Reading Bearing Failure Modes: A Visual Guide

When a bearing comes out of a machine it carries a complete record of how it lived and why it stopped. Naming that record correctly turns maintenance from part-swapping into diagnosis, because the same spall that means the end of a fair life on one shaft means wrong clearance on a second and stray current on a third. The terminology is standardised: the ISO 15243 families of rolling-bearing damage have agreed names, and the discipline of naming the mode before touching a tool is the first half of preventing the next failure.

Fatigue Spalling and What Its Edge Tells You

Material fatigue is the designed-for failure. Under a sufficient number of stress cycles - often on the order of ten million revolutions and more - a crack initiates 0.1 to 0.2 mm below the raceway surface, grows towards the surface, and lifts out a flake of metal, a spall. The contact runs on bearing steel of roughly 58 to 62 HRC, and that hardness is what sets how the flake leaves the surface. Genuine fatigue spalls have clean, bright edges, originate under the surface, and appear first in the load zone. A 6205 that reaches its calculated L10 life of, say, 8,000 hours at 1,500 rpm and fails this way has done its job.

Clean-edged spalling in the load zone

Clean edges and a single loaded patch say the bearing was loaded as intended and simply ran out of life. The engineering questions then move to the replacement: was the load estimated honestly, was the clearance right, was the fit correct?

Spalling that arrives early

Fatigue that comes far sooner than the calculation predicts is the interesting case. It usually means the load was higher, the clearance tighter, or the misalignment worse than the selection assumed, and the position of the spall around the raceway hints at which.

Subsurface origin and crack paths

Reading whether a crack began below the surface and grew outwards separates true fatigue from damage imposed from outside, such as a dent that later seeded a spall. The distinction decides whether the answer is a better bearing or a better installation.

Abrasive and Adhesive Wear Compared

Wear removes material gradually instead of in flakes, and the two mechanisms are told apart by surface texture and by where the damage sits.

Abrasive wear: contamination in the film

Abrasive wear is caused by particles in the lubricant: raceways go dull and grey, sometimes grooved along the rolling direction, and the grease turns gritty and discoloured. Its cause chain is sealing and lubrication practice - contamination past a failed seal, a dirty grease gun, or grinding dust left in a housing after a rebuild. The numbers make the mechanism plain: when operating clearance is only 10 to 20 µm, a particle of the same order is enough to bridge the film and mark the raceway every pass.

Adhesive wear: when rolling becomes sliding

Adhesive wear, or smearing, happens when the contact slides instead of rolling: at start-up under heavy load, at excessive preload, or when rollers skid under light load at 3,000 rpm or more. Its signature is galled, torn patches of metal that look almost welded. Smearing on roller ends points to axial load the bearing was not meant to carry; smearing on raceways at start-up points to a viscosity or a load-at-start problem.

Corrosion, Etching and Fretting Powder

Corrosion announces itself in colour. Red-brown rust with emulsified, milky grease is water ingress - condensation in stored or stopped equipment, or wash spray past a failed seal. Black etching or staining can be moisture working with the acidic breakdown products of aged grease. Fretting corrosion is the subtler cousin: red-brown oxide powder at fits and interfaces where tiny relative movements of 1 to 5 µm grind the surfaces, on the outer-ring outside diameter, on shaft shoulders or under a lock nut. Fretting powder in a housing is a fit problem rather than a rust problem, and fitting a new ring into the same worn seat reproduces the failure quickly.

Electrical Erosion: Frosting Before Fluting

Current crossing a rolling contact discharges as a micro-spark and craters the raceway. The early stage looks like frosting, a matte grey roughening, while established damage forms fluting - regular ridges spaced around the raceway at a rate set by the machine's vibration pattern. The grease in the contact carbonises, thickens and loses its lubricity, so electrical damage accelerates mechanical wear. Naming this one correctly matters because the cure is electrical, not mechanical: an insulated or hybrid bearing, a shaft grounding ring and drive-side measures. Replacing the bearing alone buys only a few months on an inverter-driven machine, sometimes fewer than six.

Plastic Deformation: True and False Brinelling Marked

Static overload, impact during mounting and debris pressed through the contact deform the raceway without removing material. True brinelling produces rounded dents at ball spacing and comes from a hammer blow or a dropped assembly. False brinelling produces similar but polished marks from vibration while the bearing is stationary - during transport or storage, or on a machine that sits idle beside a running one. Contamination dents are sharper and more random, and overload flattening appears on a heavily loaded roller path left static too long. Every one of these marks is a stress raiser that seeds early fatigue, so prevention is handling discipline: presses and induction heaters instead of hammers, clean assembly, and rotating stored motors every two to four weeks.

Cracked Rings and Broken Cages

Cracked rings and broken cages usually trace to abuse beyond design. Driving a ring with a hammer to seat it, pushing on the wrong ring during mounting, heating a ring above about 120 °C with an open flame, or a fit so tight the ring splits at assembly will all crack a ring. Cracks can also grow from fretting notches at a fit under a rotating bending load. The evidence is readable: shiny, fresh fracture faces point to a fast break, while oxidised, slow-growing cracks point to fatigue. A ring fractured at a fit-interface notch is a fit problem; one split by torch heating is a training problem. Cages rarely fail first - they usually report a bearing already in distress through secondary damage - so when a broken cage is found, read the raceways and the lubrication history before ordering a replacement.

Three habits separate a competent teardown from guesswork, and all of them protect the evidence rather than the parts bin.

• Keep the parts together. Bag each bearing with its own cage and rolling elements so a failure at one of four positions is never mixed up with another.
• Note the orientation. Which way the load zone faced, and which ring was on the shaft, because that fixes the geometry that caused the mode.
• Sample the lubricant. A teaspoon of used grease tells you about contamination, water and base-oil condition that the metal alone cannot.
• Measure the seats. Record the shaft and housing diameters, because a fit fault will not show in the bearing alone.

Numbers That Help You Name the Mode

A few reference figures settle arguments at the bench, because they describe the conditions that produced each mark. They are typical values for general industrial duty, not limits, but they make the naming easier to defend.

A Repeatable Teardown and Naming Routine

The practice that turns failure analysis into prevention is short and repeatable, and it costs a few minutes per event. Work through it step by step rather than in whatever order the bench allows.

1. Photograph before cleaning. Colour, position and distribution of the marks are destroyed the moment the part is wiped.
2. Identify the components and the mark distribution. Which ring, where around the raceway, and whether the pattern is local or circumferential.
3. Name the mode using the standard term. A shared vocabulary lets two technicians in different plants compare findings directly.
4. Write one line linking the mode to a cause. A fit, a seal, a grease, a current path or a storage rack.
5. Fix the cause, not just the part. Fit the right suffix, clearance and seal set, and record what was found so the next teardown compares against a baseline.

Failure terms, clearance and fit conventions, and mounting procedures are collected in the technical library, and a fuller vocabulary of the standard families is discussed in the ISO 15243 terminology notes. Replacement candidates are filterable in the bearing selector, and the same wallchart of signatures is applied to electric motors in the motor bearing field guide, with the broader method in the failure analysis guide. Life calculation, including the ISO 281 method behind the word fatigue, is set out in the L10 life guide. The bearings that carry most load in general machinery are deep groove ball bearings and cylindrical roller bearings, and replacement parts are widely stocked from makers such as FAG.