VFD Shock: A Bonding Fault Is Likely, Not Motor Logic

A drive runs for about one hour, then hums, hunts during deceleration or stopping, and gives an operator a shock from its enclosure. Current is crossing a boundary it should not cross: the protective bonding path has enough impedance, or is open, so the frame rises above the potential of the surrounding grounded surfaces. The earlier destruction of both the drive and a reactor adds a separate thermal warning. This is heat, not logic; inspect the power circuit before treating the behavior as a tuning problem.

Remove power, apply the site lockout procedure, prove absence of voltage with a properly rated instrument, and keep the equipment out of service until a qualified electrical worker finds the touch-voltage path. A motor insulation-to-ground test that passes does not prove that the drive frame, cabinet door, motor frame, cable shield, and protective-earth conductors form an effective high-frequency bonding network.

Competing fault paths and test priorities

Several mechanisms can produce part of the symptom set, but only an ineffective protective-earth path directly explains a shock from an exposed drive frame or door. Start there. Supply-reference faults, shield termination errors, and grounded feedback signals can then explain damage or hunting without displacing the first priority.

PWM current and frame-voltage mechanism

The inverter output is a rapidly switched waveform. Capacitance exists between the motor windings and motor frame, between conductors and shield, and throughout the drive power stage. Each switching edge drives common-mode current through those capacitances. That current must return to the inverter through the motor-frame bond, cable shield, protective-earth conductors, and cabinet bonding network.

The number that matters is the impedance of that return path at the switching components, not only its low-frequency resistance. A conductor can show continuity on an ordinary meter yet present significant high-frequency impedance because it is long, loosely routed, coiled, interrupted, or connected through small contact areas. Voltage then appears according to V = I × Z. Fast common-mode current multiplied by return-path impedance can elevate the drive or motor frame enough to produce a shock.

Installations with a long or defective protective-earth route have been associated with approximately 50–100 V of high-frequency voltage on a motor frame. That range is a fault clue, not an acceptable touch-voltage limit. The waveform may contain fast edges that an ordinary averaging meter reports poorly, so meter readings must be interpreted together with the bonding inspection and, where necessary, suitable high-frequency measurement equipment.

A separate protective-earth conductor does not substitute for correct shield termination where a screened motor cable is used. The shield provides a low-impedance, distributed high-frequency return when it has broad, continuous termination at both the drive and motor ends. A thin pigtail adds inductance and reduces that benefit. The motor frame must bond back to the drive installation, and the drive frame must bond to the cabinet protective-earth system.

The 347/600 V supply decision

The stated supply is 347/600 V, four-wire. In a correctly referenced system with that nominal relationship, phase-to-phase readings are approximately 600 V and phase-to-neutral readings are approximately 347 V. The installation still has to be measured; the label alone does not prove that the neutral, equipment bond, or supply reference is intact.

A pattern near 600/600/0 V from the three phases to ground signals a fundamentally different supply reference from three balanced phase-to-ground readings. Stop at that point and have the facility distribution, neutral, grounding electrode connection, and bonding arrangement examined. A replacement drive is not a correction for a supply-reference fault.

De-energized inspection procedure

1. Lock out every energy source feeding the drive, reactor, motor, control circuits, and externally powered analog signals. Follow the drive’s documented discharge indication and waiting instructions, then prove absence of voltage.
2. Identify the burned reactor’s actual location from the conductors and drawings. Record whether it is connected on the drive line side or motor side because the two positions experience different waveforms and point to different fault paths.
3. Inspect the failed drive, reactor, disconnect, terminals, cable, and motor junction box for discoloration, cracked insulation, carbon tracking, loose lugs, damaged shield hardware, and signs of localized heating. Replace damaged conductors and terminations rather than cleaning carbonized insulation for reuse.
4. Trace the protective-earth route from the facility grounding point through the cabinet, drive frame, cabinet door, motor cable, and motor frame. Give the door its own flexible bonding conductor; hinges are mechanical parts, not dependable bonding jumpers.
5. Remove paint, oxidation, sealant, and loose hardware from designated bonding surfaces. Reassemble the joints using the equipment’s intended bonding hardware and torque values read from its documentation.
6. Verify continuity between every accessible metal part and the cabinet protective-earth bar. Flex the door and accessible cable sections while observing the reading to expose intermittent bonds.
7. Inspect a screened motor cable for continuous coverage and termination at both ends. Restore broad shield contact at the drive and motor entries and keep the high-frequency return path short and direct.
8. Inspect the motor insulation and motor cable separately with the drive disconnected. Never apply an insulation-resistance test voltage to connected drive electronics.
9. Check analog-command and feedback wiring for unintended connections between signal common and protective earth. Keep protective bonding intact while correcting signal-reference or shielding errors.

Controlled electrical measurements

After the de-energized defects are corrected, a qualified person can perform controlled live measurements under the facility’s electrical-safety program. Use an instrument, leads, category rating, and personal protective equipment suitable for the 600 V circuit. Establish a known-good ground reference independently of the suspect enclosure.

1. With the motor stopped, record all three phase-to-phase, phase-to-neutral, and phase-to-ground input readings. Investigate imbalance or an unstable ground reference before commanding the drive.
2. Measure from known-good ground to the cabinet, drive frame, door, and motor frame. Any unexpected sustained voltage calls for shutdown and correction; a shock has already demonstrated that the installation is unsafe.
3. Run at the lowest practical operating condition and compare frame voltage with the inverter disabled and enabled. Voltage that appears with PWM operation directs attention to the motor cable, motor-frame bond, shield, and high-frequency return path.
4. Observe input current, output current, output frequency, speed command, speed feedback, and drive status. Read the exact parameter names from the installed drive manual rather than substituting another family’s identifiers.
5. Repeat the original slowing and stopping sequence only after frame potential remains controlled. Capture the command and feedback during the event so that mechanical load, unstable feedback, and drive output can be separated.

Hunting and stopping behavior

Hunting can occur when an analog command or feedback signal uses a different ground reference from the drive. Common-mode current flowing through a shared reference adds an unwanted signal to the intended command. An internal PID loop then reacts to that disturbance, changes motor output, and creates the audible hum associated with repeated torque corrections.

Test the grounding system first because changing loop gains cannot make an exposed frame safe. Once bonding is corrected, compare the commanded value with the measured feedback. A moving command points toward signal wiring or the upstream controller. A stable command with oscillating feedback points toward the sensor, signal reference, mechanical process, or loop adjustment. Stable command and feedback with irregular current or frequency shifts attention to power wiring, motor cable integrity, drive diagnostics, and load condition.

Keep control-cable shield practice distinct from motor-cable shielding. A signal shield may follow the control-system design needed to prevent low-frequency ground-loop current, while the motor-cable screen needs a continuous high-frequency return. Never interrupt the protective-earth conductor to solve signal noise.

Repair verification and release criteria

Replacement alone is not a verified repair. The first drive and reactor burned, and the second drive developed symptoms after about one hour. Verification therefore has to cover both immediate electrical integrity and operation long enough to include that observed delay.

1. Record the corrected bonding path, shield terminations, conductor condition, supply readings, and all measured frame-to-ground values.
2. Run the drive through start, steady speed, deceleration, and stop while watching current, command, feedback, output frequency, and temperature indications available from the drive.
3. Inspect for renewed humming, hunting, odor, discoloration, abnormal reactor noise, or rising terminal temperature. Shut down immediately if frame voltage, smoke, arcing, or shock sensation returns.
4. Continue the supervised test beyond the prior approximately one-hour symptom point under a representative load. A short no-load run cannot clear a delayed thermal or connection fault.
5. After isolation and discharge, recheck accessible terminations and bonding joints for heat damage or movement. Release the equipment only when the supply, protective bonding, motor circuit, control signals, and repeated operating sequence all remain stable.

Frequently asked questions