How switched-mode power supplies with APFC work

Some time ago, I was asked about power factor correction in switch-mode power supplies. I’ll try to briefly explain what it is and why it’s needed.

In everyday life, you’re most likely to come across a power factor corrector (PFC) in a computer power supply.
Of course, they’re found in other power supplies too—more often, in fact, than in computer power supplies. But these are usually industrial units, which are very rarely found in the home.
I think most readers of my blog and viewers of my channel know at least a little about electronics, so they’ve probably seen the inside of a computer power supply.
At first glance, a power supply with active PFC looks almost the same as a regular one.


But look more closely, and you’ll spot a large inductor on the “hot” side. Its core can come in different shapes, but toroidal ones are the most common, as in the example in the photo.
Another difference is that these power supplies usually have a single 450–500 V filter capacitor, rather than two 200–250 V capacitors. That’s because these units often support a wide input voltage range of 100–115 V, so there’s no need for an input voltage selector.


Don’t confuse the active PFC inductor with the output inductor used for cross-regulation, even though they look quite similar. The difference is that the PFC inductor usually has just one winding, while the cross-regulation inductor has several.


A power factor corrector doesn’t have to be active; it can also be passive. In that case, you’ll see an iron-core inductor with a pair of wires on the top cover of the power supply. It looks like a 10–20 W 50 Hz transformer.
This approach works too, but it can’t replace a full-fledged active PFC.

Now, a little about why any of this is needed. You probably know that the mains current has a sinusoidal waveform, with an RMS voltage of 220–230 V (here) and a peak voltage of 310–320 V. I won’t explain the difference between RMS and peak voltage now; I’ll do that another time. But if you haven’t seen it before, the sine wave looks like this:


Next, the AC is rectified and filtered by capacitors. The most common circuit looks like this: a bridge rectifier and a pair of capacitors (sometimes just one).
There’s also an input filter and a fuse, of course, but they’re not relevant here.


Under normal conditions, the voltage across the capacitors will be around 280–320 V, depending on their capacitance and the load power. I’ve already talked about this in my video on how power supplies work.
But since the mains voltage essentially goes from zero to 320 V and back to zero 100 times per second, and there’s a bridge rectifier in the circuit, current only flows to charge the filter capacitors when the peak voltage exceeds the voltage across them.
The current in the 220–230 V circuit will then look like the waveform at the top of this image. In other words, the power supply doesn’t draw power continuously, only at the peaks of the sine wave. If we assume the PSU draws power for just 20% of the time, then the current while the capacitors are charging will be five times the average current draw. For example, at 1 A and 220 W, the current peaks will reach 5 A.


This problem became impossible to ignore once the number of switch-mode power supplies passed a certain “critical mass.” Eventually, a fairly simple and effective solution was devised. By the way, in developed countries, all high-power supplies are required to have PFC. But since it adds to the cost, manufacturers of budget power supplies are the first to cut corners here.

As I said, the solution is simple and, in a sense, right there in front of us. It’s based on an ordinary step-up converter. The schematic shows an inductor, a transistor, a diode, a PWM controller, and a capacitor.
When the transistor turns on, energy builds up in the inductor. When the transistor turns off, that energy adds to the input voltage and flows to the load, charging the output capacitor. This circuit is often used in power banks to step 3.7 V up to 5 V.


Combine a regular power supply with this circuit, and you get an active power factor corrector.
The important thing is that there’s no filter capacitor after the bridge rectifier. Its job is done by a small-capacitance capacitor, usually 0.47–1.0 µF, which is only there to compensate for the pulsed current draw of the PFC.

As a result, the converter tries to “suck” as much as it can from the mains over a range of not 220–230 V, but 40–80 V. By the way, high-power supplies don’t always support a wide input voltage range, even if they do include active PFC. In these operating modes, the PFC has a hard time and cannot provide operation over a wide range, although it continues to work correctly.


Here I’ve tried to clearly show the difference between a regular power supply and one with a PFC.
The regular power supply’s operation is shown in red. The hatched area represents the period when current is drawn. You can see that this period is quite short, so the current will be high. And the larger the filter capacitor, the shorter this period becomes and the lower the power factor.
In blue and green, I’ve shown two examples of how an active PFC operates: one starts working at around 100 V of the AC voltage amplitude, the other at around 50 V. You can see that the current-draw period is longer, so the current drops proportionally and the power factor increases.
This period can actually start almost at zero, in which case the power factor will be one. In practice, though, it’s usually 0.98–1, which is more than enough.

So what are the consequences of this notorious power factor?
Current peaks cause brief overloads on the power grid, which can lead to problems in old, worn-out networks. In three-phase networks, the neutral wire may burn out, with very serious consequences.


Here’s the input stage of a computer power supply with an active PFC, highlighted in blue.
Don’t be surprised that the schematic doesn’t show the PWM controller that drives it. It’s often located on a separate board, and sometimes it’s integrated into the main PWM controller. In other words, in addition to one or two standard channels, there’s also an output for controlling the PFC transistor. This arrangement is convenient for the manufacturer, but not always for the repair technician. At the very beginning, I showed a photo of a power supply where the PFC stage had failed. Since the IC controls everything, it had burned out too. I couldn’t find a replacement, so the power supply is now dead weight and may eventually be stripped for parts, especially since they’re quite good quality.


So what does a PFC give us? First, the advantages:
1. Its current draw is almost the same as that of a resistive load, so there are no current peaks.
2. These power supplies often have a wider input voltage range and work better on poor-quality power grids.
3. A smaller filter capacitor is needed because the periods with no current are shorter.
4. The power supply’s inverter has an easier job, since it is essentially powered by a regulated voltage.

Now for the disadvantages.
1. Higher cost.
2. Lower reliability.
3. They may have compatibility issues with some UPS models.

There’s sometimes debate about the efficiency of these power supplies. I believe their efficiency is the same: although the PFC consumes some power of its own, the main inverter has an easier job, so it all balances out.

And of course, here’s a video to round things out. As always, I look forward to your questions, both in the comments here and under the video.