Diagnosing Rolling Bearing Failure: Fatigue Spalling, Wear, Corrosion or Electrical Erosion — and What Each Pattern Tells You to Fix

Rule out electrical erosion before you blame fatigue

On an inverter-driven motor, a regular fluting or washboard pattern around the raceway circumference is electrical erosion, not fatigue¹⁵. The mechanism is stray shaft current: current arcs across the thin oil film and micro-craters the raceway, and the craters progress into the fluted pattern¹⁴.

That is why a fluted race keeps failing after every lubrication and alignment correction — the current path is still there, and only breaking it stops the recurrence^15,21.

One supplier reports that stray shaft currents above 1.5 V arc through the oil film and cause micro-cratering¹⁴; treat that as one source's reported value, not a threshold you can screen against, and ask your drive or motor supplier what current evidence they can supply.

The same damage can be misread as spalling if you do not consider the drive at all, so check the motor type and drive arrangement before you commit to a fatigue diagnosis¹⁵.

Is the spalling end-of-life fatigue or premature fatigue?

Pitting or flaking on raceways and rolling elements is fatigue: cyclic stress over time eventually removes material, and that is expected at rated life^15,21. The diagnostic question is not whether spalling happened but whether it happened early — premature spalling points to overloading rather than a defective bearing²¹.

Before you accept the overload explanation, look at the wear distribution: fatigue spalling is pitting on the race from cyclic stress, while misalignment wear is heavier on one side of the race¹⁵. A one-sided wear band sends you to shaft and housing alignment, not to the load rating.

If the pattern is uniform pitting and the service hours are short of the calculated life, the load path is the next thing to check^15,21.

Corrosion or abrasive wear: two degraded surfaces, two different fixes

Corrosion shows as reddish-brown pitting or rust on the raceway or balls, and it comes from moisture ingress, seal failure, poor storage, or the wrong lubricant^15,21. Abrasive wear looks different: a dull, worn surface on raceways and rolling elements, caused by contaminated lubricant carrying abrasive particles²¹.

The two call for different corrections even though both present as a degraded raceway. Corrosion sends you to sealing and storage conditions; abrasive wear sends you to lubricant cleanliness and filtration or more frequent lubricant changes²¹.

Contamination is described as a leading cause of premature wear, so if the surface is dull rather than rusted, treat the lubricant and its environment as the suspect¹⁹.

Rule out installation and stationary-vibration damage first

Indentations are made before the bearing ever runs, so rule them out before attributing any mark to a service root cause. Hammered installation instead of a press fit leaves brinell marks at random angles¹⁵.

Vibration during storage or transport of a stationary bearing leaves elliptical dents at ball spacing — false brinelling¹⁵. The pattern geometry tells the two apart: random angles point to the mounting method, ball-spaced ellipses point to how the bearing was stored or moved¹⁵.

If you find either pattern, the corrective action is in your mounting procedure or your stores, not in the operating conditions.

Confirm the mode with the monitoring the sources support

Vibration analysis and thermography can flag a developing failure before it causes downtime, so use them to confirm a trend rather than to name the mode from a single reading¹⁰.

A bearing above 300°F in a shutdown condition is likely in its final stages of failure, and the time to react may be hours rather than days — at that point focus on limiting secondary damage to shaft and housing⁵.

• ✓Check whether the machine is inverter-driven before you accept a fatigue diagnosis
• ✓Compare observed service hours against the calculated L10 life for the actual load
• ✓Look at the wear distribution: uniform pitting versus one-sided wear
• ✓Check the raceway surface colour and texture: rust versus dull worn metal
• ✓Check for ball-spaced elliptical dents versus random-angle brinell marks
• ✓Trend bearing temperature and vibration; a change matters as much as the absolute value

Match the corrective action to the mechanism you found

Each diagnosis implies a different correction. For electrical pitting, insulated bearings or shaft grounding rings are recommended to prevent recurrence²¹.

For lubrication-driven damage, the regime itself is the variable: both over- and under-lubrication can cause failure, so correct the amount and type rather than simply adding grease¹⁹.

Where high temperature, high speed, or heavy contamination defeats the grease, oil can dissipate heat and flush contaminants, and converting from grease to oil is the direction those conditions point to². Grease that loses viscosity under those conditions allows metal-to-metal contact and eventual seizure, which is the mechanism behind the dry-out pattern².

For fatigue and misalignment wear, the correction is in the load path and alignment rather than the lubricant¹⁵.

Was the life actually short? Use the load-life relationship

L10 life follows a cubic relationship with load: doubling the equivalent dynamic load cuts calculated life to about one-eighth^11,12. The formula is L10 = (C/P)³ × 1,000,000 revolutions, where C is the basic dynamic load rating and P the equivalent dynamic load¹².

That is why a small overload produces a large life shortfall, and why a premature spall is evidence about the load rather than about the bearing's manufacture¹¹.

Reported L10 figures give you a rough expectation to compare against: one supplier cites 30,000 to 100,000 hours for tapered roller bearings under proper lubrication, alignment, and normal load, and another cites 40,000 to 100,000 hours as the theoretical figure often specified for motor bearings^10,14.

Both are supplier-published ranges, so use them as a sanity check on whether the observed life was short, not as a guaranteed life for your application.

How much of the failure belongs to the bearing itself?

The sources disagree on whether failures are predominantly application-driven or bearing-driven, and that disagreement is unresolved^14,15. What is settled in general terms is the material side: through-hardened GCr15/AISI 52100 reaches HRC60+ with high wear resistance and fatigue strength, which is why it is the common bearing steel^1,6.

What the ABEC scale does not control is just as relevant when you are deciding whether to change the bearing: it does not cover material composition, heat treatment, cage quality, surface finish, internal clearance, or lubrication quality²².

So a higher precision class is not a route to longer life against a lubrication or contamination root cause. If your diagnosis points to the application, changing the bearing grade will not fix it.

The failure-share figures the sources state differently

Two sources give different answers to the same question — what share of failures is application-caused versus bearing-caused. Both figures come from supplier-tier pages, so neither should be quoted in a report as an industry statistic without checking the basis behind it.

What the sources do not establish

• No source provides magnification-level morphology to separate fatigue spalling from wear, corrosion, and electrical erosion — you will need a laboratory or specialist reference for that.
• No source covers magnetic particle inspection, dye penetrant, BPFO/BPFI/BSF/FTF frequency analysis, shock pulse, electrical current measurement, or lubricant analysis as diagnostic methods for these modes.
• No source quantifies wear rate, corrosion rate, or time-to-failure curves under different operating conditions.
• No source describes the content of ISO 15243 or ANSI/ABMA failure-mode terminology, so neither can be used here to classify the damage.
• No source covers corrosion-resistant materials or coatings, specific seal upgrade types, alignment correction procedures, or mounting procedure revision as corrective actions.
• No source covers vibration as a root cause of fatigue spalling, or excessive temperature as a distinct root cause beyond lubricant dry-out.
• No source covers sealing or shielding type selection, or internal clearance selection, as design factors in failure prevention.
• No source provides a comprehensive root-cause distribution across all failure modes.