Assume we have designed a PFC stage that perfectly shapes the input current. Will we have a nearly unity power factor (PF)? Not necessarily as the EMI filter capacitor unfortunately degrades the PF.
If you are not convinced, run this simple simulation.
As shown by Figure 1, replace the PFC stage by a resistor which exhibits a perfect PF and add a front-end EMI filter made of capacitors and an inductor. You will note that the line current is sinusoidal but that it is not in phase with the line voltage. This current is actually the sum of the wished current drawn by the resistor Rin with the reactive current flowing through C1 and C2. This reactive current is sinusoidal but 90-degree ahead of phase.
Actually, we can define the maximum power factor achievable with given input capacitors. The result of a simple computation gives the following expression:
or:
In light of (1) or (2), we can note that:
The larger the input capacitors and line frequency, the higher the reactive current generated by the input capacitors and hence the lower PFmax.
The higher the power or the lower the line rms voltage, the higher PFmax since both conditions cause a higher current to be drawn by the PFC stage. Hence, the relative contribution of the reactive component in the total line current reduces. For the same reason, the lower Rin, the higher PFmax.
In other words, this formula expresses that the PF gets worse when the reactive current magnitude increases compared to the PFC stage one.
Figure 2 shows that the input capacitor’s influence can be illustrated by playing with the line frequency. On the left side, you can observe the line current (in blue) obtained with ON Semiconductor’s NCP1653 PFC controller evaluation board ([1]) at 300 W and 230 V rms, operated with a 50-Hz line voltage. In this case, the THD is low (7%) and PF is high (0.986).
In Figure 2 right side, the same is done but with a 400-Hz line voltage. The line current is still nearly sinusoidal but it is no longer in phase with the input voltage (black trace). The THD is unchanged but the power factor is much worse (only 0.730!).
This example illustrates that the EMI filter design is particularly critical at high-frequency line frequencies. It further instructs that the THD gives a better measure of the PFC stage performance than PF since the power factor can be affected by the EMI filter.
Please also note that an input capacitor is also generally placed across the diodes bridge to filter the rectified line voltage. Such a capacitor plays the same role as the input capacitors of the EMI filter just discussed.
However, it further degrades the THD when the PFC current is not large enough to fully discharge it near the line zero crossing. There is hence a period of time during which the PFC stage is only fed by the input capacitor and no current is absorbed from the mains.
The current is hence distorted near the line zero crossing.
([1]) The NCP1653 is a continuous conduction mode PFC controller. With its successor (NCP1654), it provides a very compact solution for applications above 300 W.