VFD Cable Sizing and Motor Cable Length Limits

What is VFD cable sizing? Motor cable between a variable frequency drive and motor must satisfy two independent requirements: conductor cross-section rated for the motor's full-load current per NEC Table 310.16 or IEC 60364-5-52, and a length short enough to keep PWM reflected-wave voltage at the motor terminals below the winding insulation rating referenced in NEMA MG-1 Part 30/31. Meet only the first requirement and a correctly ampacity-rated cable can still deliver voltage spikes near twice the DC bus voltage to the motor, degrading winding insulation over months rather than tripping a breaker on day one. This article covers ampacity sizing, reflected-wave theory, critical length calculation, voltage-class effects, shielded cable selection, and what to do once a run exceeds the rated limit.

Two Sizing Problems, Not One

Engineers used to sizing motor feeders for a fixed-speed contactor often size VFD cable the same way: pick a cross-section from an ampacity table and move on. That covers thermal capacity. It says nothing about the electrical stress a PWM inverter imposes on the same cable and on the motor winding at its far end. Two constraints, two failure modes, two different reference standards.

Undersized ampacity causes conductor overheating and, eventually, insulation breakdown from heat — a slow, predictable failure driven by current. Reflected-wave overvoltage causes partial-discharge damage inside the motor winding, a failure driven by cable length and IGBT switching speed, not by current at all. A cable run can pass one check and fail the other without any obvious symptom until the motor trips or fails early.

Ampacity: The Familiar Half

Conductor cross-section for the drive-to-motor run follows standard motor-circuit logic: full-load amps (FLA) from the motor nameplate, multiplied by 1.25 per NEC 430.22 (or the equivalent continuous-duty factor in IEC 60364-5-52), then matched to a conductor size at the applicable ambient temperature and installation method — conduit, cable tray, or direct bury. Three-phase VFD output is not a sine wave; it is a PWM pulse train. For ampacity purposes, though, RMS current is what heats the conductor, and RMS current calculation does not change because the waveform is chopped. Note that this is conductor sizing between drive and motor, separate from sizing the drive itself to the motor, which is a kW/duty decision made earlier in the same project.

Two adjustments are specific to drive circuits. First, derate for harmonic-rich, non-sinusoidal current when several VFD cables share a tray — triplen harmonics can raise neutral and, depending on routing, phase-conductor heating beyond a simple FLA calculation. Second, size the ground/PE conductor larger than a fixed-speed equivalent would need. High-frequency common-mode current requires a low-impedance return path, and an undersized ground conductor pushes some of that current onto paths it was never meant to take — signal cable shields, structural steel, instrumentation grounds.

Reflected Wave: The Length-Dependent Half

IGBT switching edges in a modern drive have rise times in the 50-400 ns range. At these speeds a motor cable behaves as a transmission line, not as a simple wire. When the cable's electrical length is a meaningful fraction of the voltage rise time, the impedance mismatch between cable and motor reflects part of the pulse back toward the source, and the incident and reflected waves add at the motor terminals. Over roughly 15-50 m, depending on rise time and cable type, that peak can approach twice the DC bus voltage. On a 480 V drive, terminal spikes near 1,600 V are possible on an otherwise healthy, correctly-ampacity-sized circuit. See IEC and NEMA standards for VFDs for where this sits in the broader standards picture.

Below L_crit, the reflected wave arrives back at the drive and decays before the next switching edge, and peak terminal voltage stays close to DC bus level. Above it, reflected and incident waves overlap and peak voltage climbs toward the 2x ceiling. Faster IGBTs — shorter t_r — push L_crit shorter, which is why newer, more efficient drives are sometimes more length-restricted than older, slower-switching ones despite being technically superior on losses.

Why the Limit Changes by Voltage Class and Drive Model

Every drive vendor publishes a maximum cable length, with and without a filter, and the numbers vary by series, not just by brand. A machinery drive built for fast dynamic response (Schneider ATV340, ABB ACS380) typically carries a shorter unfiltered limit than a general-purpose drive optimized for efficiency over dynamics, such as ABB ACS580 or Siemens G120C. Neither family is wrong; they are tuned for different jobs. See the variable frequency drives collection for the range across brands. Higher DC bus voltage — 690 V class versus 400/480 V — raises the absolute spike magnitude for the same cable length, so 690 V installations often need a filter at distances a 400 V system tolerates unfiltered. Ratings context for both voltage classes is in VFD voltage and current ratings.

What we see in the field: published limits assume shielded cable of a specific construction. Swap in unshielded cable or a different conductor geometry and the real propagation velocity shifts, sometimes enough to move a borderline installation from safe to marginal. Vendor tables are a starting point, not a substitute for checking the actual cable type against the datasheet footnote.

Shielded Cable, Grounding, and Bearing Currents

Symmetrical shielded cable (three phase conductors plus PE, wrapped in a copper braid or foil-plus-drain construction, terminated 360 degrees at both ends) is the standard VFD motor cable design for a reason beyond EMC compliance alone. The shield gives common-mode current a low-impedance return path back to the drive instead of through the motor frame, plant ground grid, or bearing races. Left unmanaged, that current path contributes to conducted and radiated emissions covered under harmonics and EMC filter requirements, and it can pit bearing races through electrical discharge machining over time — a failure mode unrelated to motor load or duty cycle.

Terminate the shield at both ends with a 360-degree clamp, not a pigtail. A pigtail — a twisted shield tail crimped to a ring lug — adds inductance that defeats most of the shield's high-frequency benefit. It stops working exactly where it matters most, above a few hundred kHz. Some integrators skip the clamp for convenience and only revisit the decision after an unexplained bearing failure.

When the Cable Run Exceeds the Limit

Three remedies exist once a run exceeds the drive's rated unfiltered length, and they are not interchangeable. A dV/dt reactor slows the voltage rise time, pushing L_crit longer without materially reshaping the waveform — adequate for moderate overruns and the cheaper fix. A sine-wave filter, larger and more expensive, reconstructs something close to a true sine wave at the drive output, eliminating reflected-wave stress regardless of cable length, at the cost of size, price, and a small efficiency penalty. Reducing carrier frequency lowers switching count but does not change rise time, so it does not resolve reflected wave — a common misunderstanding worth stating plainly. Full selection logic between the two real options is in VFD output filters.

Some integrators default to a sine-wave filter on every long run to avoid the length calculation entirely. It works. It is not the cheapest fix when a dV/dt reactor would have covered the actual overrun.

Conclusion

VFD cable sizing is two checks, not one: ampacity for thermal capacity, reflected-wave length for winding insulation. Skipping the second check is the more common mistake, since it produces no immediate fault and the ampacity table looks satisfied. Confirm the drive's rated cable length with and without a filter, use symmetrical shielded cable terminated correctly at both ends, and add a dV/dt reactor or sine-wave filter when the physical run exceeds the datasheet limit. For the broader picture across sizing, ratings, standards, and control methods, see the VFD engineering guide and the variable frequency drive selection checklist.