Keyway Shear: Answered by the Shop Floor
A 25 mm steel shaft driving a conveyor pulley. Standard 8×7 mm square key. One week into production, the key sheared clean across. Torque was only 30 N·m — nothing dramatic. Here’s how that conversation went with the shop foreman, and what the numbers said.
“We used a standard key. Why did it break?”
Because “standard” means it fits the shaft diameter, not that it survives the torque. The shear stress on a square key is:
τ = 2T / (d · b · L)
Plug in the numbers: T = 30,000 N·mm (30 N·m), d = 25 mm shaft, b = 8 mm key width, L = 20 mm engaged length. τ = 60,000 / (25 × 8 × 20) = 150 N/mm². Mild steel key allows about 80–100 N/mm² with a safety factor. You were 50% over the limit. The key didn’t fail randomly — the math predicted it.
“Would a bigger key have saved it?”
Yes, but it’s not just width. Upsizing to a 10×8 mm key (b = 10 mm, L = 25 mm): τ = 60,000 / (25 × 10 × 25) = 96 N/mm². That’s in the acceptable range. Alternatively, two keys 180° apart share the load and double the effective area. A spline (integral teeth) handles even more torque but costs more and needs a spline bore in the hub.
“What about crushing? Isn’t that the weak link?”
For a square key, crushing rarely governs. The bearing stress on the key edge:
σ = 2T / (d · 0.5h · L)
With h = 7 mm: σ = 60,000 / (25 × 3.5 × 20) = 34 N/mm². Steel allows ~150 N/mm² in compression. Crushing is fine. Shear is always the failure mode on a standard square key.
“What about the shaft itself? Is it strong enough?”
The shaft torsion check:
τ_shaft = 16T / (π · d³)
τ = 16 × 30,000 / (π × 15,625) = 9.8 N/mm². Steel allows 40–50 N/mm² in shear. The shaft was fine. The key was the weak link — and that’s normal. Keys are deliberately the weak component so they fail before the shaft or the expensive hub. When you oversize the key, make sure you’re not pushing the failure mode into something harder to replace.
“When does the keyway itself cause trouble?”
Under fatigue. The sharp corners of a milled keyway are stress raisers. A reversing torque (like a servo axis that speeds up, stops, and reverses every cycle) loads those corners in bending. Crack initiation at the keyway root is a classic fatigue failure mode. For reversing service:
- Use a keyway with generous corner radii (not a sharp square).
- Ground the keyway rather than mill it (smoother surface = lower stress concentration).
- Add 20–30% to the safety factor on shear.
For a conveyor running at constant torque in one direction, fatigue isn’t on the table. The key just sees steady shear.
“What key types are there, and when?”
| Type | Torque capacity | Use case |
|---|---|---|
| Square parallel key | Standard | Most shafts, general drive |
| Woodruff (half-moon) | Low (small shafts) | Ø20 mm and under, self-aligning |
| Spline (integral teeth) | High | High torque, frequent shifting |
| Two keys at 180° | ~2× single key | High torque, no room for spline |
| Tapered key | Medium, axial lock | Shafts without hub screws |
“How did we fix it?”
We upsized the key from 8×7 to 10×8, matched the keyway in the pulley hub, and specified the key material as 1045 steel (not cold-rolled mild). We also added a 1 mm radius at the keyway corners on the next revision. The replacement ran for 18 months without issue. The shop foreman’s takeaway: “Always size the key, don’t just pull one off the shelf.”
The one formula to remember: τ = 2T/(d·b·L). If that number is over 100 N/mm² for a steel key, upsizing is non-negotiable. The shaft almost never fails in torsion before the key does — that’s by design.
