Current sensing in boost PFC controllers

Question

I was going through the datasheets of a few active PFC controller ICs (like this: https://www.ti.com/lit/ds/symlink/ucc28019a.pdf?HQS=dis-mous-null-mousermode-dsf-pf-null-wwe&ts=1619350669153&ref_url=https%253A%252F%252Fwww.mouser.in%252F) when I noticed something odd about the way the switch current is sensed.

Because of the location of the current sense resistor, two things happen: the load current gets added on to the switch current and the voltage across the current sense resistor is negative, as seen by the IC.

As I scrolled through the datasheet, I found the functional block diagram of the IC.

You can clearly see that the sensed current is fed to a -1x amplifier!

So my questions are:

• If the voltage across the current sense resistor, which is negative, is fed to a -1x amplifier, then why sense it in that manner instead of doing it like this:

• And why is the load current added to the switch current?

I would like to build a PFC circuit, but at a low voltage (like 24V AC) for safety reasons (I have zero experience building high-voltage circuits). These ICs are designed and optimized for mains voltage applications, so will it lockout at low AC voltages or will it still function normally?

3)

Could I just use a regular boost converter (without the input filter cap) to perform PFC, or do I really need to use a dedicated PFC chip for this?

Answer 1

For power factor correction (PFC) controllers, you have two options for sensing the current: positive or negative sensing.

1. Positive sensing: this is when the resistance is in series with the MOSFET (or power switch) source. In this case, the voltage sent to the controller is positive and corresponds to the inductor current image during the on-time. This information includes the gate-source current also and thus needs to be appropriately cleansed with a leading-edge blanking (LEB) circuitry. The below application circuit shows how a simple positive sensing is implemented in a classical borderline conduction mode (BCM) type of corrector, NCP1608:

The advantage of positive sensing are indeed ease of implementation but you only see the inductor current during the on-time while the circuit is blind during the off-time. Also, the sense resistance as it is shown does not limit the in-rush current and the controller can adversely start pulsing while this power-on current still circulates and potentially saturates the inductor.

2. negative sensing: depending on the adopted control law - the way the switching pattern is elaborated to perform power factor correction - there can be a need to sense the entire inductor current during the on- and off-times. In this case, you need to insert the resistance in the return path, just before going back to the diodes bridge:

The internal controller circuitry depends on the designer choice but with this controller, the NCP1654, it is designed to maintain 0 V at the CS pin and you program the peak current via the resistance \$R_{CS}\$: \$i_{CS}(t)=\frac{R_{sense}}{R_{CS}}i_L(t)\$.

In this configuration, the controller monitors the entire inductor cycle but knows also when the in-rush current circulates. The sensing information is also very clean because there is no gate-source drive current spike. This technique is very popular in high-power CCM PFCs.

For your low-voltage ac experiments, I encourage you to start with a simple circuit like NCP1608 or even a venerable MC33262 if you can find a few. These are simple self-relaxing controllers and when you master PFC operations, you can upgrade to a more complicated CCM PFC type of controller.