Air Compressor Electrical Requirements Including Voltage Phase Amperage and Wiring

The pump always gets blamed first. Valves, rings, gaskets, head bolts. A mechanic will tear a pump down twice before anyone puts a meter on the incoming power.

230V on a motor nameplate means the stator was wound so that at 230V, flux density in the lamination stack sits just below the saturation knee of the core steel's B-H curve. This is where the motor produces rated torque at rated efficiency and where all the thermal testing was done for the UL or CSA listing.

Drop to 208V and everything shifts. Flux density drops roughly ten percent. Slip increases to maintain torque against the load. Stator current rises, and it rises by more than ten percent because the motor is also losing efficiency as it moves off its design point on the torque-speed curve. The winding runs hotter.

How much hotter depends on so many variables that quoting a single number is irresponsible, and yet the number matters, so: EASA's technical documentation and IEEE Std 117 both describe the resistance-rise method for measuring winding temperature under load. Published test data from motor manufacturers' thermal certifications show a wide spread depending on frame size, enclosure type, ambient, and loading fraction. On a TEFC 184T frame at full load and 35°C ambient the penalty from running at 208V versus 230V can eat most or all of the thermal margin between operating temperature and the insulation class limit. On an ODP motor in a cool ventilated shop the same voltage deficit might leave adequate margin. The point is that 208V operation does not produce a fixed, predictable temperature penalty across all motors. It produces a penalty that interacts with every other thermal factor in the installation, and on a bad day the sum exceeds the insulation rating.

L = L₀ × e^(Ea/k × (1/T₂ - 1/T₁))

Ea for polyester and polyesterimide insulation films is 0.8 to 1.2 eV per EASA's 2003 technical handbook on motor repair effects on efficiency. k is Boltzmann's constant. The "ten degree half-life" rule that shows up in every motor textbook is a linearization of this curve. It is roughly correct for small increments near the class rating. Over larger temperature spans the exponential steepens. A twenty-degree sustained increase does not halve insulation life twice; it roughly quarters it in one step because the curve is not linear. The implications for a compressor motor running 208V on a 230V winding for years should be obvious.

208V comes from wye-wound service transformers in commercial buildings. 120V line-to-neutral times root-three. The installer connects the 230V compressor, starts it, watches the gauge climb, checks for leaks, leaves.

The thermal overload trips for the first time maybe four months later on a hot afternoon.

Gets reset. Trips a few more times over the summer. Someone replaces the overload with a higher-rated one, which is exactly the wrong response but a completely understandable one because the compressor ran fine for months and the overload is the thing that is acting up. The motor burns the following summer. Warranty claim says defective motor. The voltage was never measured.

A motor stamped 208-230V is wound differently. More turns of different gauge magnet wire. Same frame, same shaft, same HP. Different motor inside.

NEC 2023 Article 430.110(A) addresses motor controller horsepower rating. It does not require verification of supply voltage against motor voltage rating. Measure at the disconnect with a true-RMS multimeter under load before ordering the compressor.

Electrolytic start capacitors. This is where the article is going to spend a lot of time because the failure sequence is subtle, slow, self-reinforcing, and poorly understood even by people who service compressors regularly. The rest of the topics in this article are comparatively simple. Breakers get sized per a table. Wire gets sized per a table and a voltage drop calculation. Grounding follows a recipe. Capacitor degradation on single-phase compressor motors is not a recipe. It is a process that unfolds over years with almost no visible indication until the end stage.

A single-phase supply produces a pulsating magnetic field. No rotation. Zero starting torque on a squirrel-cage rotor. The start winding, physically offset in the stator slots from the main winding, carries current that is phase-shifted by a series capacitor. The combination produces an elliptical approximation of a rotating field.

The start capacitor is an electrolytic device. Etched aluminum foil, liquid electrolyte, rolled into a cylinder in a Bakelite shell. The etching increases surface area to achieve high capacitance in a small package. The electrolyte is what makes it work and the electrolyte is what eventually kills it.

Electrolyte evaporates through the end seals. This happens even on a shelf. In service, the high-current pulse at every motor start heats the capacitor internally and accelerates the evaporation. The evaporation rate is not constant over the life of the capacitor, and this nonlinearity is the part that makes the failure mode so insidious.

Early in life the electrolyte reservoir is full. The liquid level is above the foil edges. The vapor pressure differential across the seals is low relative to the total liquid volume and the loss rate, measured as percent of remaining electrolyte per year, is small. The capacitor might lose a few percent of its capacitance per year in this phase. At a starting frequency of ten cycles per day in a busy shop, the capacitor is accumulating maybe 3,500 start-heat-cool cycles per year and each cycle drives off a tiny amount of electrolyte, but the reservoir is deep and the rate of change is slow.

Then the electrolyte level drops below a threshold where the remaining liquid exists as thin films on the foil surfaces instead of a pool at the bottom of the can. The surface-area-to-volume ratio of the remaining electrolyte jumps. Evaporation rate per cycle increases. Capacitance decline accelerates. This is the knee of the curve. Before the knee, the capacitor might go five or six years losing maybe 15 to 20 percent total capacitance. After the knee, it might lose another 20 percent in the next twelve to eighteen months.

What does a capacitor at 75 percent of rated value do to the starting event? The phase shift between main winding current and start winding current decreases. The elliptical rotating field becomes more eccentric. Starting torque drops. The rotor takes longer to accelerate from standstill to the centrifugal switch cutout speed, which is roughly 75 percent of synchronous speed on most motors.

The start winding was designed for heavy current at near-LRA levels for a duty cycle of about 1.2 to 1.5 seconds per start on a properly functioning motor with a good capacitor. Extend the start to two seconds and the thermal energy deposited in the start winding per cycle goes up. Not enough to damage anything on one start. Enough to matter over two thousand starts. Extend to 2.5 seconds and it matters over a thousand starts. Extend to three seconds and the start winding insulation is taking thermal hits that exceed its design duty every single time the compressor cycles.

Meanwhile the capacitor degradation is accelerating because the longer starts heat the capacitor more per cycle, driving off more electrolyte, dropping capacitance further.

The compressor starts every time through this entire decline. The lights might dim a beat longer than they used to. The startup might have a slightly heavier sound. In a working shop with ambient noise, a radio, a welder, other equipment, there is no practical way to notice a quarter-second change in startup duration without instrumentation.

By the time the motor is visibly struggling to start, the capacitor might be at 50 or 55 percent of its original capacitance and the start winding insulation has taken cumulative thermal damage from a year or more of extended-inrush starts. Replacing the capacitor at this point sometimes restores clean starting. And then the start winding fails weeks later because the insulation damage was already done and the new capacitor did not undo it.

Capacitance test. Pull the leads off the cap. Put a meter across the terminals. Read the number. Compare to the case marking. Two minutes. Replace below 80 percent.

Centrifugal switch. Contacts weld shut from accumulated arc erosion: start winding stays energized continuously, burns in minutes, sometimes with smoke visible from the motor bell housing. Spring mechanism sticks open: motor hums at standstill drawing LRA. Behind the rear bell housing on most NEMA frames.

Run capacitors are oil-filled, rated for continuous duty, degrade over years to decades. Not the same concern.

Three phases, rotating field, no start cap, no centrifugal switch.

Single-phasing: lose one of three supply phases, motor stays running on two, current on the surviving phases jumps, heating goes as I-squared. Thermal overload should trip. If the overload was upsized at some point, it might not trip fast enough. Phase-loss relays catch it in electrical cycles. Compressors do not ship with them.

VFD output is PWM. Common-mode voltage from IGBT switching couples capacitively through stator insulation to the rotor shaft. Shaft voltage exceeds the dielectric breakdown of the bearing grease film. Discharge through the ball-to-race contact. Cumulative erosion at kilohertz repetition rates grooves the inner race. Bearing whines, then seizes.

Shaft grounding rings, insulated bearings, or common-mode chokes prevent it. None ship standard with a VFD.

Documented in IEEE and EPRI literature. The information exists. It sits in application notes and technical papers. The VFD application guide that ships with the drive covers motor compatibility, cable length limits, carrier frequency selection. Bearing protection gets a mention in passing. The person buying a VFD to run a compressor motor in a fabrication shop or an auto body shop does not read IEEE papers. The VFD gets installed, the compressor runs well, and somewhere between six and eighteen months later a bearing fails with fluting marks on the race that nobody looks for because the bearing gets thrown in the trash.

NEC 2023 Table 430.52. Inverse-time breakers: ceiling 250 percent of FLA, start at 175 percent. Breaker protects the conductor. Thermal overload protects the motor.

Magnetic trip thresholds vary between breaker product lines. Motor-circuit or HACR-listed breakers are designed around the inrush current profile that motors produce.

Running current creates modest voltage drop on a long conductor run. Startup current at five to seven times running creates proportionally worse drop. On a 150-foot run, wire sized to the ampacity minimum from NEC Table 310.16 may produce voltage drop at startup that collapses motor terminal voltage below the point where the motor can complete its start. Calculate drop at startup current and size the conductor for the run length. On runs over 80 feet, go at least two AWG sizes above the ampacity minimum. The code addresses voltage drop only in informational notes, not as a mandatory limit.

Copper THHN/THWN in EMT or rigid conduit. Flex whip at the compressor. Dedicated insulated EGC inside the conduit. Phase rotation check on three-phase. Rotary screw compressors seize on reverse rotation because rotor profiles are directional.

Aluminum: anti-oxidant compound on every termination, torque to spec, annual retorque. Annual retorque does not happen in practice and the connection loosens because aluminum cold-flows under compression. This section could be ten pages long or it could be one paragraph because it is standard electrical installation practice that applies to any motor load. The compressor-specific electrical issues are the voltage mismatch and starting circuit degradation described above. The wiring is wiring.

Internal compressor wiring from junction box to motor terminals on imported units sometimes uses metric gauge conductors with lower insulation ratings and inconsistent crimp terminals. Branch circuit engineering stops at the junction box. Look inside during commissioning.

Humid climates only. Gulf Coast, Southeast, Pacific Northwest coast, Great Lakes in shoulder seasons.

Motor cools to ambient over a weekend. Moisture condenses on winding surfaces inside the housing. On a motor with healthy insulation the moisture flashes off during the first seconds of Monday's start and nothing happens.

On a motor whose insulation has been thermally weakened over years, the moisture reduces dielectric strength at an already marginal spot. Monday inrush current punches through. Turn-to-turn short. Circulating current in the shorted loop propagates the failure within seconds. EASA classifies this as moisture-assisted dielectric failure.

Strip heaters. 30 to 100 watts. Wired through the starter aux contact to energize when the motor is off. Factory option on NEMA frame motors 5 HP and up, has to be specifically ordered. Aftermarket installation is an hour.

In arid climates this does not apply.

Service factor is 1.15 on most compressor motors. The pump often loads the motor to 108 to 112 percent of nameplate HP at rated discharge pressure. Running current exceeds nameplate FLA accordingly. Clamp the leads under operating conditions. The measured number is what matters for conductor sizing and for understanding how much thermal margin remains in the motor, which connects back to the 208V discussion and the insulation life discussion and the capacitor discussion. Everything in this article is about winding temperature. Voltage affects it. Starting circuit condition affects it. Loading affects it. Ambient and enclosure and altitude affect it. The insulation has a budget and every thermal input draws from it. The budget is invisible and the inputs are cumulative.

— end —