A 400 W driver with a power factor of 0.6 draws 3.6 A from a 230 V supply. The same 400 W driver with active power factor correction and a PF of 0.95 draws 1.83 A. Same output. Same efficiency, roughly. Same heat in the cabinet. Half the input current.
That difference shows up in four places, and three of them are on your invoice rather than the driver’s: the breaker, the cable, the generator, and — in a growing number of markets — the electricity bill itself. Yet power factor is the one specification that buyers skip past on the way to the efficiency figure, because they assume the two are the same thing. They are not, and the confusion is expensive in a way that is easy to demonstrate.
Power factor is not efficiency
Efficiency tells you how much of the power that enters the driver reaches the LEDs. A driver at 88% efficiency turns 12% of its input into heat. That number is about waste.
Power factor tells you how much of the current the utility has to deliver is actually doing work. A driver with a PF of 0.6 needs the supply to deliver 40% more current than the real power alone would require, and that extra current sloshes back and forth between the driver’s input capacitor and the mains sixty times a second without doing anything useful. It is not lost as heat in your cabinet — it is lost as capacity in the grid, the cable, the transformer and the generator.
Which means a driver can be 88% efficient and have a power factor of 0.6 at the same time. Ours does, and so do most low-cost constant voltage drivers above 100 W. The two numbers describe orthogonal properties, and a datasheet that lists only one of them is telling you half a story.
Here is the arithmetic that matters, using a 400 W unit at 230 V:
| PF 0.6 | PF 0.95 | |
| Real power in | 455 W | 448 W |
| Apparent power | 758 VA | 471 VA |
| Input current | 3.60 A | 1.83 A |
| Heat in cabinet (88% eff.) | 55 W | 54 W |
| Cable copper for 100 units | 360 A design | 183 A design |
The heat is identical. The current is not. If you are building a cabinet with a hundred of these, the difference is 177 A of copper, breaker capacity and transformer rating for the same light output.
Why the number is 0.6 and not 0.95
A driver without power factor correction has a bridge rectifier feeding a bulk capacitor directly. Current only flows when the instantaneous mains voltage exceeds the capacitor voltage, so it arrives in short, tall pulses near the peak of each half cycle instead of following the voltage sinusoid. That pulse shape is what a low power factor is: the current waveform is out of phase with the voltage and rich in harmonics.
Add a passive PFC stage — usually a valley-fill circuit with two capacitors and a diode — and the current is drawn over a wider portion of the cycle. Passive PFC typically gets you to 0.85–0.90 at 230 V, costs perhaps $0.80, and is what most mid-range drivers use. It has a weakness that surprises people: at 120 V it degrades badly, often to 0.70 or below, because the valley-fill topology is tuned around a 230 V input. If you buy a passive-PFC driver in Europe and ship the same luminaire to the US, the power factor you certified is not the power factor you get.
Active PFC uses a boost converter ahead of the main stage. It forces the input current to track the voltage sinusoid, gets you 0.95 or better from 90 V to 264 V, and costs $3 to $6 depending on the power level. It also adds an inductor and a switcher to the bill of materials, which is why the 150 W active-PFC version of a driver is usually 8–12 mm taller than the passive version. In a 35 mm deep lightbox, that 10 mm decides whether the product exists.
The regulation, and the bit everyone misreads
In the EU, EN 61000-3-2 sets the harmonic current limits. Lighting equipment above 25 W falls into Class C, and Class C above 25 W is strict enough that in practice you need a power factor around 0.9 or better to pass. Below 25 W the limits are per-watt rather than proportional, and a 15 W driver can pass with a much lower power factor. This is why small drivers so often have no PFC stage at all: the regulation does not require one, and the customer is not asking.
Two things get misread constantly.
First, EN 61000-3-2 is an emissions standard, not a performance standard. Passing it means your harmonics are below a limit at the test conditions. It does not mean your power factor is 0.95 at 30% load. Almost every driver’s PF collapses at light load, because the PFC stage’s own consumption becomes a significant fraction of a small real power. We have measured 0.93 at full load dropping to 0.52 at 20% load on drivers that carry a 0.95 badge. If your luminaire dims to 10% every evening, the number on the datasheet describes a condition your installation never sees.
Second, DOE Level VI does not apply to your built-in driver. Level VI is a US efficiency specification for external power supplies — the brick on your laptop cable. It says nothing about a driver that is wired into a luminaire. Suppliers quote it anyway because buyers recognise the logo, and buyers accept it because they have seen it before. It is not a certification of anything relevant to a built-in LED driver. The relevant US safety standard is UL 8750 for LED equipment, and for a Class 2 output, UL 1310.
Where low power factor actually costs money
Four places, in ascending order of how often they surprise people.
Breakers and cable. Covered above. At 400 W, PF 0.6 costs you double the current. On a 16 A final circuit you fit four units instead of eight, and the extra circuits are the panel builder’s problem until they become your problem.
Generators. A 100 kVA generator is rated at 0.8 power factor, which means 80 kW of real power. Load it with drivers at PF 0.6 and you hit the alternator’s current limit before you hit its kW limit: the usable real power drops to about 60 kW. On a temporary lighting installation — an event, a film set, a construction site — that is a 25% shortfall against the number on the hire contract. The symptom is not a trip; it is voltage sag and flicker every time a contactor closes, and a generator that hunts.
Neutral current in three-phase boards. Single-phase non-linear loads on a three-phase four-wire system produce triplen harmonics — the 3rd, 9th, 15th — that add in the neutral instead of cancelling. With a badly balanced lighting board full of low-PF drivers, the neutral current can reach 1.7 times the phase current. Neutrals are often sized the same as phases, and sometimes not even brought out to the board. This is a genuine fire risk and it is not theoretical: it is the reason some European specifiers now demand PF ≥0.9 as a contractual condition rather than a regulatory one.
Tariffs. In India, parts of Brazil, and a number of Middle Eastern utility contracts, reactive power is billed or penalised above a threshold, typically PF 0.9. A hotel retrofitting 800 downlights at PF 0.6 can find itself paying a reactive power charge it never had before, and the facilities manager will trace it back to the lighting supplier. This is the scenario where the customer asks for a compensation report and nobody in the supply chain knows how to produce one.
What a good RFQ says
If you are writing a specification and you want to avoid all of the above, four lines will do it:
- Power factor ≥0.90 at 100% load, ≥0.85 at 50% load, at both 120 V and 230 V. The 50% clause is what separates drivers that were designed for the requirement from drivers that were tested for it. It also stops the passive-PFC-at-120 V problem dead.
- THD ≤20% at full load. Total harmonic distortion of the input current. A driver with PF 0.95 and THD 15% behaves very differently on a generator from one with PF 0.95 and THD 40%.
- Test standard and class stated explicitly — EN 61000-3-2 Class C, with the test report number. Not “meets CE”.
- Inrush and leakage stated with the PF figure, because a high-PF stage with a large input capacitor can have a worse inrush than the low-PF driver it replaced. We have seen exactly this: a customer upgraded to active PFC for the generator and then had to change the breakers, because the boost stage’s capacitor was bigger.
That last point is worth dwelling on. Active PFC improves steady-state current but can worsen the first 2 ms, because the EMI filter and the PFC bulk capacitor together present a lower impedance at switch-on. If your panel is already marginal on inrush, upgrading the driver for power factor reasons can push it over. The two problems have to be solved together, not one after the other.
What we actually build, and why
Our standard 150W IP20 constant voltage driver and the 400W indoor power supply are passive-PFC designs, with a power factor of 0.6 on the 400 W unit at full load. That is a deliberate choice for the markets these models serve — fixed indoor installations on 220–240 V supplies, sold on cost per watt, where the circuit is designed by an electrician who has the headroom and the customer is not billed for reactive power.
It is the wrong choice for a generator-fed installation, for a three-phase board with a shared neutral, or for a project in a market that charges for reactive power. For those, we build the active-PFC version on an OEM basis: 0.95 across the load range, 120–277 V input if the project needs it, and the specification written into the test report rather than onto a marketing page. The cost delta is real — roughly 6–9% on a 150 W unit at 500 pieces — and it is smaller than the cost of a reactive power penalty or a re-cabled lighting board.
The honest framing is this: power factor is a system-level number that gets sold as a component-level badge. If your project has a generator, a shared neutral, or a utility that meters reactive power, PF belongs in the RFQ next to the wattage. If it has none of those, a passive front end is a legitimate engineering decision and you should not pay 8% more for a number that never reaches your invoice.
FAQ
Does a low power factor mean the driver wastes energy? No. Efficiency and power factor measure different things. A driver can be 88% efficient and have a PF of 0.6; the 12% is heat, and the 40% extra current is circulating reactive power that the utility has to generate and deliver but which you do not consume. You pay for it through cable, breaker and generator capacity, and sometimes through a tariff.
Is PF 0.6 legal in Europe? For lighting above 25 W, EN 61000-3-2 Class C in practice requires a power factor near 0.9 to pass the harmonic limits, so a 400 W driver at PF 0.6 would not normally be placed on the EU market as compliant. Many low-cost 400 W units are sold into markets where the standard is not enforced, and buyers discover the gap at customs or at the first audit.
Why does the datasheet say 0.95 when my meter reads 0.72? Because the datasheet number is measured at 100% load, and your installation runs at 30–40%. PFC stages lose effectiveness at light load. Ask for the PF-versus-load curve, not a single number.
Can I add a capacitor in the panel to correct the power factor? You can, and it is what industrial plants do for motors. For an LED driver with a rectifier front end, a fixed correction capacitor interacts with the driver’s input filter and can cause resonance and overvoltage at switch-off. Correct at the driver, or specify a driver with active PFC. Panel-level correction of a distributed non-linear load needs a proper study, not a capacitor.
Does active PFC make the driver bigger? Usually 8–12 mm taller at 150 W, because of the boost inductor. In a 35 mm lightbox that matters. Ask for the mechanical drawing before you design the housing, not after.
Specifying for a generator, a shared neutral, or a market that meters reactive power? Our OEM LED driver manufacturing line builds active-PFC versions of the standard indoor range with the PF curve written into the test report. Send the load profile and the supply details and we will come back with the driver and the panel implications together — because on a generator, they are the same decision.