Gearboxes

A gearbox trades speed for torque. Nothing else it does matters as much as that, and every number on its nameplate is an expression of it. Put 1,750 rpm and a modest twist in one end of a 20:1 reducer and you get 87.5 rpm and roughly twenty times the twist out the other, minus what the gear teeth lose to friction.

Why it matters on the job

A gearbox is one of the few machines that will not tell you it is dying until it is nearly dead. It is closed and quiet compared with a fan or a compressor, and by the time an operator reports noise the damage inside is usually done. So gearbox maintenance is mostly indirect: you diagnose through the oil, through the temperature of the case, and through the vibration spectrum, and you open it only when you already know what you expect to find.

Gear types and what each one costs you

Spur gears have teeth cut straight across the face. They are simple, efficient, and produce no thrust load. They are also noisy, because each tooth pair engages all at once.

Helical gears have teeth cut at an angle, so each tooth rolls into mesh gradually. That makes them quieter and stronger for the same size, and it produces an axial thrust load that the bearings have to carry. Most industrial reducers are helical.

Bevel gears carry power between shafts whose centerlines intersect, usually at 90 degrees. Spiral bevel is the quieter version, with the same thrust penalty as helical.

Worm gears put a screw thread against a wheel to turn a right angle with a large ratio in one stage. The contact is mostly sliding rather than rolling, so a worm set runs hot and is far less efficient than a comparable helical train, and it usually needs a different (often compounded) oil. Some worm sets are self-locking, which means the output cannot drive the input. That is a useful safety property, and it is a property of the specific set, never something to assume.

Ratio is teeth, and stages multiply

Ratio comes from tooth counts, not from diameters you measure with a tape. A pinion of 17 teeth driving a gear of 68 teeth is 68 ÷ 17 = 4:1, and that is exact.

Stages multiply. Two stages of 4:1 and 5:1 give 20:1 overall. That multiplication is why a compact reducer can achieve ratios that no single gear pair could, and why the intermediate shaft in a two-stage box turns at a speed that appears nowhere on the nameplate.

Worked example: a two-stage helical reducer

A 10 hp motor at 1,750 rpm drives a two-stage helical reducer. Stage one is a 17-tooth pinion into a 68-tooth gear. Stage two is a 15-tooth pinion into a 75-tooth gear.

Ratio. Stage one is 68 ÷ 17 = 4.00. Stage two is 75 ÷ 15 = 5.00. Overall 4.00 × 5.00 = 20:1.

Speeds. Intermediate shaft: 1,750 ÷ 4 = 437.5 rpm. Output shaft: 437.5 ÷ 5 = 87.5 rpm.

Input torque. Torque in pound-feet is 5,252 × hp ÷ rpm, so 5,252 × 10 ÷ 1,750 = 30.01 lb-ft.

Output torque. Ideal output is input torque × ratio = 30.01 × 20 = 600.23 lb-ft. A helical stage is efficient but not free; take 98% per stage for this example (the manufacturer’s rating governs on a real box), so 0.98 × 0.98 = 0.9604 overall. Output = 600.23 × 0.9604 = 576.5 lb-ft.

Twenty times slower, close to twenty times the torque, and the missing 4% left as heat in the case

Check it the other way to be sure. Output power is 10 hp × 0.9604 = 9.604 hp, and 5,252 × 9.604 ÷ 87.5 = 576.5 lb-ft. The two routes agree, and that cross-check is a habit worth building.

Now notice what the losses became. Roughly 4% of 10 hp is 0.4 hp of heat, going into the oil and out through the case. That is why a gearbox case runs warm, why the oil level and the breather matter, and why a hot case is a real symptom rather than a nuisance.

The oil is the instrument

Most of what you will ever learn about a running gearbox comes from its oil.

• Level is read on a stopped, cooled box, at the mark, not by eye through a plug hole. Overfilling churns oil and raises temperature; underfilling starves the top-side bearings.
• Grade is on the nameplate as an ISO viscosity grade or an AGMA lubricant number. Wrong viscosity is a slow failure, and a worm box that gets a standard gear oil is a fast one.
• Milky or cloudy oil is water. Find where it came from (a failed seal, a washdown, or condensation through a breather) before you refill.
• Metal on the magnetic drain plug is normal in fine fuzz and abnormal in flakes. Flakes are pieces of a tooth surface.
• Oil analysis turns all of the above into trended numbers, and it is a high-value predictive task on a gearbox.

The breather deserves its own line. A gearbox heats up, the air inside expands, and it has to go somewhere. Plug the breather (with paint, with dirt, or with a well-meaning plug) and the box pressurizes and pushes oil past the seals. Then the box runs low, and the seals get blamed.

Listening to the mesh

Gear faults appear in vibration at gear mesh frequency, which is the number of teeth on a gear multiplied by that shaft’s speed in revolutions per second.

For the input pinion above: 1,750 rpm ÷ 60 = 29.17 rev/s, and 17 teeth × 29.17 = 495.8 Hz. A rise at that frequency, with sidebands spaced at shaft speed, points at the input stage rather than the output stage. That is how a vibration analyst tells you which stage to open before anybody removes a bolt.

Opening one up

• Match-mark everything before it comes apart: case halves, bearing caps, gear-to-shaft position.
• Keep shim packs with their own caps. Shims set bearing preload and gear position, and a swapped pack changes both.
• Record backlash before disassembly with a dial indicator against a tooth flank, holding the mating gear still. It is your reference for reassembly and a measurement of wear.
• Photograph the tooth contact pattern before you clean anything. The wear pattern on the flank tells you whether the mesh was correct.

Where it bites

• A gearbox rarely fails on its own. Overhung load from an overtight belt drive, a misaligned coupling, and shock loading from the driven machine cause more gearbox failures than gear-tooth fatigue does.
• Service factor is not a safety margin you may spend. It is the allowance already made for the character of the load, and a box selected on nameplate horsepower alone is undersized for a shock-loaded application.
• Do not top up with whatever is on the shelf. Mixing incompatible additive packages can drop a lubricant out of solution, and the box then runs on something neither product was.
• The intermediate shaft speed matters for analysis. Vibration data collected against nameplate input speed alone will not locate a second-stage fault.
• A worm box is not a helical box with a corner in it. Its efficiency, its heat, and its oil are all different, and the ratio you can get in one stage tempts people into ignoring that.

Exam relevance

Expect tooth-count ratio arithmetic, multi-stage multiplication, the 5,252 torque relationship, and identification questions on the four gear types and their thrust behavior. The CMRT’s hands-on domains and the NCCER Millwright mechanical levels both treat reducers as core equipment. The reliable discriminator question is the oil one: given a symptom (milky oil, oil at the seals, a hot case), name the likely cause. Answer it with the breather and the level before you reach for the gears.