separate under operating load. The rule: preload must exceed the external load by a safety factor. For a tension joint (like a cylinder end bracket), I keep the external separating load under 50% of preload. If the cylinder pulls with 2,000 N, I need preload of at least 4,000 N. The M8 at 17,800 N is way over that — fine. But if I’d used M6 (9,000 N), it’s still fine. The joint doesn’t separate.
For shear joints (like a bracket holding a rail), the bolts don’t see tension — they see shear. The preload creates friction that holds the joint. The friction force is μ × preload. For μ = 0.15 (steel on steel), an M8 preloaded to 17,800 N gives friction of 2,670 N. If the shear load is 1,500 N, the joint holds. If it’s 3,000 N, it slips. I either use two bolts or go to M10.
When torque isn’t enough
For critical joints (high-cycle, safety-related), I use the turn-of-nut method or a torque + angle approach. Torque to snug (say 10 N·m), then turn an additional 90 degrees. This gives more consistent preload because it measures elongation, not friction. But for 90% of fixture and machine frame bolts, the torque table above is good enough. I don’t overthink it.
The number I check before I specify a bolt: target preload at 70% yield, back-calculated torque with the correct K factor for the surface condition. If the clamped material is soft, add a hardened washer. If the joint separates under load, I either go up a bolt size or add another bolt. The torque wrench doesn’t lie — it just doesn’t tell you the whole story.
