ISO 15243: Naming Bearing Failure Modes

A failed bearing is a scene with evidence scattered across it, and the first mistake most teams make is to argue about the story before they have described the scene. One person says the part "wore out", another blames the grease, a third blames the supplier. The words are doing none of the work. ISO 15243 exists to remove that ambiguity: it is the international standard that classifies rolling bearing damage by the appearance and mechanism of the surfaces involved, giving maintenance teams, suppliers and warranty desks a single, checkable vocabulary for the most expensive conversation in the trade.

The standard's central discipline is an order of operations. Describe the damage first, interpret the mechanism second, and assign the cause only last. Follow that order and the argument shrinks to a manageable size. Reverse it - settle on a cause first and then read the evidence through that conclusion - and teams misdiagnose in exactly the pattern the standard was written to break.

Six modes, one shared vocabulary

ISO 15243 groups rolling bearing damage into six primary modes, each defined by what can be observed rather than by a guess at what caused it: material fatigue, wear, corrosion, electrical erosion, plastic deformation, and cracking or fracture. The structure is deliberately diagnostic. You begin with what you can see on the damaged surface, and the mode narrows the mechanism, which narrows the cause, which points to a corrective action.

Failure proper versus impairment

The standard also separates damage that stops a bearing from working - failure proper - from damage that impairs but does not stop it. That distinction decides whether a claim is about capability or about a life cut short, and it matters in a planned-replacement decision as much as in a warranty dispute.

The six ISO 15243 families, described by appearance and by the mechanism that usually produces it.

Size alone does not classify a mode, but the scales are worth knowing. A subsurface fatigue spall on a 6205 inner ring may begin as a pit well under 1 millimetre across before it spreads; electrical fluting forms ridges roughly 0.5 to 1 millimetre apart around the full 360 degrees of the raceway; an impact brinell is often 0.1 to 0.5 millimetres deep; and a fretting oxide band at a loose fit can run 2 to 3 millimetres wide. Because those ranges overlap, appearance and location carry the classification, not scale by itself.

Why a shared vocabulary settles disputes

The terms are not academic. When a plant, a supplier and an insurer all use the same mode names, three things become possible at once. A maintenance log turns into data that aggregates across a fleet instead of a pile of anecdotes. A warranty conversation becomes auditable, because two parties who both describe the marks as fluting from electrical erosion are already most of the way to agreeing on the remedy. And a supplier's quality record becomes comparable, because a failure signature can be tied to a design, a fit or a duty rather than to an adjective. An entry reading "bearing failed" teaches the next engineer nothing; an entry reading "fatigue, subsurface-initiated, in the load zone, at roughly 60 percent of the calculated life" opens a real investigation with real questions.

Classify first, diagnose second, blame last

The method embeds a professional discipline that is easy to state and hard to practise. Classify the observable damage first, interpret the mechanism second, and assign the cause only last. Workshops that reverse the order - deciding it must be poor lubrication and then reading the evidence through that lens - misdiagnose systematically, because the conclusion is fixed before the observation begins. The correct sequence fits the field in minutes: photograph the bearing before cleaning, note which components are damaged and where the marks sit, match the appearance to a mode, and only then reason about the mechanism and its cause.

Location is part of the description

Where a mark sits is evidence in its own right. Damage confined to the loaded zone points in one direction, damage spread around the full circumference in another, and marks spaced at the ball pitch in a third. Recording location as carefully as appearance is what lets a second engineer reproduce the classification.

Cleaning destroys evidence

Grease condition, corrosion products and trapped debris are part of the story. Washing a part before it is photographed removes exactly the information an investigation needs, so in every workflow the camera comes before the solvent.

Photograph before you clean

A camera and a scale are the two cheapest diagnostic instruments in the plant, and they are the two most often left in the toolbox. A photograph taken before cleaning preserves the position, spread and colour of the damage, and colour carries meaning: the darkened, carbonised grease around a fluted raceway tells a different story from the clean grey of a fatigue spall. Photographs also settle arguments months later, when memory has faded and only the picture can confirm which face of which ring carried the marks.

• Keep the part intact. Do not cut cages or press rings apart until the external evidence is captured.
• Use a scale. A rule across the frame turns a photograph into a measurement.
• Shoot the assembly first. The position on the shaft is evidence the loose part cannot supply.
• Record both faces of both rings. The unworn face is the control that proves the worn face is abnormal.

The teardown sheet that carries the mode

Adopting the vocabulary costs almost nothing when it is built into a one-page teardown sheet that drives a step-by-step routine. The sheet is deliberately short, and it uses the standard's terms, so the record is portable between shifts, sites and companies.

1. Record the identification. Designation, position, running hours, and the calculated life at the point of removal.
2. Photograph in place and in the hand. Against a scale, before any cleaning, catching grease colour and debris.
3. Name the modes observed. One or more of the six, described by appearance and location, not by suspected cause.
4. State the mechanism hypothesis. A single line, offered for review rather than as a finding.
5. File the sheet with the work order. Keep the record where the next teardown can find it.

Turning mode logs into fleet statistics

A single classified failure is a story; a hundred are a database. Once records use the standard's terms, patterns appear that were invisible in prose - the same logic that lets a vibration trend, read against the severity bands of ISO 10816, show a developing fault before it becomes a stoppage. The value is in reading the pattern rather than the individual entry.

• Repeat offenders. A motor line accumulating electrical-erosion entries is telling the plant to look at its drives and its grounding.
• Storage problems. Corrosion entries in low-duty positions often point at storage and washdown rather than at the bearing.
• Capability questions. A supplier appearing in fatigue entries at short lives is being asked a capability question, not a price question.
• Handling losses. Deformation entries frequently trace back to packaging and transport rather than to the running duty.

Where ISO 15243 sits in the standards family

ISO 15243 completes a set. ISO 281 defines the life calculation that sets the expectation against which an early failure is judged, ISO 15 and ISO 199 fix the boundary dimensions that make parts interchangeable, and ISO 492 grades the tolerances. Failure classification is the standard you reach for when reality departs from what the others assume. Manufacturer atlases and this library's own guides on failure analysis and telling failure modes apart apply the same vocabulary, with photographs, to real parts. The measuring accuracy notes show how the same discipline of observation is applied to a new part, and the deep groove ball range and self-aligning roller range show where each mode most often appears in practice. A bearing from a maker such as SKF arrives with the same designation the standard's terms refer to, which is what makes the vocabulary portable across the whole technical library.

Training a team to speak the standard

Adoption lives or dies on the people who handle the parts. A half-day session built around photographs of the plant's own failures, classified together against the six modes, does more than any manual, because the team learns to argue about the cause only after agreeing on the description. Add a laminated teardown sheet to each maintenance trolley and a mode pick-list to the work-order software, and make the quarterly review a habit.

The record becomes the training material

Within a year the plant's own log is the most useful training material the next cohort can have, because it carries the accumulated pattern recognition of the team that came before. The standard supplies the words, and the plant supplies the memory.

A vocabulary that pays rent

Applied consistently, the terms are the cheapest reliability programme in the plant. They cost a sheet of paper and a habit, and they turn every subsequent teardown into a data point instead of an argument.