The symptoms started three weeks after commissioning. A bottling line with four VFD-driven conveyors and two VFD-driven pump systems began throwing problems that defied diagnosis. The proximity sensors on conveyor two reported false triggers, so the reject gate fired on perfectly good product. The analog level transmitter on the fill tank drifted 3 to 5 percent at random intervals, enough to overfill some bottles and underfill others. The PLC link to the HMI dropped out every few hours, forcing the operator to cycle power to restore the display. And at about the six-week mark, the drive-end bearing on one of the pump motors developed a high-pitched whine, far ahead of any reasonable replacement interval.
Two integrators worked the problem. The first spent two days swapping drives, checking parameters, and replacing the proximity sensors. Nothing helped. The second examined the cable routing and found everything correct on paper: motor and control cables properly separated, shielded VFD cable, shield terminated at both ends.
The third person to look was an electrician with twenty years of drive work behind him. He walked straight to the back of the panel, put a multimeter between the drive’s PE terminal and the panel ground bus, and read 4.7 ohms. It should have been less than 0.1. He traced the ground conductor to the sub-plate and found the mounting bolt sitting on top of a freshly painted surface. He scraped the paint to bare metal, retorqued the bolt, and measured again: 0.02 ohms. Within a day, the false triggers stopped, the analog drift disappeared, and the communication dropouts ceased.
The ground was connected the whole time. It just was not there.
What Paint Does at PWM Frequencies
The back panel of a VFD enclosure does two jobs, and the second one is easy to overlook. It is the mechanical mounting surface for every component, and it is the primary electrical bonding surface that puts every metallic part at a common ground potential. Every drive, contactor, DIN rail, terminal block, and wireway bolted to that panel depends on metal-to-metal contact for its ground reference.
Paint is an insulator. A powder-coated back panel can measure hundreds or even thousands of ohms at a connection point where the mounting bolt lands on the coating. At 60 Hz, that barely matters. But a VFD’s output is not 60 Hz. The PWM carrier runs at 2 to 16 kHz, and the switching transients contain frequency components far higher still. Every switching event pushes a pulse of common-mode current through the system’s parasitic capacitances, and that current must return to the drive’s DC bus to complete the circuit.
When a layer of paint blocks the intended path, the current does not stop. It finds its own way home through whatever conductor is available: signal cables, encoder housings, bearing races, adjacent equipment frames. That is the whole mystery of the bottling line in one sentence. The sensor false triggers, the analog drift, the dropped communications, and the whining bearing were the same fault expressed through four different detours. Radiated interference from a poorly grounded drive can reach equipment tens of meters away, which is precisely the noise the European EMC limits in EN 61800-3 were written to contain. If you are chasing symptoms like these, the VFD Troubleshooting Guide walks the diagnostic sequence, but the fix almost always lands back here.
The Two Clean Solutions
The first solution is the one to specify before the panel is ever built: galvanized steel for the back panel and sub-plates instead of painted. Galvanized steel keeps its electrical conductivity at every bolted connection with no surface preparation at all. It costs about the same as a painted panel, and it deletes an entire category of grounding problems before the first component is mounted. The full context for that decision lives in the VFD Installation Guide.
The second solution is for the panel you did not get to specify, because the customer required painted, a corporate standard mandated it, or the panel was already purchased. In that case the paint must be scraped to bare metal at every single connection point before the component is mounted. That means the area under every drive mounting foot, every DIN rail bracket, every wireway support, every ground lug, and every terminal block standoff. Rockwell Automation’s industrial wiring and grounding guidelines make the same demand: scrape any non-conductive coating away around the mounting hole and put a star washer at the connection. The washer helps cut through residual coating, but it is not a substitute for proper paint removal. The washer is insurance. The bare metal is the connection.
Bonding Sub-Plates and Doors Into One Sheet
Large enclosures often carry two or more sub-plates joined together, and every joint is a potential break in the ground plane. Sub-plates bolted together through their paint, or through a single narrow bolt, can present enough high-frequency impedance at the joint to compromise the bonding across the entire panel.
The fix is braided copper or tinned-copper straps at each joint, with the paint removed at both connection points and the strap secured with star washers or serrated-flange bolts. Braid earns its place through skin effect. At 1 MHz, current penetrates copper to a depth of only about 0.066 mm, so a round wire conducts high-frequency current only in a thin shell at its surface while the copper in the center rides along as dead weight. A braided strap of the same cross-section is woven from many fine wires, each contributing its own surface, which gives it dramatically lower impedance at exactly the frequencies a drive generates. The goal is to make a multi-piece sub-plate behave electrically like one continuous sheet of metal.
The same logic covers the enclosure door. Hinges do not provide a reliable ground path, because the hinge pin is a high-impedance rotational contact. A braided bonding jumper from door to frame keeps the door at enclosure potential regardless of hinge condition, and UL 508A specifically calls for it.
The Ground Bus and the No-Daisy-Chain Rule
The ground bus is the backbone of the panel’s grounding system: a single solid copper or tinned-copper bar, mounted directly to the galvanized or bare-metal panel with paint removed at every mounting point. How conductors reach that bus matters as much as the bus itself.
The cardinal rule is no daisy chains. Running a ground wire from the first drive to the second, the second to the third, and then to the bus creates a shared conductor where impedance accumulates at every hop and one drive’s high-frequency noise appears as a voltage on its neighbors’ ground reference. Worse, one loose or corroded connection strands every drive downstream of it. Each drive gets its own dedicated conductor running directly from its PE terminal to the bus, a star topology, sometimes called home runs.
Size those ground conductors the same as the power conductors they parallel. The NEC equipment grounding tables answer a different question, sizing for 60 Hz fault duty. A VFD ground conductor carries high-frequency common-mode current continuously during normal operation, and skin effect shrinks its effective cross-section at those frequencies. If the motor cable runs 6 AWG phases, the ground is 6 AWG.
Verification: Numbers, Not Appearances
A panel can look immaculate and still be a ghost factory. The bottling line panel would have failed a thirty-second meter check on day one, so make the meter check routine. Measure resistance between every ground connection point and the main ground bus: every reading should be less than 0.1 ohms. At the system level, the transformer’s Wye neutral should be solidly grounded with less than 1 ohm of ground resistance.
A note to the reader on that last figure. Less than 1 ohm is the stricter number, the drive-system practice target, and it is the one to hold a new VFD installation to. The looser number is the code minimum: the NEC accepts a single ground rod at 25 ohms before requiring a supplemental electrode. Both are real, and they answer different questions. “It meets code” and “it is quiet enough for a drive system” are two different sentences.
The full grounding verification runs seven checks: the system ground at the transformer, the panel ground plane (galvanized or scraped, sub-plates strapped), a dedicated ground conductor from each drive to the bus, motor grounding with the cable shield terminated 360 degrees at both ends, PE terminals positioned so input ground sits near input wiring and motor ground near output wiring, potential equalization between motor and driven equipment for motors 100 kW and above, and grounding on every surge suppressor. Run the sequence before energization and the phantom problems never get their six-week head start.
The Foundation Under Everything
Grounding is the single most important factor in a successful VFD installation, and if you get it wrong, every other best practice becomes irrelevant. That is what the twenty-year electrician told the maintenance manager at the bottling plant, and thirty years of drive work has not produced a counterexample. Get the back panel right and the rest of the installation has a foundation; get it wrong and you have signed up for phantom problems that show up one detour at a time.
The panel-level practices here sit inside the larger discipline covered in the VFD Installation Guide, and keeping the ground system healthy over the life of the drive belongs to the VFD Maintenance and Reliability Guide. The complete treatment, including the field-form grounding checklist with space to record your measured values, is in Before the First Fault, and the companion VFD training course puts a meter in your hand and makes you prove it.
Author: Dr. Carl Lee Tolbert, PhD, CMRP, Wayward Leaders LLC, waywardleaders.com
Frequently Asked Questions
Why do painted back panels cause VFD problems?
Paint is an insulator, so every component bolted to a painted panel has a nonconductive layer between it and the ground plane. The high-frequency common-mode current from PWM switching cannot cross that layer, so it detours through signal cables, bearings, and adjacent equipment, producing sensor faults, analog drift, communication dropouts, and premature bearing failure.
Do star washers make a painted panel acceptable?
No. A star washer helps cut through residual coating and maintain contact, but it is not a substitute for scraping the paint to bare metal at the connection point. Use both: bare metal first, star washer as insurance.
What resistance should I measure between a drive's PE terminal and the ground bus?
Less than 0.1 ohms. A reading in whole ohms means a compromised connection, and paint under a mounting bolt is the first suspect. The bottling line case in this article read 4.7 ohms before the fix and 0.02 ohms after.
Is a galvanized back panel more expensive than a painted one?
No, it costs about the same. The difference is that galvanized steel stays conductive at every bolted connection without surface preparation, which eliminates the entire category of paint- related grounding failures before the first component is mounted.
Why must the VFD ground conductor match the power conductor size?
Because it carries high-frequency common-mode current continuously, not just fault current. Skin effect reduces a conductor's effective cross-section at high frequencies, so the standard NEC sizing tables, written for 60 Hz fault duty, undersize it for drive service.
Can I daisy chain ground wires between drives in the same panel?
No. A daisy chain shares one conductor among several drives, so each drive's noise appears on its neighbors' ground reference, and a single loose connection strands every drive downstream. Run a dedicated ground conductor from each drive's PE terminal directly to the panel ground bus.