Are VFD bearing currents causing fluting? Measure and verify the remedy

Are VFD bearing currents causing this fluting?

Separate bearing condition, present electrical exposure and remedy performance. A replacement bearing can be quiet while discharge continues; a protected bearing can remain noisy from existing damage. Assign electrical correction and mechanical repair separate dispositions.

ISO 15243:2017, edition 2, classifies rolling-bearing damage by appearance and possible causes; its public abstract calls for further investigation when appearance cannot reliably establish the cause. It is not a shaft-voltage acceptance standard. SKF's September 2011 maintenance handbook also identifies electrical erosion from flux asymmetry and process or belt static charging, so establish when failures began relative to drive conversion and service changes.

ISO public scope; SKF handbook, pp. 319 to 322.

What turns a suspicion into a supported diagnosis?

Preserve tagged bearing photographs, orientation and lubrication history, operating shaft traces, and a drawing of the likely loop. A sharp voltage collapse supports a discharge mechanism, while craters and melted features support electrical erosion in the bearing. Correlating operating history and current path connects those findings; an external shaft trace alone does not identify which internal bearing discharged.

Where does the current go?

ABB Technical guide No. 5, Rev C (2011), distinguishes capacitive discharge, circulating bearing current and shaft-grounding current, so a remedy must address the identified loop rather than shaft voltage alone. A grounded frame still permits capacitive shaft charging. Record bearing insulation and shaft-grounding hardware separately from the motor's inverter-duty winding rating.

A variable frequency drive controls motor speed; the PWM drives discussed here switch their output in rapid voltage pulses. Common-mode voltage is the average of the phase voltages relative to the reference, which can change even while the commanded motor speed stays steady. High-frequency (HF) return impedance determines which available paths carry its fast switching currents. The distinction between frame bonding and shaft grounding follows the separate circuits explained in grounding versus bonding.

Three loops the field drawing must distinguish

Capacitive coupling charges the shaft; discharge crosses a bearing to the frame and returns toward the inverter. A brush provides a parallel branch. Circulating current follows the shaft, both bearings and frame, so a brush alone need not interrupt that loop.

In the driven-equipment loop, elevated motor-frame potential drives current through the motor bearing, shaft, conductive coupling and load bearing toward the inverter return. Include pump, gearbox and encoder branches on the site drawing. A frame equalizing bond and bearing insulation change different parts of that circuit.

Label the test points before choosing a remedy

Mark TP1 at the drive-end (DE) shaft-to-frame point, TP2 at the non-drive end (NDE), and TP3 between motor and driven frames. Add accessible brush branches and any specialist test jumper to the job drawing. Label inaccessible paths explicitly; a brush-current reading measures that branch, not the internal bearing.

Keep the cable/shield arrangement with the VFD installation and commissioning record. Record drive-fed operation and any bypass operation separately using the actual motor starting methods.

Is it electrical erosion or lubrication damage?

Fluting, dark grease, noise and a hot housing justify investigation; none alone proves VFD attribution. Electrical erosion can degrade lubricant and produce secondary fatigue alongside lubrication damage.

Before cleaning, preserve grease and orientation and photograph both rings, rolling elements, load tracks, seals and fit surfaces. SKF's Bearing maintenance handbook, PUB SR/P7 10001/1 EN (September 2011), shows current-leakage craters at 500× and 1,000× magnification and describes thermally altered material in cross section. Use laboratory examination when ordinary magnification cannot distinguish the surface mechanisms. Preserve an unwashed lubricant sample and the least-damaged raceway areas; heavy spalling can erase the initiating damage.

What the damage pattern supports

NSK Electrical Erosion describes local melting and crater-like depressions; NSK False Brinelling connects repeated contact-position damage with vibration and small oscillations, including transport. NSK Handling Instructions for Bearings, copyright 2010 supplies the broader mechanical differential.

Compare spacing and surface texture. False brinelling follows repeated rolling-element contact positions; electrical fluting develops from erosion and rolling contact. Seek crater or melt evidence when transport vibration and poor lubrication are also plausible.

What heat and vibration add

Vibration trends establish deterioration and its rate; thermal images record housing temperature at measured load and ambient conditions. Neither measures bearing discharge. Use electrical thermography inspection for the thermal record.

A damaged bearing still needs a mechanical disposition under the machine's established alarm settings, deterioration rate and failure consequence. A low shaft-voltage result does not authorize continued operation of a spalling bearing. Use motor bearing lubrication and shaft grounding for grease amount, interval and service procedure after identifying the damage mechanism.

How do you prepare a safe, repeatable shaft test?

Stop and isolate the machine before preparing the shaft, attaching the probe or changing its position. Secure the brush and leads, establish guarding, and operate the recorder outside the rotating-parts hazard area. Helwig's BPK-Probe V2 guide warns that even a smooth, slowly rotating shaft can catch clothing or skin; PPE does not replace guarding.

OSHA 29 CFR 1910.333(a)(1), July 1, 2025 edition, requires deenergizing exposed live parts unless its stated exceptions apply; subsection (c)(2) limits work on energized circuit parts or equipment to qualified persons. Subsections (b)(2)(ii) and (iv) require isolation from electrical sources, control of stored energy and verification of the deenergized condition. A drive stop command or interlock alone is not isolation; follow the actual drive's capacitor-discharge instructions and the site's electrical lockout procedure.

For machine servicing covered by OSHA 1910.147, July 1, 2025 edition, subsection (f)(1) governs temporary energization for testing: clear tools and materials, remove employees from the machine area, remove energy controls under the prescribed procedure, test, then deenergize and reapply controls before further servicing. OSHA 1910.212(a)(1), the same edition, requires guarding against rotating-part hazards. If guarded access cannot be arranged, record that shaft end as inaccessible and plan another measurement arrangement.

Use manufacturer-approved isolated inputs or a suitably rated differential probe where the reference connections require them. Shared oscilloscope ground clips can connect frames together and bypass an insulated path; map every channel return before connection. Tektronix's floating-measurement safety brief prohibits defeating oscilloscope protective grounding. Keep connections within the probe's and instrument's voltage, category and isolation ratings.

Helwig operating guide; OSHA electrical work practices; OSHA energy-control testing sequence; OSHA guarding; Fluke instrument and setup instructions.

Pretest asset and operating record

Read the motor nameplate and record model, serial, kW/hp, voltage, current and nominal speed. Add bearing IDs and insulation, coupling and driven machine, drive model and firmware, switching settings, output frequency, cable type and routed length. These identify which manufacturer application instructions govern the remedy.

Fluke's September 2018 MDA manual calls for a motor at normal operating temperature; ABB TG5 Rev C requires at least 10% nominal speed for its bearing-current measurement method so the bearings rise on the oil film. Record actual speed, load indication, housing temperatures and time since start. Thermal stabilization and minimum measurement speed are separate conditions.

Fluke manual, pp. 21 to 23; ABB TG5, p. 20.

What bandwidth, probe and capture settings do you need?

EST Q&A Part 16 gives the shaft-voltage bandwidth recommendation below. Select the measurement chain for the event: analog bandwidth determines which frequencies reach the recorder, actual sample rate sets sample spacing, and memory length determines capture duration. Record the rate used at the selected timebase.

Shaft-voltage instrument checklist

Compensate the probe as instructed and keep the brush clear of keyways. Record contact and setup photographs with the checklist settings.

Follow the exact probe assembly instructions: Fluke's September 2018 manual removes the specified tip sleeve for its brush, while AEGIS Handbook edition 4 (2024), p. 31, retains the PP510 insulated sleeve. Save a broad PWM view and a close view of event edges. Identify clipping and noise floor; a visually vertical edge does not establish a numerical transition time.

AEGIS edition 4; Fluke application note.

For current measurements, specify sensor, transfer ratio, recorder and accessible conductor together. Verify pulse range and noise floor; the current-chain figures above are instrument capabilities.

How do you measure shaft voltage in the field?

Capture the drive end, then the accessible non-drive end, and repeat at each declared duty point before and after correction. Verify brush contact in ordinary waveform view before relying on an event display. Save maximum, minimum and peak-to-peak separately: for an asymmetric trace, Vpp is maximum minus minimum and is not automatically twice its positive peak.

Fluke's September 2018 manual describes approximately 20 seconds before the event-rate reading appears; its specifications call that rate indicative. Record the configured voltage-change and transition-time definition with events/second. An empty EVENTS display is not a flat zero-voltage waveform: return to ordinary waveform view to confirm acquisition and contact.

Use matched duty points and an explicit observation window

Anvilfield's proposed sampling plan uses normal low, operating and high duty points permitted by the process, with three 60-second observation windows at each accessible end after stabilization. Retain repeated short high-speed captures during those windows and repeat the matrix after correction. Record any unavailable duty point as untested operating coverage.

Anvilfield estimates 36 minutes of observation for two ends, three duty points, three windows and before/after testing: 2 × 3 × 3 × 60 × 2 seconds, divided by 60. This method counts observation only; setup, warm-up, transitions and reporting extend the assignment. Retain the actual captured duration and known trigger dead time separately from observation time.

Divide record length by actual sample rate to obtain acquired duration. Anvilfield estimates that a selected 1-million-sample record at 1 Gsample/s covers 1 ms, while continuous recording for 60 seconds at that rate requires 60 billion samples per channel. This calculation multiplies selected sample rate by duration and explains why a minute watching the display is not a minute of saved high-speed data.

Disable the BPK embedded grounding simulator for the untreated baseline and final permanent-hardware test, following Helwig's operating guide. Store a simulator-on comparison as its own test state. Change one item at a time during diagnostic trials; when several repairs are installed together, attribute the result to the repair package.

Fluke manual, pp. 22, 23 and 33; Helwig simulator instructions.

Which waveforms indicate discharge, and what voltage is too high?

Read waveform shape before assigning meaning to amplitude. The AEGIS ranges below describe typical patterns, not universal acceptance limits. Common-mode steps show shaft excitation, while repeated charge-and-collapse patterns support electrical discharge machining (EDM). When the lubricant film breaks down, an arc locally melts the contact surface, leaving the craters described by NSK. Check apparent collapse events against switching edges, brush bounce, reference-lead pickup and clipping; repeat with verified contact and consistent geometry before assigning them to a bearing.

Fluke's MDA-550/MDA-510 Users Manual (September 2018) identifies shaft-voltage changes above 15 V with transition times below 50 ns as possible grease flashover indicators and states that no fixed voltage defines bearing damage for every motor.

Fluke manual, p. 22.

Why lower shaft voltage can accompany a damaged bearing

The low-amplitude discharge pattern in the AEGIS comparison can accompany lubricant whose impedance has fallen, allowing more continuous conduction. Carbon and metal contamination can create that lower-impedance condition. Speed, load, temperature and film condition also change the circuit.

AEGIS edition 4, p. 30; Helwig Pocket Guide GDE-009/2023; EST resistance article, January 9, 2024.

Can you measure actual bearing current?

Direct internal bearing-current measurement normally requires an isolated or modified bearing path or a purpose-designed test arrangement; ABB TG5 Rev C cautions that it is unavailable on a standard motor. A current probe on a brush conductor measures diverted brush current. Identify the enclosed path, sensor bandwidth, transfer ratio, polarity and range on every saved trace.

Zero signal in one frame bond does not clear parallel paths, and drive protective-earth current is not a direct bearing-current reading. Expand the test scope when damage recurs despite sound brush contact, the two motor ends disagree, a circulating loop is suspected, or pump, gearbox or encoder bearings remain exposed.

When the shaft-voltage result needs a specialist current-path test

The manufacturer-authored ASHRAE Transactions 114(2) paper SL-08-025 (2008), Figure 6, describes isolated bearings in a 5 hp, 460 V, three-phase test motor and reports 7 V peak shaft voltage with no measurable bearing current after microfiber grounding. The published detection limit is unspecified. The test distinguishes remaining shaft voltage from measured bearing current.

ASHRAE published test, p. 250; ABB TG5, p. 19.

Which remedy fits the measured current path?

Start with the installation's high-frequency return circuit. ABB TG5 Rev C (2011), pp. 17 to 18, calls for 360° shield connections at both ends and describes 50 to 100 mm wide copper braid for high-frequency potential equalization. These are installation practices separate from required protective-earth sizing.

A brush diverts current, insulation interrupts a selected loop and a common-mode filter reduces excitation. Correct damaged bearings, alignment and lubricant problems identified in the same investigation.

Compare the mechanisms before comparing products

Specify the filter's intended function and its motor, drive and cable application. A line-side harmonic reactor or improved phase-voltage trace does not establish bearing-current mitigation. Keep harmonics and power quality results separate from the shaft/path verification.

Where manufacturers differ and which instruction controls

Follow the installed motor/drive and accessory instructions; resolve incompatible configurations with the supplier. These manufacturer boundaries serve different applications.

ABB's both-end insulation option requires the actual motor-series option table: the October 2025 catalog lists variant 702 unavailable across the table on p. 55 and available for frames 100 through 250 in the table on p. 61. Keep special drive-end roller or angular-contact bearings and vertical thrust arrangements separate from the general radial-ball-bearing example. Verify the coupled equipment after any insulation change.

ABB 9AKK105944 EN 10-2025; AEGIS edition 4, pp. 16 to 19.

How do you verify shaft grounding after installation?

Test the permanent remedy at the baseline temperature, speed, load and drive settings. Establish conductive shaft contact and mount conductivity before interpreting a low reading. Retest both motor ends and any driven-machine path identified in the diagnosis; a good result beside the brush does not clear a remote bearing.

Why an ohmmeter reading does not close the job

EST's January 9, 2024 resistance article explains that operating high-frequency performance cannot be inferred from static ring resistance alone. Continuity checks locate gross open connections; operating waveforms establish protection under the tested conditions.

When the final trace remains poor, stop and isolate before checking for paint beneath the mount, shaft coating or corrosion, loose connections, damaged fibers or brush wear, and accidental bypasses around insulated bearings. Follow the installed accessory instructions for surface preparation and attachment. Correct contact defects and repeat the original measurement before changing the remedy design.

Accept, repeat or reopen the diagnosis

Set the acceptance criterion before repair: identify the equipment instruction, its peak or Vpp metric, and the required locations and duty matrix. Where no OEM numerical criterion exists, agree a project criterion combining valid contact, suppression of the implicated discharge pattern and verification of the identified return path. A percentage voltage reduction alone is insufficient.

Accept when the equipment-specific criterion is met, implicated discharge behavior is suppressed in the declared windows and accessible adjacent-bearing exposure has been checked. Record an untested required path as incomplete. Repeat invalid captures after correcting contact, attenuation, references or operating differences; reopen the loop diagnosis when discharge persists with a sound installation.

EST operating-verification guidance; Fluke test method.

What should the before-and-after waveform record contain?

Complete one paired record for each shaft end and duty point. Record the measured values in the fields below and attach native data, screenshots and setup photos. “Not accessed” and “not recorded” remain incomplete coverage.

The field closeout checklist

Archive this package with the MCC commissioning and machine turnover records.

• Resolve missing motor, drive, bearing and coupled-equipment IDs.
• Preserve bearing orientation, grease and pre-cleaning photographs.
• Record lubrication, alignment, seal and failure history.
• Draw the suspected loop and mark the measured branch.
• Isolate for setup; secure guarding, probe and leads.
• Verify conductive shaft contact and the local frame reference.
• Check attenuation, compensation, bandwidth and acquisition settings.
• Record warm operating conditions and drive settings.
• Save broad and edge-detail traces; keep the simulator off.
• Record observation time, acquired duration and known dead time.
• Identify current-sensor path, ratio, range, bandwidth and polarity.
• Record remedy hardware, bearing work and instructions followed.
• Repeat the operating matrix and adjacent-path checks.
• Apply the agreed criterion; repeat invalid captures and flag incomplete coverage.
• Archive traces, photos, disposition and next inspection trigger.

Worked example: what does the published compressor remedy show?

Helwig Carbon's CSE-414 compressor case reports a substantial shaft-voltage decrease and no subsequent bearing trouble at its reported follow-up after a combined bearing replacement and grounding-kit repair. The manufacturer supplies a damaged-bearing photograph, instrument readings and a later oscilloscope image.

Check the reduction

Method: subtract the upper bound of the final reading from the baseline, divide by the baseline, and multiply by 100. Anvilfield estimates the percentage bound from (6.3 − 1) ÷ 6.3 × 100, rounded down to a whole percent; the actual final reading was below the value used in that calculation.

Keep the source's immediate peak-voltage comparison and later peak-to-peak reading in separate dated records. Add operating conditions, contact checks and current-path evidence using the paired field form.

Helwig Carbon, Saving Compressor Motor Bearings, CSE-414.

How much time should diagnosis and verification take?

Anvilfield estimates 2 to 8 elapsed hours, or 4 to 16 person-hours for a two-person team, for an accessible-motor diagnostic assignment including baseline testing, matched retesting and reporting. Use the larger allowance for more duty points or current-path investigation. Schedule remedy installation, bearing replacement and extended shutdown waiting separately.

Method: these November 5, 2026 Anvilfield estimates multiply assigned people by elapsed hours. The team size and task durations are planning inputs. Add time for guard modifications, warm-up beyond the visit allowance, process transitions, travel and repeat mobilization; an internal bearing-current test needs its own engineered scope.

Build the work order around the evidence package

Assign baseline diagnosis, remedy installation and matched verification as explicit tasks. Identify who supplies operating access, who changes bearings or filters, and who accepts the final electrical record. Keep production outage time separate from technician time.

Manufacturing service programs group motors by failure consequence, repair history, drive conversion and practical shaft access, then obtain the equipment inventory from the operator. Service managers use PorchRocket Ground Control to select manufacturing buildings by business type and building size when defining a motor-diagnostics territory.

How often should the remedy be rechecked?

Retest at correction commissioning; for a critical-motor program, Anvilfield estimates a first follow-up at 30 days and quarterly reviews until stable site history supports another interval. These are program planning intervals, adjusted to the installed equipment instructions and failure consequence. Method: the first month checks service exposure; quarterly reviews build a first-year record alongside the condition route.

Retest early after motor/bearing replacement, drive/cable/filter or coupling changes, disturbed grounding hardware, contamination or new vibration/temperature trends. Compare archived traces and teardown findings when failures recur.

AEGIS Handbook edition 4 (2024), p. 14, describes the ring as requiring no maintenance while requiring a conductive shaft and recommending a new ring when bearings are replaced. Helwig's GDE-009/2023 guide also describes its BPK as maintenance-free. Inspect conductivity and attachment when the condition route or service work exposes the assembly, following the installed product's instructions.

AEGIS maintenance guidance; Helwig Pocket Guide.

Sources, review date and how to cite this page

By Anvilfield. Reviewed November 5, 2026. Sources checked: manufacturer manuals, damage references and the official OSHA texts below; ISO scope and status checked through its public page. Handbook edition 4 governs the AEGIS comparison; edition 3 is retained as a cross-check.

1. S1: ABB, Technical guide No. 5, Rev C, April 27, 2011, in Technical Guide Book Rev I (2016).
2. S2: ABB, IEC low-voltage Process performance motors, 9AKK105944 EN 10-2025, pp. 15, 55, 61.
3. S3: Fluke, MDA-550/MDA-510 Users Manual, September 2018, pp. 21 to 23, 33.
4. S4: ISO 15243:2017, edition 2, public abstract and status; confirmed 2022.
5. S5: SKF, Bearing maintenance handbook, PUB SR/P7 10001/1 EN, September 2011, pp. 319 to 322.
6. S6: NSK, Electrical Erosion, undated manufacturer page.
7. S7: NSK, False Brinelling, undated manufacturer page.
8. S8: NSK, Handling Instructions for Bearings, copyright 2010; document reference HIB/C/E/07.21.
9. S9: Electro Static Technology, Electrical Bearing Protection Q&A, Part 16, October 5, 2021.
10. S10: Helwig Carbon, BPK-Probe V2 Operating Guide, undated, seven pages.
11. S11: Electro Static Technology, Industrial Bearing Protection Handbook, edition 4, 2024; manufacturer PDF hosted by HSA.
12. S12: Fluke, Measuring motor shaft voltage discharges with Fluke MDA-550, undated application note.
13. S13: ASHRAE Transactions 114(2), SL-08-025, 2008, Shaft Grounding: A Solution to Motor Bearing Currents; manufacturer-authored paper.
14. S14: Electro Static Technology, AEGIS Rings' Resistance Isn't Everything, January 9, 2024.
15. S15: OSHA, 29 CFR 1910.333, July 1, 2025 annual CFR edition, official text.
16. S16: Helwig Carbon, Saving Compressor Motor Bearings, CSE-414, undated; February 2020 follow-up image.
17. S17: Helwig Carbon, BPK Pocket Guide, GDE-009/2023.
18. S18: Electro Static Technology, AEGIS Shaft Voltage Testing, handbook edition 3 extract, copyright 2018; cross-check.
19. S19: OSHA, 29 CFR 1910.147, July 1, 2025 annual CFR edition, especially (f)(1), testing or positioning.
20. S20: OSHA, 29 CFR 1910.212, July 1, 2025 annual CFR edition, (a)(1), machine guarding.
21. S21: Tektronix, Floating Oscilloscope Measurements and Operator Protection, 51W-10640-1, November 1998.

Update log, 2026-11-05: Rechecked primary sources; expanded isolation and guarded testing; clarified damage discrimination and manufacturer selection; added paired waveform fields; consolidated the compressor calculation and labor-hour planning.

Citable units: current-path diagram; probe bandwidth checklist; bearing damage differential; remedy comparison; post-remedy waveform record. Cite the unit name and reviewed date with the page citation.

How to cite this page: Anvilfield, Are VFD bearing currents causing fluting? Measure and verify the remedy, 2026, https://anvilfield.com/field-guides/electrical/vfd-bearing-current-fluting-diagnosis/

Related tools

FAQ

People also ask

• How often should you grease a motor?
• What does a VFD do?
• What does an infrared inspection find?
• What is a motor control center?