Umm, complexity = cost. At a lot of companies in my area, circuit design costs upwards of $100 / hr. So you see, the cost of passive components quickly becomes a non-issue compared to the labor cost. You also ignored the cost of assembly, which increases with complexity. And since when were large and bulky passive components required? There are 20 F or larger "super" caps that take up less room than that active PFC circuit.
The whole concept of PFC is really no different than impedance matching in RF. We don't have any of these problems with harmonic currents with an RF source driving an impedance matched diode. In RF, there are well defined equations, based on physics, governing impedance matching and filtering. But when it comes to 60 Hz AC, all I see is a lot of hype and little to no science.
For experienced power supply designer, designing a PFC pre-regulator should not take a long time. Neither the amount of components is usually significant issue since active PFC's are quite common. PFC controller chip and associated passives can be assembled in no time with SMD components. Every PC power supply should have one to comply with EMC harmonic current requirements. Of course, for very small power supplies, one may get away without one. Can you represent a simple passive circuit which filters out every other current harmonic other than just mains frequency current and can handle, say, 500 watts of power, which would be required for the power supply that dave showed in his blog?
I'd like to see that 20 F capacitor which tolerates mains voltages (230 VAC, or if truely universal supply, up to 275 VAC), it surely is not very small, more likely it is equivalent to size of an refrigerator
Have you actually looked inside PC power supplies which have a passive PFC? PFC inductor is not small component. And like I said, complying to EN61000-3-2 will most likely require an active PFC, so whole thing is only for academic interest.
What makes you think that there is no science and physics behind a PFC? I used to think that all things obey very same physics laws around here in this universe. Even RF does, although it is often thought to be somehow different kind of electricity. I think there are dozen papers published on the subject via IEEE etc. Are you actually familiar with the problem and why PFC is required? Your comment makes me think that might not be the case.
RF impedance matching is a very different process. Just put in some passives or transmission line pieces to adjust the impedance around some frequency. Or put an appropriate filter to remove the unwanted harmonics. The whole process is usually simple for RF because the center frequency is usually very high compared to required bandwidth, so one can think just about the center frequency. Do you take any significant amounts of power via that impedance matched diode? Something like mixer or detector diode hardly does that. I bet that the RF amplitude is very small compared to diode V-I characteristic, especially for the mixer diode case.
For mains applications, we want to pull significant amounts of power via the PFC (tens to hundreds or thousand watts). Main purpose of the PFC pre-regulator is to make AC into DC in such a way that current waveform remains sinusoidal. Yes, passive components can be used for power factor compensation if the load is linear but rectifier+filtering capacitor is a non-linear load. Even linear load is problematic to compensate passively if load varies greatly.
Regards,
Janne