Case note: fluted raceways on VFD-driven motors

Twelve identical motors, all on VFDs, all losing the non-drive-end bearing within about a year. The drive end, which carries the load, was fine. That asymmetry is the first clue: a load problem attacks the loaded bearing, not the free one.

The outer raceway had regular, evenly spaced dark bands across the rolling path — the washboard pattern that gives fluting its name. The spacing corresponds to the ball passage frequency, which is why it looks machined rather than worn. This is electrical erosion: current passes through the rolling contact, the lubricant film breaks down momentarily, and a discharge crater forms. Repeated, the craters line up into bands.

Once the raceway is fluted, the bearing is finished — the surface is a series of stress raisers and the noise rises before the failure. The question is not how to repair it but how to stop the current.

• “Electrically conductive” grease: at the voltages involved the film still breaks down, and conductive greases trade erosion for other problems.
• Ceramic-coated outer ring: works, but the coating is a wear surface and can be damaged on fitting.
• Insulated housings and non-conductive spacers: effective only if every current path is broken, which is rarely the case in a real machine.
• Shaft grounding ring: addresses the current at the source, but on these motors there was no room at the NDE.

Hybrid bearings with silicon nitride balls were fitted on the non-drive end. The steel rings kept the standard fit and the standard housing; the ceramic balls broke the circuit at the contact. Over the following two years, none of the twelve motors lost a bearing, and the vibration readings on those positions dropped to the level of the drive-end bearings.