Fluting, Frosting and the Voltage You Cannot Hear: A Bearing-Current Checklist for Retrofitted 400 V Drives

A drive-end bearing that fails after eighteen months on a retrofitted 400 V motor is rarely a bearing problem. By the time the workshop report reads mechanical wear, the electrical damage is already months old, and the replacement bearing is on the same trajectory as the one it replaced.

The mechanism, stated plainly

A Bosch drive rectifies the incoming three-phase supply to a DC link of roughly 565 V, then switches the output through an IGBT bridge. That switching pattern is not symmetrical about earth: the instantaneous sum of the three output voltages is not zero, so a common-mode voltage is imposed on the stator winding. Each switching edge moves that voltage by several hundred volts in a few microseconds.

The motor behaves as a capacitive divider \u2014 winding to rotor, rotor to frame, frame to protective earth. The rotor picks up a voltage capacitively, and the only low-impedance route to earth is through the bearings, across the lubricating film. Where the film is thin, it breaks down and a discharge occurs. Individually these events are trivial; in aggregate, tens of thousands per hour, they machine the raceway. Frosting appears first, then the characteristic washboard fluting, then noise, then seizure.

A second mechanism appears on larger frames, broadly from 100 kW upward: circulating current. Asymmetry in the magnetic circuit induces a low-frequency voltage along the shaft, driving current axially through both bearings and the frame. It is a different fault with a different remedy, and treating one as the other is a common and expensive error.

Why conventional monitoring reports it too late

Vibration analysis watches for bearing defect frequencies \u2014 BPFO, BPFI, cage and roller tones. Those signatures develop once damage exists. Electrical erosion is silent in its early phase: the motor sounds normal, runs at normal temperature, and the only visible clue is darkened grease at service. If the maintenance strategy rests on vibration alone, the first reliable notification of a bearing-current problem is a failure.

That is not an argument against vibration monitoring. It is an argument for knowing what the technique cannot see, and for adding a measurement that does.

Part one: measure before mitigating

Anything fitted without a before-and-after reading is a guess. The measurements worth taking are:

• Shaft voltage. Oscilloscope with a high-impedance probe, carbon or copper-graphite brush riding the shaft, reference to motor frame earth. Capture peak-to-peak and look specifically for sharp discharge steps rather than a smooth waveform.
• Common-mode current. Rogowski coil or split-core RF current transformer encircling all three motor phases together at the drive output.
• Bearing current. Split-core RF current transformer around the shaft adjacent to each bearing housing.
• Service conditions. Switching frequency, motor cable length and type, earthing arrangement, motor frame size, and the complete drive parameter set. A shaft-voltage figure without these is a number, not a baseline.

Interpretation is comparative. A sister machine, or the same machine before the retrofit, gives the reference that a generic threshold cannot. Sustained discharge activity, or shaft voltages rising into the tens of volts, warrant mitigation.

Probing a rotating shaft requires a permit to work, effective guarding and a second person present. This is not a task to improvise at the terminal box.

Part two: cable, screen and glanding

• Use screened variable-speed cable, not single cores pulled through trunking.
• Terminate the screen at both ends in 360-degree EMC glands. Pigtail drain wires defeat the purpose.
• Bond the motor frame to the drive enclosure earth with a dedicated conductor, sized and installed in accordance with BS 7671.
• Keep unshielded conductor tails inside the terminal box as short as practicable.
• Check the drive manufacturer's maximum motor cable length at the switching frequency in use, and fit a dv/dt or sine filter where runs are long or where several motors are paralleled on one drive.

Part three: bearing protection

• An insulated bearing \u2014 hybrid ceramic or ceramic-coated \u2014 at the non-drive end interrupts circulating current.
• A shaft grounding ring or brush at the drive end gives rotor-to-earth current a deliberate path, ahead of the bearings.
• The brush must contact bare metal, sit correctly in its housing, remain spring loaded, and stay clear of paint, grease and corrosion.
• Fitting one without the other is the classic error. An insulated bearing with no grounding path does nothing about discharge currents, and a brush alone does not stop circulating current.

Part four: the record kept between visits

A bearing-current problem is a trend, not an event, so the file matters as much as the fix. Record the shaft-voltage and common-mode baselines with instrument details; measure brush continuity and wear at every service; read and log the insulated bearing's resistance; trend bearing temperature; photograph grease condition at each strip-down; and store the drive parameter set alongside the readings.

Re-measure after any change to switching frequency, cable route, motor or firmware. Raising switching frequency to quieten audible noise increases common-mode activity, and a baseline taken before that change describes a machine that no longer exists.

The practical conclusion

Bearing currents are cheap to diagnose and cheap to mitigate, provided they are suspected. They are expensive when they are not, because the symptom \u2014 a failed bearing \u2014 invites a like-for-like replacement that reproduces the cause. Treat every bearing failure on a retrofitted 400 V drive as an invitation to measure the shaft, not simply to fit another bearing.