Can You Run a CNC Machine Off a Phase Converter?

A phase converter takes the two hot legs of a single-phase supply and manufactures a third. A static converter does this with capacitors and a starting circuit. It has no rotating mass, so the manufactured leg sags the moment a motor pulls hard. A rotary converter couples a real idler motor to the line, and that spinning rotor supplies the third leg with some inertia behind it. A VFD rectifies single-phase AC to DC and inverts it back to three-phase at whatever frequency you set.

The distinction matters because a CNC is not one load. It is a whole cabinet of them. The spindle motor, the axis servo drives, the coolant pump, the tool changer, and the control's own power supply all sit on the same bus, and they draw current in completely different ways.

An induction motor is forgiving. It will run on a slightly unbalanced three-phase supply, with a bit more heat and a bit less torque. A servo drive is not forgiving. Its rectifier and DC bus expect a reasonably symmetric input, and the drive will trip on phase loss or bus ripple long before a plain motor would notice anything wrong.

• 1
  Rotary converterSpinning idler motor supplies the third leg; good inertia, poor voltage balance under load.
• 2
  Static converterCapacitors only; cheap, no rotating mass, sags badly on spindle ramp-up.
• 3
  VFDRectifier plus inverter; clean output, but sizing and derating rules are strict.

The hardest moment for any converter is spindle acceleration. A spindle motor that ramps from zero to 12,000 rpm in a second or two can pull several times its nameplate current for that second. On a rotary converter the manufactured leg dips during the ramp, and the drive sees an unbalanced input. On a static converter the dip is worse, and on a badly sized rotary it can drop enough to fault the drive.

Rigid tapping is the other common failure point. The spindle reverses direction under load, synchronised with the Z axis, several times a second on small threads. Each reversal is a current transient. Converters that handle steady milling fine will sometimes trip mid-thread, and a broken tap in a finished part is an expensive way to learn that.

Regenerative braking is the third problem. When a large spindle decelerates quickly, the motor becomes a generator and pushes energy back into the DC bus. On a proper three-phase supply the drive has headroom to absorb this. Behind a converter the bus can rise until the drive trips on overvoltage. Some builders add a braking resistor; if yours has one, it needs to be sized for the real decel profile, not the catalogue default.

Voltage balance is the number that decides most of this. Measure line-to-line voltage on all three pairs at the machine terminals, at idle and again with the spindle at full speed in a cut. A rotary converter that reads 2% imbalance at idle can read 8% or 10% under load. Most CNC builders want imbalance under 2%, and some servo drives fault above 4%.

Then check frequency stability and waveform. A rotary converter's manufactured leg is not a clean sine wave; it carries harmonics from the idler motor. Drives with a passive front end tolerate this better than drives with an active front end, which measure the input directly and will reject a distorted supply. Ask the machine builder which front end your drive has.

Finally, check the neutral and grounding arrangement. Some converters produce a floating or high-impedance neutral that confuses the machine's insulation monitoring. If the control faults on ground leakage with no obvious cause, the converter's neutral is often the reason.

• 1
  Imbalance targetUnder 2% line-to-line at the machine terminals, measured under load.
• 2
  Measure twiceIdle reading alone tells you almost nothing about spindle ramp behaviour.
• 3
  WaveformActive front-end drives are far less tolerant of harmonic distortion.

Rotary converters are usually sized at two to three times the largest motor's horsepower. That sounds generous until you remember the converter also has to carry the coolant pump, the tool changer, and the servo supply at the same time. Add the continuous loads to the spindle rating, then apply the multiplier to the total, not just to the spindle.

A single VFD feeding a whole machine is a different problem. The VFD has to be rated for the total connected load, and most manufacturers require derating when the input is single-phase. A drive sold as 10 kW three-phase input might only be good for 5 or 6 kW on single-phase input, because the rectifier has to carry the full DC bus current through fewer input legs. Read the derating table, not the headline number.

Retrofits often work better with one VFD per axis drive and one for the spindle. Each drive then sees a clean, dedicated three-phase supply, and a fault on one does not disturb the others. It costs more up front and takes cabinet space, but it removes the shared-bus interactions that cause most intermittent faults.

• 1
  Add continuous loadsSize on spindle plus pumps plus servo supply, not spindle alone.
• 2
  Derate for single-phaseMost VFDs lose 40-50% of rated output on a single-phase input.
• 3
  One VFD per driveCleaner supply, easier fault isolation, more cabinet space used.

High-speed spindles are the clearest case against converters. A 20,000 rpm spindle with a fast ramp profile demands a stiff supply. Behind a converter the DC bus sags on every acceleration, and the drive spends its life near a fault threshold. If the machine is doing production work at high spindle speeds, the converter will cost more in downtime than a three-phase service would have cost to install.

Multi-machine shops hit the same wall. One rotary converter shared across two or three machines works until two of them start at once. Then the manufactured leg collapses, both drives fault, and the operator learns to stagger starts by hand. That is not a process you can run unattended.

If the building already has three-phase at the service entrance and the issue is just the run to the machine, the money goes further on a proper feeder and a transformer than on a converter. Converters make sense for one machine, modest spindle power, and a shop that genuinely cannot get three-phase.

• 1
  High-speed spindlesFast ramps need supply stiffness a converter cannot provide.
• 2
  Shared convertersSimultaneous starts collapse the manufactured leg.
• 3
  Existing three-phaseA feeder and transformer usually beat a converter on cost per machine.