By letting the load tighten the spring onto the hubs, so the grip increases with the torque applied.

A helical spring is fitted closely around two hubs sitting end to end - one on the input, one on the output. When the input turns in the direction that winds the spring tighter, the coils constrict onto both hubs, grip them, and drive the output. The harder the load pulls, the tighter the spring wraps and the more firmly it grips.

That self-energising action is what gives the clutch high torque capacity from a small, simple mechanism - no external clamping force is needed at all.

Turn the input the other way and the spring is unwound rather than tightened, so it opens slightly and slips.

Engagement is controlled by restraining the spring's outer end. A control collar held stationary keeps the spring open and the clutch disengaged; releasing the collar lets the spring wrap down almost instantly.

The control collar can be given a single stop, so the clutch disengages itself at exactly the same angular position every cycle.

When the pawl releases the collar, the spring wraps and the output turns with the input. The collar rotates with it. After one full revolution the stop on the collar comes back round to the pawl, which has re-extended, and is caught - which unwraps the spring and disengages the clutch at precisely that point.

The result is exactly one output revolution per command signal, stopping in the same position each time with no timing control, no encoder and no braking required.

Collars with two, four or more stops give correspondingly precise part-revolution indexing.

That repeatability, from a purely mechanical arrangement, is why wrap-spring clutches persist in feeders, cut-to-length machinery, indexing tables, print and stamping cycles, and packaging equipment despite the availability of servo alternatives.

Yes - inherently so, and it is fundamental to how they work rather than a limitation that can be designed around.

The spring grips when torque winds it tighter and releases when torque unwinds it. Reverse the direction of drive and the mechanism does the opposite of what is wanted.

That directionality is exploited deliberately in some products: a wrap-spring arrangement makes an effective one-way clutch or backstop, driving in one direction and overrunning in the other.

For a clutch application it means the direction of rotation must be specified when ordering, and it must not change. A drive that reverses needs a different clutch type.

It also means the clutch provides no braking or holding in the reverse direction - the output can freewheel backwards when disengaged, which on an inclined or gravity-loaded application is something the design must account for separately.

Much lower, because there is no sustained slipping - the spring wraps and grips in a fraction of a turn.

A friction clutch dissipates energy every time it engages, slipping while the driven side accelerates, and that slip is what wears the linings. A wrap-spring clutch engages almost instantaneously: the spring closes onto the hubs and grips, with very little relative motion between the surfaces.

So there is very little friction wear and very little heat, and cycle lives in the millions of engagements are normal.

The corollary is that the shock of engagement goes into the drive train instead of being absorbed as heat. There is no soft start - the output is picked up abruptly - so the driven machine and the input drive must be able to take that impulse. On a high-inertia load that can be punishing, which is why wrap-spring clutches suit light, low-inertia, high-cycle indexing rather than heavy drives.

Modest by industrial clutch standards, and set by the spring, the hub diameter and the materials rather than by an applied force.

Because the grip is self-energising, capacity scales with hub diameter and with the spring's section and number of active coils. Units cover the range needed for feeders, indexers, packaging and light machinery comfortably, but they do not reach the torques a pneumatic or hydraulic clutch handles on a press or a mill.

Capacity is also direction- and design-specific and should be taken from the manufacturer's rating rather than estimated.

A useful characteristic is that overload tends to produce slip rather than breakage: exceed the capacity and the spring slips on the hubs, which is noisy and generates heat but is often survivable. That said, repeated slipping polishes the hubs and reduces grip, so it should not be treated as a designed-in overload protection.

By a solenoid operating a pawl that catches or releases the control collar - so the solenoid does almost no work.

The solenoid does not engage the clutch. All it does is move a small pawl into or out of the path of a stop on the control collar. When the pawl is extended it catches the collar, holding the spring unwrapped and the clutch disengaged. Retracting the pawl frees the collar and the spring wraps under its own action.

Because the solenoid only has to move a light pawl against a small load, it can be small, fast and low-powered, while the clutch itself transmits far more torque than that solenoid could ever generate directly.

For single-revolution operation the solenoid is pulsed: energise briefly to release, and the pawl re-extends to catch the collar at the end of the turn. Timing the pulse is the only control requirement, and it is not critical - anything shorter than one revolution works.

Where the requirement is exactly one repeatable revolution, cheaply, reliably and without programming.

A servo can do everything a wrap-spring clutch does and more, but it requires a drive, a motor, an encoder, configuration and a control system. For a machine whose only requirement is 'turn this shaft once per command, stopping in the same place', that is a great deal of apparatus.

A wrap-spring clutch achieves it with a spring, a collar, a pawl and a small solenoid, running from a continuously turning line shaft. There is nothing to program, nothing to lose its home position, and cycle life runs into the millions.

They remain common in feeders, cut-off mechanisms, packaging, print finishing and any machine with a continuously running main drive and intermittent auxiliary motions. Where the motion profile needs to change, or where positioning is arbitrary rather than fixed, a servo is the right answer instead.