It is often self-locking, and it should never be treated as a brake.

Self-locking means the output cannot back-drive the input. In a high-ratio worm the sliding contact angle is such that friction prevents the wheel turning the worm, so a load holds when power is removed. Low-ratio worms are frequently not self-locking at all.

The reason it is not a brake is that it depends entirely on friction, and friction is not dependable. It falls as the gearbox warms and the oil thins, it falls as the gear set wears and beds in, and vibration can allow creep that would not occur in a static test. A unit that holds perfectly on the bench can creep in service.

Static self-locking and dynamic self-locking are also different: a unit that will not start moving under load may still run on when already in motion.

Where a suspended load, a person or a safety function is involved, fit a proper mechanical brake. Use the self-locking property as a convenience, never as the safety measure.

Because the worm slides across the wheel tooth rather than rolling on it, and sliding friction is high.

In a helical or bevel mesh the tooth surfaces roll against each other with only a small sliding component. In a worm mesh the thread of the worm slides continuously along the wheel tooth - that sliding IS the mechanism by which the reduction happens. Friction at that interface converts a significant fraction of the input power directly into heat.

Efficiency also falls as ratio rises: the higher the ratio, the shallower the lead angle, and the more sliding there is per unit of output. A high-ratio worm can lose a third or more of its input.

That heat has to be dissipated through the housing, which is why worm units often have to be selected on thermal capacity rather than torque, and why the same gearbox is rated much lower for continuous duty than for intermittent.

On intermittent duty, where the high ratio and low cost matter and the running hours are too few for efficiency to add up.

Typical cases are gate and barrier drives, damper and valve actuators, hoists and winches used occasionally, positioning and adjustment mechanisms, packaging machinery, and light conveyors. In all of these the drive runs for seconds or minutes at a time, so wasted energy is negligible and heat has time to dissipate between operations.

The self-locking characteristic is frequently what actually decides it - a gate or damper that holds position without a brake is simpler and cheaper.

The reverse case is continuous duty. A worm running sixteen hours a day on a process line wastes enough electricity to pay for a helical-bevel unit within a few years, and runs hot enough to need oversizing. That is precisely where the more efficient geometries earn their extra cost.

Because the two surfaces slide against each other continuously, and a dissimilar, softer material pairing is what makes that survivable.

The worm is hardened steel. If the wheel were also hardened steel, two hard surfaces sliding under load would gall and seize. Bronze is softer, has good bearing properties, and allows a little conformity - the wheel beds in to the worm during running-in and the contact area grows, which spreads the load and reduces stress.

Bronze also holds a lubricant film well under sliding contact, which steel on steel does not.

The consequence is that the wheel is the sacrificial component and wears in preference to the worm, which is deliberate: replacing a wheel is cheaper than replacing a hardened and ground worm shaft. It also means running-in matters - a new worm unit should not be put straight to full load, and manufacturers publish a running-in procedure that is worth following.

On thermal capacity as well as torque, and the thermal rating is frequently the one that governs.

Every worm gearbox has two ratings. The mechanical rating is the torque the gears and bearings can carry. The thermal rating is the power the housing can dissipate as heat without the oil exceeding its temperature limit. Because a worm converts so much input into heat, the thermal rating on continuous duty is often lower than the mechanical one.

Selecting on torque alone therefore produces a unit that is mechanically adequate and runs too hot, which degrades the oil, softens the bronze and shortens life dramatically.

Ambient temperature matters directly, since dissipation depends on the difference between the housing and the surrounding air. A unit adequate in a cool plant room may not be adequate in a hot one or in an enclosure.

Where the thermal rating is limiting, the options are a larger unit, a fan, or moving to a more efficient geometry.

Correct oil, watched temperature, and respect for the running-in period - with oil the dominant factor.

Worm gearing depends on a lubricant that survives sliding contact under high pressure, and the oil specification is not interchangeable with that for helical gearboxes. Synthetic polyglycol oils are widely specified for worm drives because they reduce friction measurably, which both lowers running temperature and raises efficiency. Note that polyglycols are not compatible with mineral oils or with some seal materials, so switching is not a simple drain and refill.

Monitor running temperature and treat a rise as information: it generally means the oil has degraded, the level is wrong, or the load has increased.

Follow the running-in schedule on a new unit. The bronze wheel beds to the worm during those first hours, and loading it hard immediately produces a poor contact pattern that limits capacity for the rest of its life.

Only if it is a low-ratio unit specified as non-self-locking, and it should be confirmed with the manufacturer rather than assumed.

Some applications genuinely need back-driving - handwheel operation of a powered drive, manual release, or a system where the load must be movable when power is off. Low-ratio worm sets with a steeper lead angle will back-drive, though at poor efficiency in that direction.

The difficulty is that self-locking is not a clean yes or no. It varies with ratio, with temperature, with lubricant, and with whether the unit is starting from rest or already moving. A unit that back-drives when warm may not when cold.

If back-driving is a functional requirement, state it when ordering and get a written confirmation for the specific ratio and lubricant. If holding is the requirement, do not rely on self-locking - specify a brake. Designing around an assumption in either direction is where worm drives cause trouble.