A conveyor drive sheared its coupling hub bolts three times in a year. The drive was a 2.2 kW motor through a worm reducer to a belt conveyor carrying packaged goods. The jams came from product wedging under the belt. Each jam snapped the bolts and stopped the line for forty minutes. The maintenance team replaced the bolts with hardened ones. The next jam bent the coupling and bent the reducer’s output shaft. The hardened bolts moved the damage from the cheap part to the expensive one.
The problem with a shear pin
A shear pin is the classic overload protection — a small pin sized to break at a set torque. It works, once. After it breaks, the drive is down until someone finds the right pin, drills out the old one, and taps in a new one. If the maintenance crew can’t find the exact pin, they use whatever fits. A slightly bigger pin doesn’t shear at the set torque — it transmits the overload to the next component, which is exactly what happened on the conveyor. The hardened bolts were a bigger shear pin that moved the failure to the reducer.
Shear pins are also imprecise. The shear torque scatter is typically ±15%. A pin set at 50 Nm might shear at 43 or 57 Nm. If the process torque occasionally peaks at 45 Nm, the pin trips on normal operation. The line stops for nothing.
What a torque limiter does differently
A torque limiter is a coupling that slips at a set torque and re-engages when the overload clears. Two common types. The friction type — a clutch plate pressed by springs — slips continuously while overloaded, then grabs again. Its trip torque accuracy is about ±20%, and it’s best for constant-duty protection where occasional slip is acceptable. The ball-detent type — spring-loaded balls seated in detents — releases completely at the trip point, disconnects the drive, and re-engages when the torque drops. Trip accuracy is about ±5%, and it gives a clean disconnect with no rubbing wear.
For the conveyor, the ball-detent type was right. Set at 60 Nm (about 30% above the normal 45 Nm peak), it releases on a jam, the motor keeps running (the limiter is on the low-speed side, so the motor just spins free), and the line operator clears the jam and pushes the limiter back in. Downtime per jam dropped from forty minutes to four.
Where to mount the limiter
This is the detail people get wrong. The limiter protects whatever is downstream of it. Mounted on the motor side (high-speed side), it protects the reducer and everything after. Mounted on the load side (low-speed side, after the reducer), it protects only the belt and the components after the limiter — the reducer itself is exposed. For a conveyor where the reducer is the expensive part, the limiter goes on the motor side, before the reducer.
The torque setting also needs to account for the ratio. A limiter on the motor side of a 20:1 reducer is set at 1/20 of the load torque. The 45 Nm process peak at the load is 2.25 Nm at the motor. The 60 Nm trip setting at the load is 3 Nm at the motor. The limiter is selected for that range — a small unit, cheaper than a load-side limiter that must handle the full output torque.
Friction vs ball-detent, the practical decision
| Duty | Type | Why |
|---|---|---|
| Frequent, brief overloads, constant duty | Friction | Slips through, self-resets, wears slowly |
| Jams and blocks, clean disconnect wanted | Ball-detent | ±5% accuracy, full release, manual reset |
| High-speed protection, low inertia | Ball-detent | Releases fast, no drag |
| Where reset must be automatic | Friction | No operator action needed |
The friction type has a hidden cost: it generates heat when it slips. A friction limiter slipping for a minute at full motor power turns the slip energy into heat at the friction faces. If jams are frequent, the limiter gets hot and the friction coefficient drifts. The set torque drifts with it. The ball-detent type disconnects completely — no heat, no drift, no wear while tripped.
The setting procedure
Set the limiter with a torque wrench, not by feel. With the drive disconnected, turn the output (or input) by hand using a torque wrench on the shaft and read the torque at which the limiter trips. Set it at 1.3 to 1.5x the maximum normal operating torque. Then test it under real load — jam the conveyor by hand at the belt and confirm the limiter trips before anything bends. If it trips on normal peaks, raise it 10% and test again. The setting is a compromise: high enough not to trip on normal peaks, low enough to protect the weakest component. The weakest component is the one you’re willing to replace cheaply.
Shear pins protect once, then cost you the line. Hardened bolts protect nothing — they just move the damage downstream. A ball-detent torque limiter set at 1.3x normal torque, mounted before the reducer, turns a forty-minute jam stop into a four-minute reset. The bent reducer was the price of protecting a coupling with hardened bolts.
