Most industrial gearbox repairs trace back to seven failure modes, and almost every one of them leaves evidence that points at its own root cause. A disciplined gearbox failure analysis turns a broken drive into a better specification for the next one - which is why the mode matters more than the invoice. This guide walks through the seven modes in the order they are usually met, the physical evidence each leaves behind, and the root causes that hide under the symptoms.

Why Failure Analysis Pays for Itself

A failed gearbox is expensive twice: once when it stops the line, and once when the wrong lesson is drawn from it. The expensive habit is to order a like-for-like unit, refit it, and send the failed case to the scrap bin. When that happens, the next failure arrives on the same schedule, because the root cause - a duty the rating never covered, an oil programme nobody owned, a foundation that was never level - is still in place.

Analysis changes the economics. An hour spent reading the bearings, the tooth flanks and the oil tells you whether the unit was undersized, misapplied, starved of lubrication or simply old. Each answer leads to a different action: a larger frame, a better breather, a sampling programme, or nothing at all beyond the scheduled renewal. The seven modes below cover the overwhelming majority of what a maintenance crew will ever open a case to find.

Bearing Failure - The Most Common Mode

Bearings account for more gearbox repairs than any other component, and they are also the most misread. A gearbox bearing failure is rarely a bearing problem; it is an electrical, lubrication or load-distribution problem that happened to surface in a bearing. The evidence sorts it. Fluting patterns - regular ripples around the raceway - point at shaft currents, usually from a variable frequency drive without insulated bearings or a shaft earthing brush. Discolouration blue or grey raceways point at heat, from oil starvation or a badly specified viscosity. Spalling concentrated on one side of a raceway points at overload or misalignment pushing the load to one flank.

The root-cause discipline is to ask what the bearing experienced rather than what it is. A bearing renewed without answering that question runs out its life on the same clock as its predecessor. Where the analysis points at lubrication, the fix is the oil programme; where it points at currents, the fix is earthing or insulation; where it points at load, the fix belongs in the selection calculation for the renewal.

Tooth Damage - Pitting, Scuffing and Breakage

Gear tooth damage arrives in three grades, and the open-up tells you which one you have. Pitting is surface fatigue: contact stress repeated millions of times until the flank sheds small particles. Destructive pitting - pits that grow, join and hollow the flank - means the contact stress exceeded the material's endurance for the duty, usually because the service factor was too small for the real load spectrum, the surface hardness sat below what the duty required, or an alignment error concentrated the load at one end of the tooth. Not all pitting is a verdict, though: initial pitting on a new unit that stops growing after the running-in period is the mesh correcting itself, and monitoring it through a second inspection interval is the correct response.

Scuffing and breakage are the grades that stop a line without warning. Scuffing - radial tears in the flank where the oil film collapsed and metal welded and tore - follows high speed, high load or a lubricant that lost viscosity to heat. Breakage at the tooth root follows overload, and the fracture face tells the story: a single clean fracture means one event far beyond the rating, while a beach-marked face means a fatigue crack that grew over weeks from a root stressed just past its endurance. Neither grade forgives a loose specification; the duty description that omits the worst surge, the breakaway torque or the starts per hour is the common ancestor of both.

Seal Leakage and the Loss That Follows

A leaking seal looks like a housekeeping problem and behaves like a bearing-killer-in-waiting. Oil loss lowers the sump level, a lower level starves the upper bearings, and the failure that ends up on the report is a bearing - mode 1 - with the seal as its unrecorded root cause. Seal analysis therefore asks two questions: why the lip failed, and why the case pressurised. A hardened, cracked lip points at heat; a worn, grooved shaft surface at the lip points at the shaft finish or a slight misalignment; oil escaping at multiple points points at a blocked breather building pressure inside the case.

The pattern across a plant matters too. When every unit of a family weeps at the same corner, the cause is usually systemic - a breather spec, an ambient that exceeds the seal's temperature band, or a washdown practice the seal was never chosen for. Renewing one seal at a time in that situation is a subscription, not a fix.

Oil Contamination and Degradation

Gearbox oil contamination is the mode that never presents itself directly - there is no broken part to photograph, only a sample that tells you what is coming. Water in the sample names its own sources: a tired breather breathing humid air, a washdown hose aimed at a plug seal, condensation from a cold start regime. Wear metals carry a direction: iron rising points at gears or bearings in wear, and the ratio of elements hints at which. Silicon arriving with iron means dust is finding a path in, and the analysis has just linked two modes - contamination now, abrasion later.

Degradation is the quieter half. Oil that has oxidised past its viscosity band, or shed its additive package, protects nothing regardless of how clean it is. This is why the sampling programme, not the calendar, should own the change interval - and why a plant that samples on schedule almost never meets the advanced stages of any mode in this guide.

Misalignment, Soft Foot and Fretting

The last mode is structural, and it is the one that most often gets recorded as somebody else's failure. Misalignment between gearbox and driven machine loads bearings and seals sideways, and shows up as edge loading on tooth flanks and one-sided bearing wear. Soft foot - a case pulled out of square by an unleveled pad - distorts the internal clearances the bearings were designed around. Fretting corrosion at the bore or the feet, the orange powder that looks like rust, is the record of micro-movement that the load path was never meant to make.

The evidence is unglamorous but conclusive, and the fixes are close to free: a level check of the pads, a laser alignment of the train, a torque arm fitted at its designed angle rather than wherever it reached. A drive renewed without these checks inherits the failure that ended its predecessor.

The Seven Modes at a Glance - and What Each Buys You

The purpose of gearbox failure analysis is not the post-mortem; it is the purchase order that follows it. Each mode converts into a line of specification: bearing modes into insulated bearings and an oil programme, tooth modes into an honest service factor, seal modes into a breather and seal spec that match the ambient, structural modes into a levelled base and an aligned train. A supplier who receives a duty description carrying those lines can quote a unit that will not repeat the history - and the pages for our MC3RLSF08 bevel-helical gear unit show the kind of duty detail that turns a renewal into an improvement.

The fastest way to check a renewal against its duty is the same division every selection starts with: motor speed over target output speed for the ratio, then the service factor applied to the torque for the frame. You can work the numbers yourself in the online selection system, or send us the failed unit's history and let our engineers propose the specification that breaks the cycle.

FAQ

Which gearbox failure mode is the most common?
Bearing damage tops the list in most repair statistics, but it is better described as the most common symptom - the majority of gearbox bearing failure cases trace back to lubrication, shaft currents or load distribution rather than the bearing itself. Treating the raceway evidence as the beginning of the analysis rather than the end is what separates a repair from a repeat.

Can a gearbox be repaired after tooth pitting is found?
It depends on whether the pitting is arrested or progressive. Initial pitting that stops growing after running-in can be monitored through normal inspection intervals. Destructive pitting that is deepening, joining or hollowing the flank means the contact stress exceeds the material's endurance for the duty, and the realistic options are derating the drive, correcting the alignment, or renewing with a unit selected against the true load spectrum.

How does oil analysis detect gearbox problems early?
An oil sample reports the sump's health in three directions: contamination (water, dust silicon), wear (iron, bronze and their trends) and degradation (viscosity and additive levels). Taken at a fixed interval and tracked across samples, the trends surface most of the seven failure modes months before a symptom appears at the machine - which is why the sample, not the calendar, should own the oil change.

Is it worth analysing a gearbox that is already scrap?
Yes - that unit is the cheapest consultant the plant will ever hire. The raceway patterns, flank condition, fracture faces and remaining oil together name the root cause, and the specification written from that evidence is what stops the history repeating. The alternative - a like-for-like renewal and a scrapped case - is how plants buy the same failure twice.

Read the Evidence Before You Order the Repair

Seven modes, seven sets of evidence, seven root causes - and one habit that covers most of them: specify the duty as it really runs, and let the sample programme keep the score between overhauls. If a failed drive is on your bench right now, send us what the parts are showing and the duty the unit was supposed to see; we will help you read the evidence and quote the renewal that breaks the pattern. Our online selection system cross-checks the ratio and service factor in minutes.