Field note

VFD Cable: How to Choose It and Terminate the Shield

The most common thing I find wrong on a drive job is not the drive and not the parameters. It is the 3 or 4 inches at each end of the motor cable where somebody took a shielded cable, twisted the shield into a pigtail and landed it on the nearest ground lug, or cut it back and taped it. The cable cost a lot more than building wire, and with that termination it is doing about what building wire would do.

This note is the decision on what cable to run between a drive and a motor, and the procedure for terminating it so the money spent on it shows up as fewer faults, quieter instruments and longer bearing life.

Why a drive output needs different cable

A drive switches the full DC bus onto the motor leads in a fraction of a microsecond, thousands of times a second. Each of those edges does 3 things that 60 Hz power never did. It charges the cable’s capacitance, and on a small drive that charging current can be a big enough share of the output to trip it on overcurrent or ground fault. It reflects off the motor terminals and can nearly double the voltage the winding sees. And it pushes high frequency common mode current out through the motor frame, the bearings and the building ground looking for a way back to the drive.

Purpose built VFD cable handles 2 of those 3 well and helps with the third. It uses XLPE insulation with low capacitance and a high corona inception voltage. It lays the conductors symmetrically so the fields balance. And it wraps everything in a continuous shield that gives the common mode current a low impedance path home, as long as that shield is bonded properly at both ends. What it does not do is stop the reflected wave. The peak at the motor depends on the drive, the run length and the motor, and a long run can still need an output reactor or filter on the best cable made.

Choosing the cable

ABB frames the choice as tolerable, better and best, and that is the honest way to put it.

OptionInsulationWhat you getRating
THHN in conduitThin PVCLowest cost, highest capacitance (up to 10 times VFD cable), absorbs moisture, deforms under pull pressureTolerable
XHHW-2 in conduitXLPE, thicker wallHigher corona inception voltage, resists moisture and cold flow, small premium over THHNBetter
VFD cableXLPE with continuous shieldLowest capacitance, symmetrical construction, a controlled return path for common mode currentBest

The insulation is the reason. Lutze puts the corona inception voltage of PVC at only 70 to 80 percent of XLPE’s, and moisture can cut PVC’s in half while barely moving XLPE’s. If the budget will not stretch to VFD cable, specify XHHW-2 and not THHN. The price gap is small and the first service call on a THHN run that has broken down inside a wet conduit costs more than the difference.

When it is VFD cable, these are the lines to put in the spec:

  • Standard: VFD cable per ANSI/ICEA S-138-738, the first U.S. standard written for cable between drives and motors, published in 2024.
  • Construction: 3+3 symmetrical preferred, or 4 conductor with an insulated ground.
  • Shield: a continuous metallic shield built for drive output, such as foil plus copper braid at 85 percent coverage or more, overlapping copper tape, or a continuously welded metal armor the cable maker rates for the job.
  • Voltage: 600 V minimum. 2,000 V cable is available for long runs and harsh locations.
  • Temperature: 90 degrees C minimum.
  • Termination: 360 degree EMC gland or shield clamp at both ends, per the cable maker’s instructions.

The 3+3 preference comes from how the jobs actually go. In 3+3 cable the 3 bare grounds ride against the shield the whole length of the run, so the shield stays tied to ground along the whole run even when the termination is less than perfect. It still needs terminating properly. It just forgives more. A 4 conductor cable depends completely on the installer. Southwire’s test program put numbers on what a bad termination costs: VFD cable with no shield termination, and plain building wire, both carried roughly 6 to 7 times the high frequency current of a properly terminated VFD cable. An unterminated shield is expensive building wire.

Terminating the shield, step by step

  1. Pick the gland or clamp before the job starts. An EMC gland sized to the cable’s shield diameter at each enclosure entry, or a shield clamp bar inside the drive enclosure. Order them with the cable, because the job that is missing them on install day gets a pigtail.
  2. Strip back only the outer jacket at the clamp point. Leave the shield intact under the gland or clamp. The idea is shield metal pressed against grounded metal all the way around.
  3. Land it 360 degrees at the drive end. Into the EMC gland on a gland plate, or under the shield clamp the drive maker provides. The gland plate or clamp has to be bonded to the drive’s PE with bare metal, which means paint scraped where the plate meets the enclosure. The note on painted back panels is about exactly this.
  4. Land it 360 degrees at the motor end. Into a metallic EMC gland on the motor terminal box, not a plastic one. If the box has a plastic entry, change it. If the cable maker calls for isolating glands instead, follow that, as covered below.
  5. Land the ground conductors. On 3+3 cable, the 3 grounds go together to PE at the drive and to the motor’s ground lug in the terminal box. Do not leave them floating at either end.
  6. Keep the shield path continuous through a disconnect. If there is a disconnect switch at the motor, use VFD cable on both sides of it and carry the shield’s high frequency return path through, using the method the cable and drive makers call for, whether that is an insulated pass through or properly bonded EMC glands. ABB, for one, wants the switch in a metal enclosure, a continuous connection in the shield or conduit from drive to motor, and the shield tied to the protective ground at both the drive and the motor. A shield dumped onto the disconnect box and restarted on the far side leaves the common mode current looking for another way home, and the motor bearings are usually what it finds. Keep that separate from safety grounding: the disconnect enclosure and the motor still get their protective ground per code, and the protective earth conductor is never interrupted to cut noise.
  7. Keep conductor tails short. The phase conductors should leave the shield as close to their terminals as the enclosure allows.
  8. Megger before the drive is connected. Lock out, then disconnect the cable from the drive and from the motor before testing conductor to conductor and conductor to ground, because a megger on drive connected leads can damage the drive. Test the cable and the motor separately, discharge the conductors afterward, and record the readings. The note on megger minimums for 480 V motors covers what the number should be.

Never pigtail the shield. A short pigtail looks like a ground connection to an ohmmeter and like a choke to a fast drive edge. At hundreds of kilohertz, where the shield does its work, a 6 inch pigtail can have impedance orders of magnitude higher than a proper clamp.

Where the cable makers differ

Everyone agrees on the goal: the shield has to carry the common mode current all the way back to the drive, bonded at both ends. Most drive makers ask for it, and Southwire’s test program confirmed it. Southwire also notes that a few drive makers do not, so the drive manual for the drive in front of you comes first.

Where they split is the gland. Southwire calls for a 360 degree EMC gland at the drive and at the motor, bonding the shield to the enclosure metal at both ends. Belden prefers isolating glands, which keep the shield off the enclosure and the machine, because its cables keep the shield in contact with the ground conductors along the whole run, so landing the grounds lands the shield. Belden’s reason is bearings: it says a grounding gland, at a disconnect most of all, can set up a common mode current loop through the motor bearings.

In practice, follow the cable maker’s termination instructions for the cable you bought, and the drive maker’s for the drive. Parts from different makers can work together, but a termination method that does not match the cable’s design is how a good cable ends up terminated badly.

Routing

Run the motor cable on its own, in its own conduit or its own tray section, away from control, analog and network cable. Where a motor cable has to cross signal cable, cross it at a right angle. On armored cable, follow the cable maker’s instructions and the code for bonding the armor. Intermediate bonds along the run can set up ground loops for high frequency current, but the equipment grounding the code requires comes first.

If one drive feeds more than one motor, add up the lead lengths. The drive sees the total, not the run to the nearest motor.

Cable length and the small drive problem

Every motor cable has capacitance, typically 50 to 150 picofarads per foot depending on construction, and it grows with length. Every fast edge from the drive charges it, and those charging pulses add to the drive’s output current. How big they get depends on the cable, the length, the drive’s switching speed and its output impedance, so there is no single amp figure that fits every job. To give a sense of scale, Lutze says even low capacitance VFD cable can draw charging current up to about 0.6 amps per meter, more than 20 amps on a longer run, and it calls this most serious on drives under 10 kW. A 460 V drive draws more of it than a 230 V drive. A 100 HP drive never notices them. On a 3 to 5 HP drive delivering 5 to 8 amps, they can be a real share of the output, and long runs on small drives are where nuisance overcurrent and ground fault trips come from with nothing wrong in the motor.

The drive maker’s cable length table, read at the motor’s actual insulation rating, is the number for a given drive. Rockwell’s tables for the PowerFlex line show how much the motor rating matters: the same drive is allowed 25 feet on a 1,000 V winding and 1,200 feet on a 1,488 V winding. The note on where the 1600 volt number comes from works through that table. When the run is longer than the table allows, the answer is an output reactor or a filter, and inverter duty motors covers how the motor’s rating changes that answer. The reflected wave itself is explained in reflected wave and long motor leads.

What a good cable job buys you

Good cable, properly terminated, is the foundation of bearing protection. When the common mode current has a clean way back to the drive, the shaft voltage often stays low enough that it never arcs through the bearing. It is not a guarantee, and large motors and hard installs can still need a shaft grounding ring, insulated bearings or a common mode filter, but none of those work well on top of a bad cable job, which is the point of a grounding ring cannot outrun a bad install. Cable, grounding and termination are covered in order in the VFD Installation Guide, and the free First Power course on the online training page teaches them hands on.

Frequently Asked Questions

Do I need VFD cable or can I use THHN?

THHN in conduit is tolerated by drive makers, not recommended. Its thin PVC insulation, high capacitance and habit of absorbing moisture all work against it on a drive output. If VFD cable is not in the budget, XHHW-2 is the better fallback for a small extra cost, and purpose built VFD cable with a properly terminated shield is the best choice.

Should a VFD cable shield be grounded at both ends?

Yes, on the motor cable, and most drive makers say so. The shield gives high frequency common mode current a low impedance path back to the drive, and that only works when it is bonded at the drive end and at the motor end. The cable makers differ on how: Southwire calls for a 360 degree EMC gland at both ends, while Belden prefers isolating glands, with the shield bonded through the cable's own ground conductors. A shield bonded at neither end performs about like building wire.

What is the difference between 3+3 and 4 conductor VFD cable?

3+3 cable has 3 phase conductors and 3 bare ground conductors laid symmetrically between them, and the bare grounds touch the shield along its whole length. 4 conductor cable has one insulated ground, so its shield only works if the installer terminates it correctly. 3+3 is the more forgiving construction when the terminations get done late and in a hurry.

Why is a pigtail shield termination bad on a VFD cable?

Twisting the shield into a wire tail and landing it on a lug adds inductance right at the frequencies the shield is supposed to handle, so the high frequency current sees a high impedance and looks for another path home, often through the motor bearings or the building steel. A 360 degree clamp or EMC gland is the fix.

How long can a VFD motor cable be?

It depends on the drive size, the motor's insulation rating and what sits on the drive output. The drive maker's cable length table, read at the motor's actual insulation rating, is the answer for a given drive. Small drives are the most limited, because cable charging current is a bigger share of their rated output.

Online start-up courses, filmed on real drives

One drive per course, filmed uncut with a clock on screen from first power to a saved backup. The install course is free, and the ABB ACS380 and ACS580 start-up course is open now.

See the online courses

Wayward Leaders® is a veteran-owned VFD training practice. We teach maintenance teams and plant electricians to install, commission, and troubleshoot variable frequency drives correctly.

Every class ends with a scored competency assessment. The plant gets a way to prove what each technician can do, not a roster of who sat in the room.

The instruction is field-trained. It is drawn from nearly 8,000 VFD, power, and motor documents and from more than 750 commissioned drives captured through personal field experience and data collection. Carl Lee Tolbert, PhD, CMRP, and ATD Master Trainer™, leads it. He has delivered more than 5,000 hours of VFD instruction across 30 years and trained 8,000 industrial professionals, from International Paper to the U.S. Navy.

The premise is simple: VFDs do not fail. Installations fail them. The curriculum is built backward from that.

Installation, commissioning, troubleshooting, and fault diagnosis guides are published openly at waywardleaders.com.