Power Factor Correction (PFC)

Power Factor Correction (PFC) is the front-end circuitry inside a switch-mode power supply that forces the current drawn from the AC line to track the line voltage waveform. Without it, the bulk rectifier and capacitor pull current in short, high peaks at the crest of each half-cycle — current that distorts the line, produces harmonic pollution, and forces the utility and your wiring to carry more apparent power than the real power actually consumed. For a high-draw load like a Bitcoin mining ASIC pulling 3kW or more, a poor power factor wastes capacity on every branch circuit it shares, which is why every serious mining PSU puts PFC first in the conversion chain.

Active vs. passive PFC

Passive PFC uses bulk line-frequency inductors to smear the current peaks and is rarely seen in modern high-density supplies — the magnetics are heavy, and the correction is mediocre. Active PFC is the standard in ASIC and server PSUs: a boost pre-converter switches at high frequency, and its controller modulates the switch so the average input current stays in phase with, and shaped like, the line voltage. The stage simultaneously produces a regulated high-voltage DC bus for the downstream converter. A well-designed active PFC front end reaches near-unity power factor with single-digit total harmonic distortion, and because the boost stage regulates its output regardless of input, it is also what gives many supplies their wide universal-input range.

Inside the APW12

Bitmain's APW12 makes a good concrete example. It takes two independent AC inputs, each feeding its own EMI filter and its own PFC stage, and the two PFC outputs charge large bulk capacitors to a DC bus of 410-420V. That bus then feeds the isolated LLC resonant converter that steps down to the 12-15V hashboard rail. On the repair bench, the PFC bus is the first waypoint in dead-supply diagnosis: measuring DC 410-420V across the bulk caps tells you the EMI filters, rectification, and both PFC stages are alive, so the fault lies downstream in the LLC or output stages — while a missing or sagging bus points you at the front end. (Standard high-voltage precautions apply: those capacitors store lethal energy well after unplugging, so verify discharge before touching anything.)

Why miners should care

A power factor near 1.0 means the breaker rating you paid for is fully usable: amps drawn match watts consumed, fewer machines trip a shared circuit, and the facility presents a clean load to the grid. That matters at home — a 15A/120V circuit has little headroom to waste on reactive current — and it matters at scale, where utility demand charges and three-phase balancing calculations key off power factor and harmonics. Poor power factor also heats neutral conductors and transformers with current that does no work. When you compare PSUs, the power-factor and THD specifications are quiet indicators of front-end quality; a supply that corrects well is usually a supply that was engineered well throughout, with synchronous rectification and soft-switching finishing the efficiency story on the output side.

Measuring it yourself

Power factor is easy to observe firsthand, and the measurement is worth doing. Any decent power meter — from a wall-plug energy monitor on a small miner to a clamp-style power analyzer on a 240V feed — reports real watts, apparent volt-amperes, and their ratio directly. A modern ASIC supply in good health should read in the high 0.9s under load; a reading that has drifted noticeably lower on a machine that used to measure better is diagnostic information, hinting at aging bulk capacitors or a struggling PFC stage before the supply fails outright. Note that power factor is a loaded-condition spec — at idle or light load, even a healthy supply corrects less effectively, so compare like with like. For fleet operators, logging power factor per circuit turns the electrical panel into an early-warning system that catches supply degradation while it is still a cheap repair.