A maintenance supervisor pulled a failed bearing from a conveyor drive and sent it to the shop for analysis. The bearing was a 6210 deep-groove ball, running at 720 RPM, carrying a 2 kN radial load. It failed after 8 months. The shop said “fatigue failure, replace with same model.” But the supervisor wanted to know why it failed prematurely. The raceway appearance told a different story. This article walks through how to read a failed bearing the way a tribologist does — by examining the raceway, rolling elements, cage, and lubricant residue.
The raceway tell
When a bearing fails, the appearance of the raceway surface points to the cause. A uniform, smooth, mirror-like track indicates normal rolling contact — the bearing reached its rated life. A patchy track with spalling (small pits) indicates fatigue — the bearing lived as expected. But there are four failure modes that leave distinct marks.
Brinelling: indentations on the raceway, equally spaced at the ball spacing. The marks look like small dents. This happens when the bearing is statically loaded beyond its elastic limit — typically from an impact during installation or when the machine is stationary and vibrates (ship vibration during transport). The bearing was hammered. Not an operating failure — a handling failure.
Fretting corrosion: reddish-brown powder around the bearing. The bearing ring crept on the shaft or in the housing. The microscopic sliding motion between the ring and its seat oxidized the metal. The failed bearing had this on the outer diameter. The housing bore was H7 clearance — too loose. The outer ring crept. The fretting powder entered the bearing and ground the raceway.
Electrical erosion: a wavy, frosted pattern on the raceway, like a washboard. This happens when electric current passes through the bearing (VFD motor shaft voltage). The current arcs across the oil film, pitting the raceway. The bearing was on a VFD-driven motor without shaft grounding. The washboard pattern confirmed electrical damage.
Lubrication failure: the raceway is discolored (blue or brown) from overheating. The rolling elements have a tempered appearance. The grease is black and carbonized. The bearing ran dry or too hot. The original grease was general-purpose lithium — not the high-temperature grease specified for the application.
Reading the actual failure
The 6210 had fretting on the OD and a discolored (blue) raceway. The cage was broken on one side. The grease was black and gritty. The diagnosis: the outer ring crept in the housing (fretting powder entered), and the contaminated grease ran hot (discoloration). The cage broke because it was grinding against contaminated races. The root cause was the housing fit, not the bearing quality.
The corrections
1. Housings re-bored to K7. The original H7 bore allowed the outer ring to creep. A K7 bore has 0.005-0.025 mm interference. The outer ring stays put. No fretting. The bore was reamed to K7 (52 +0.002/+0.018 mm for a 6210 OD of 52 mm). The bearing was pressed in. No creeping after 2 years.
2. Shaft grounding brush. The VFD motor had a shaft voltage. A grounding brush was installed on the non-drive end. The shaft voltage bled to ground through the brush, not through the bearing. The electrical erosion stopped. The bearing raceway stayed smooth.
3. Correct grease and interval. The general-purpose lithium was replaced with a polyurea high-temperature grease (rated to 180°C). The relubrication interval was set to 6 months (from the manufacturer chart). The grease stayed clean. The bearing ran cool (45°C surface). No overheating.
The grease analysis
When a bearing fails, the grease tells a story. Rub a sample between your fingers. If it feels gritty, contaminants entered. If it’s black and watery, oxidation from heat. If it has water droplets, washdown or condensation. If it smells burnt, overheated. A sample sent to a lab can measure particle count, moisture content, and oxidation. For critical bearings, take a grease sample every 6 months. The analysis predicts failure 3 months before it happens.
The bearing life recalculation
The 6210 was rated for L10 life of 20,000 hours at 2 kN load (dynamic rating 35 kN: L10 = (35/2)^3 × 10,000 = 5.3 million hours = 600 years). The bearing should never have failed from fatigue. It failed from contamination and creeping — both preventable. A bearing that fails at 10% of its calculated life is almost always a mounting, lubrication, or contamination issue, not a bearing selection error.
| Symptom | Likely cause | Check |
|---|---|---|
| Indentations at ball spacing | Brinelling (impact) | Installation procedure, transport |
| Reddish powder around bearing | Fretting (ring creep) | Housing fit, shaft fit |
| Washboard raceway | Electrical erosion | Shaft grounding, VFD |
| Blue/brown discoloration | Overheating | Grease type, load, alignment |
| Spalling after rated life | Normal fatigue | Nothing wrong — replace |
| Balls loose, cage broken | Contamination or misalignment | Seals, alignment |
The failure analysis rule: inspect the raceway before condemning the bearing. A premature failure is almost always mounting, lubrication, or contamination. The 6210 wasn’t undersized — it had a creeping outer ring and wrong grease. Re-bore the housing, ground the shaft, and use the correct grease. For critical machines, grease analysis predicts failure months ahead. Don’t just replace the bearing — find out why it failed.
