VFD cable: when it saves the motor, and when you are paying for nothing

What I get asked about VFD cable lands in one of two piles, opposite mistakes. One shop pulls four THHN conductors through conduit for a long run and the motor bearings are chewed up inside a year. The next specs 2000V armored cable with three symmetrical grounds for a 2 HP fan fifteen feet off the drive. Both cost real money. Here is where the line sits, using numbers the drive builders publish.

What the drive is doing to the wire

The inverter does not put a sine wave on the motor leads. It sends DC pulses out of the bus, they run down the cable, hit the motor and bounce back. Rockwell puts that transient stress at up to twice the DC bus voltage between one drive's own output wires.

The voltage does not punch through insulation directly. It ionizes air in the gap between wires that barely touch, the discharge makes ozone, and ozone eats PVC. Carbon tracking, then a failure nobody can point at.

THHN goes first, and Rockwell's own testing says why. Air voids form between the nylon jacket and the PVC underneath. Wall thickness is inconsistent: wire specified at 15 mil measured 10 mil in places. Heat the conductor and the PVC cold-flows where the wire hangs down from a wireway.

Then add water. Damp THHN has a corona inception voltage less than half the same wire dry. That ends the argument about building wire to a drive in a washdown area.

XLPE does none of this. It is why every real VFD cable is built on it.

The bearing current nobody can measure

Reflected wave kills insulation. Common mode current kills bearings. Two mechanisms, folded into the same bucket constantly.

A PWM inverter cannot make its three output voltages sum to zero at every instant. The leftover behaves as a common mode voltage source at the motor winding's star point, and it has to get home: out through stray capacitance in the motor, back through the frame, the grounds and the shield.

ABB puts that band at 50 kHz to 1 MHz and says inductance in the return path can hold the motor frame more than 100 volts above the inverter frame. Ground the shaft through a coupling into a gearbox and part of the current takes the shortcut through the bearing. Peaks run 3 to 20 amps depending on motor size. Each pulse moves a little metal off the ball into the grease. That is EDM, and the fluting on the race at month four is the receipt.

What should change how you plan a job: you cannot check for this. Measuring it takes 10 kHz to 2 MHz of bandwidth, 150 to 200 amps of peak capability and 10 mA RMS resolution, on a turning shaft. Cable and grounding are not a diagnostic. They are the only prevention you get.

Where common mode current goes when it cannot get back through the cable shield.

 

How far you can actually run VFD cable

The length limit people quote, 300 feet or whatever the last job used, is not one number. It moves with the motor's winding insulation and the carrier frequency set in the drive, and it moves hard. Rockwell's limits for one PowerFlex 753 and 755 wall mount drive at 400V, frame 3, 15 kW:

Carrier frequency	Motor winding insulation	Plain cable, nothing added	With a load reactor
2 kHz	1000V	7.6 m (25 ft)	91.4 m (300 ft)
2 kHz	1200V	137.2 m (450 ft)	365.8 m (1200 ft)
2 kHz	1488V	365.8 m (1200 ft)	365.8 m (1200 ft)
4 kHz	1000V	7.6 m (25 ft)	18.3 m (60 ft)
4 kHz	1200V	91.4 m (300 ft)	91.4 m (300 ft)
4 kHz	1488V	152.4 m (500 ft)	365.8 m (1200 ft)

Motor lead length limits by winding insulation and carrier frequency. No VFD cable moves these numbers if the motor insulation is wrong.

Read the two 1488V rows against each other. Same drive, same cable, same motor. Take the carrier from 4 kHz to 2 kHz and the allowed run goes from 500 feet to 1200 feet, for a parameter change that costs nothing. A 1000V motor is capped at 25 feet on plain wire no matter what you spend.

Typical ratings: inverter-duty 1488V, a newer standard motor 1200V, an older one 1000V. To confirm 1488V, ask the motor manufacturer whether it meets NEMA MG1 Part 31.

When VFD cable is genuinely overkill

Three symmetrically placed ground conductors is the design every datasheet leads with, and it works: the geometry cancels induced current instead of shoving it through the shaft.

But Rockwell's own guidance is that a single ground conductor is sufficient up to and including 200 HP (150 kW), with three grounds recommended above that. Most drives in most plants are nowhere near 200 HP. On a 10 HP pump the symmetrical ground set is not buying much over a full-size ground and a properly terminated shield.

Voltage rating goes the same way: peak on the motor leads runs roughly 2.5 times nominal, which is why 600V systems get 2000V cable and why 600V-rated cable covers 480V and below.

Shielded cable is not free electrically either. It carries more capacitance line to line and line to ground than loose conductors in conduit, and the drive supplies that charging current. On small drives that pushes you toward the lead length limit, not away. Rockwell's fixes: drop the carrier to 2 kHz, shorten the run, oversize the drive, or add an output reactor.

Conductor size	Capacitance to shield (pF/ft)	Ampacity at 90 C	Min. bend radius
16 AWG	36.34	10 A	6 in
14 AWG	44.10	15 A	7 in
12 AWG	46.93	20 A	8 in
10 AWG	52.52	30 A	8 in
8 AWG	50.72	55 A	10 in
6 AWG	56.81	75 A	11 in
4 AWG	67.95	95 A	14 in
2 AWG	75.96	130 A	14 in

One 600V XLPE VFD cable line, ampacity per NEC 310.15(B)(16). Capacitance roughly doubles from 16 AWG to 2 AWG, and the drive pays for it.

 

Carrier frequency inside a drive like this PowerFlex 525 changes the allowed VFD cable run more than the cable does.

 

Picking the shield, and what each one costs you

Construction (all with XLPE conductors)	High frequency bonding	Flexibility	Self-supporting	Mechanical protection
Continuous aluminum sheath	Best	Good	Best	Best
Copper tape	Better	Better	Good	Good
Braided wire	Good	Best	Poor	Good
Building wire or cable in conduit	Poor	Varies	Varies	Varies

Shield tradeoffs for VFD cable.

Aluminum sheath is the one you will fight inside a tight panel. Copper tape sits in the middle, which is why most general purpose VFD cable uses it.

Continuous welded armor beats interlocked, though interlocked is fine on shorter runs. Put a PVC jacket over the armor either way: bare armor picking up incidental grounds against building steel undoes the noise containment you paid for.

One to walk away from: four THHN conductors twisted together and wrapped tight in foil, sold as VFD cable. Rockwell will not endorse it: the charging current goes up and its performance against their lead length tables is unknown.

The install that ruins good VFD cable

The shield only works if it is continuous and terminated 360 degrees at both ends. Not a pigtail into a ground lug: a gland or bushing clamping the whole circumference, drive end and motor end. Break the shield outside a shielded enclosure and you have bought an expensive jacket.

Four more that come up constantly:

·       Route power and control cables separately.

·       Three drive output wires per conduit maximum, or the drives couple into each other.

·       Ground conductors sized for full drive ampacity, spaced symmetrically around the phases.

·       Terminator or output filter in the run: its connector insulation has to be XLPE, not PVC.

The right-hand termination is what a VFD cable shield is designed for. The left is how it gets thrown away.

 

FAQ

How long a run before VFD cable stops being optional?

There is no single number, and anyone handing you one is guessing. The limit comes from the motor's winding insulation and your carrier frequency before it comes from the cable. Check your drive's own lead length table first.

Will VFD cable stop bearing fluting by itself?

It cuts the current that causes fluting, it does not zero it. Symmetrical grounds and a low impedance shield give common mode current an easy way home so less goes through the shaft. On a big motor, or a shaft solidly grounded through the driven machine, add insulated bearings, a shaft grounding ring or an output filter.

Is 2000V VFD cable worth it on a 480V drive?

Usually not. Peak runs around 2.5 times nominal, so 600V-rated cable covers 480V and below. Save 2000V for 600V systems and long runs.

Can I reuse VFD cable pulled off a decommissioned machine?

Yes, if the jacket and shield came out intact and it was never kinked past its bend radius, which is where used cable dies. A shield crushed flat in one spot has a hole in it, and that will not show up on a megger.

What I would actually spend the money on

In order: XLPE insulation, a real 360 degree termination at both ends, the right carrier frequency, and only then the exotic construction. Most failures I get called out on were never going to be fixed by better VFD cable. They were baked in by the motor's insulation rating or a pigtailed shield, and the cable took the blame.

When the drive on the far end of that VFD cable finally quits, a tested surplus unit is usually a straight swap at a fraction of new. We keep PowerFlex, Altivar and similar drives on the shelf at IQElectro, so the run you already pulled does not have to change.