The third bearing died on a Tuesday.
Same machine — a 500 kVA set at a packaging plant, six years old. Same drive-end bearing on the alternator. Eighteen months, three bearings. The maintenance crew had a theory they were very attached to: bad grease. Then bad bearings. Then bad luck, in that order.
I cut the dead one open on their workbench. The outer race had a pattern like corduroy — fine, even ridges running all the way around the track, as if someone had taken a miniature file to it. The crew chief nodded like this proved his point. “See? Defective batch.”
Three bearings from three different batches don’t share a defect. That pattern has a name: fluting. And fluting isn’t made by friction. It’s made by electricity — tiny arcs jumping through the bearing thousands of times a second, each one blasting a microscopic crater in the steel. Give it enough months and the craters line up into ridges. Nobody wears out a bearing in six months. That one was electrocuted.
This is the most expensive mistake on the list, because it feels like action. Bearing fails, you press in a new one, the machine runs quiet for six months, and everyone moves on. Until the whine comes back.
Fluting takes months to carve those ridges. A bearing that dies of it inside a year was being arced from the day it went in. So the replacement you just installed is getting the same treatment, starting now. You didn’t fix anything. You reset a timer.
The dead bearing is trying to tell you, if you read it before it hits the scrap bin. Bad grease leaves smearing, overheating, a bluish tint. Overload spalls — flakes of steel peeling off one side of the race. Shaft current leaves the washboard, sometimes with a frosted matte finish on the balls. Five minutes with a magnifier. It would have saved that packaging plant two bearings.
Grease does kill bearings — I’ve written about the five greasing mistakes that kill alternator bearings before, and every word of it stands. This just isn’t a grease story. Engine crank bearings have their own separate ways of dying (main bearing diagnosis here), but the alternator’s bearings live in their own electrical weather system.
The crew’s second theory was starvation, so they packed the replacement absolutely full of grease. I understand the instinct. More protection.
Wrong direction. The voltage builds on the shaft until it’s high enough to punch through the oil film, then it arcs — vaporizing a speck of steel and carbonizing a speck of grease with every jump. Pack in more grease and you’ve handed the arcs more material to cook. I’ve opened electrically killed bearings and found the grease turned to black grit. Collateral damage. The grease never had a vote.
An old-timer I work with disagrees with me here, by the way. Thirty years on rotating machines, and he still says nine times out of ten it’s lubrication or fit. He’s right nine times out of ten. This is the tenth — and the washboard pattern is the giveaway, which, ironically, he taught me to look for.
A shaft shouldn’t carry voltage. In a perfect machine it doesn’t. It gets there two ways.
The first is the machine itself. No alternator is magnetically perfect — small asymmetries in the iron push a little circulating current around the shaft-bearing-frame loop. On most sets this stays small enough to ignore for the life of the machine. Background radiation.
The second way is the one that actually kills bearings, and it usually arrives on a truck: variable-frequency drives. A VFD chops DC into fast PWM pulses, and the switching creates common-mode voltage — high-frequency noise that couples capacitively from the stator winding straight onto the rotor and shaft. Run VFD-driven pumps, fans, or compressors off your generator and the shaft can carry spikes the bearing’s oil film can’t stand off.
A water treatment plant taught me this one. Eight months after they put VFDs on the aeration blowers, the alternator bearing started singing — that high, unhappy whine. We put a scope on the shaft, carbon brush riding it, probe to frame, and watched spikes dance across the screen every time the drives ramped. The old bearing had fluting you could feel with a fingernail. The drives were installed correctly. The generator was installed correctly. Nobody had asked what the drives were doing to the generator’s bearings. Why would you? That’s the trap.
The longer version of what drives do to generator power quality lives in the harmonics article. Same family, different victim.
Look, the catalog has answers, and two of them are traps.
An insulated bearing — ceramic-coated race, usually — blocks the current path through that bearing. Good. Now put insulated bearings on both ends with no shaft grounding, and the current doesn’t vanish. It reroutes: through the coupling into the engine crankshaft, through the exciter bearings. You’ve moved the execution to a more expensive neighborhood.
The honest setup is one insulated bearing, typically the non-drive end, plus a real grounding path on the other end — a grounding brush or a conductive microfiber ring riding the shaft, giving the current a low-resistance road home that isn’t your bearing. Two devices, two different jobs.
And then there’s the grounding brush nobody ever looks at again. I found one at a hospital — holder still bolted on, brush worn to a stub, spring doing nothing, pigtail green with corrosion. It had been “protecting” that shaft for about four years. A grounding device you never inspect is a decoration. Put it on the checklist with everything else (mine lives here).
A handheld multimeter tells you a partial truth here. It’ll catch DC offset and low-frequency content, and it’ll sail right past the high-frequency spikes from a drive. If you can borrow a scope, put it on the shaft and look for hash riding the waveform. No scope? The five-minute version:
While the end bracket is off, megger the windings too — you’re already in there (insulation testing procedure). Two birds.
We scoped the shaft, found the spikes, and traced them to a VFD-driven dust-collection fan installed the year the failures started. The fix: a grounding ring on the shaft, one insulated bearing at the non-drive end, and the drive’s cable shielding grounded properly. Three years ago. The bearing in there now is the one from that week.
It cost a fraction of one unplanned outage. The plant had been budgeting “a bearing every six months” like a consumable. It never was one.
If your alternator bearings keep dying young and the grease looks fine, stop feeding the machine bearings. Measure the shaft. The voltage is small — tenths of a volt will start the damage — but it’s patient, and it never takes a day off.
If your generator’s due for a look — or its bearings are dying on schedule — talk to us. We’ve read this corpse before.
Email：[email protected]
Work：Monday to Friday,9:30-18:30
