When the layout will not accept a drive arriving on the driven shaft's own axis.

An in-line reducer plus its motor forms one long straight assembly, and that whole length has to be clear behind the driven shaft. Very often it is not - the shaft end is close to a wall, a structural member, a guard, or the next machine in the line.

A parallel-shaft unit folds the train sideways. The motor sits beside, above or below the driven shaft rather than behind it, and the drive occupies space that is usually available even when axial space is not.

Because both types use helical gearing, the efficiency and noise are broadly comparable, so this is a packaging decision rather than a performance one. Where the axial space genuinely exists, in-line is simpler and marginally shorter in the gear train; where it does not, parallel-shaft is the answer without a technical penalty.

Yes, and it needs to be, because a parallel-shaft unit is not symmetrical about its axis.

Unlike an in-line reducer, which can generally be rotated about its own axis to suit, a parallel-shaft unit has a definite geometry: the output is offset in a particular direction relative to the input and to the mounting feet. Which side the output projects from, and whether it is a single or double-ended shaft, are ordering options.

The mounting position matters for the same reason. Foot-mounted units are supplied for a specific orientation, and that orientation determines the oil level and the breather position. Fitting a unit in an orientation it was not supplied for risks starving bearings or flooding a seal.

Work out the physical arrangement on the machine before ordering: which way the output shaft points, which face the feet bolt to, and where the motor sits.

Yes, and they are common on parallel-shaft units because the offset geometry accommodates them naturally.

A hollow output shaft lets the reducer slide directly onto the driven machine's shaft rather than driving it through a coupling. That removes the coupling, removes the alignment task, and removes the separate base the reducer would otherwise need - the unit is carried by the shaft itself and prevented from rotating by a torque arm.

The hollow bore is supplied with a keyway, a shrink disc or a taper bush depending on the torque and on whether the connection must be removable.

It is worth noting that this arrangement is common enough to be sold as its own type, the shaft-mounted reducer, which is optimised around it. Where the shaft-mounted arrangement is the intention from the start, that product is usually the better fit; a hollow-shaft parallel unit suits cases where the same gearbox range is wanted in both configurations.

Through the feet or flange on a conventionally mounted unit, and through a torque arm on a hollow-shaft one.

Every reducer produces a reaction: the torque it delivers to the output is matched by an equal and opposite torque trying to rotate the housing. On a foot-mounted unit that reaction goes into the bedplate through the holding-down bolts, which is why the base must be rigid and flat and the bolts correctly sized.

On a hollow-shaft unit there is no base. The housing would simply rotate around the driven shaft, so a torque arm links it to a fixed point on the machine or structure. The arm should be mounted so it works in tension where possible, and it needs a resilient bush at the anchor end to absorb shock rather than transmitting it into the frame.

A torque arm that is too short, badly angled or rigidly anchored is a frequent source of vibration and of cracked mounting brackets.

A gearbox is more compact, more efficient and needs less attention; a belt drive is cheaper, more forgiving and easier to change ratio.

A parallel-shaft reducer transmits through gears in an oil bath. Efficiency is high, there is nothing to tension or replace routinely, ratios are exact, and the unit occupies little space. It also holds its ratio precisely, which matters where speed must be predictable.

A belt drive between two pulleys achieves an offset drive far more cheaply, absorbs shock through belt slip, and lets the ratio be changed by swapping a pulley. But it needs guarding, tensioning and periodic belt replacement, loses efficiency as belts wear, and takes up more space for the same torque.

For continuous industrial duty the gearbox usually wins on running cost and maintenance. For light duty, frequent ratio changes or deliberate shock absorption, a belt drive remains a sound choice.

One chosen from the driven machine's actual behaviour, not from a general figure - and it is the single most consequential number in the selection.

The service factor multiplies the calculated torque before the reducer is selected, and it accounts for how the load really behaves: hours run per day, starts and stops per hour, whether the load is smooth or shock-loaded, and whether it reverses under load.

A smooth centrifugal pump running eight hours takes a modest factor. A crusher, a reciprocating compressor or a heavily loaded conveyor that starts under full load takes a much higher one, and applying the pump's factor to the crusher will destroy the gearbox in months.

Manufacturers publish tables of factors by machine type and duty. Use them rather than estimating, and where the duty is genuinely unusual, say so when specifying - a supplier can advise, but only if they know what the machine does.

Overload from an under-applied service factor, poor mounting, lubrication faults and shaft misalignment - roughly in that order.

Under-sizing is the commonest. A reducer selected on nominal torque for a shock-loaded duty is loaded well beyond its design intent at every start, and gear teeth or bearings fail long before their rated life.

Mounting comes next. A base that is not flat distorts the housing as the bolts are tightened, which misaligns the gears inside it. A foot-mounted unit on a flexible bedplate suffers the same.

Lubrication faults are usually level or breather related rather than oil quality: wrong level for the mounting position, or a blocked breather pushing oil past the seals and being read as seal failure.

Misalignment at the coupling loads the output bearing radially and shortens its life. Align properly, and use a coupling that accommodates the residual misalignment rather than transmitting it.