1 Conducting Diode PFC/Boost

[TheDood]

I've been trying to build a low wattage high efficiency + good PF power supply and curious if the attached schematic showing a generalized topology is currently used or if it'd work with a traditional PFC IC that senses and controls the boost/PFC. I couldn't find any JFETs preloaded into LTSpice that were greater than 50Vds, and not sure exactly where to place ground nodes, or if it's simulating correctly?

The general idea is to use "On-until-acted-upon" switches on the ground leg of a traditional rectifier (as well as adding a coil to each AC leg) so that during the coil charging time period the current doesn't have to go through the rectifier (dissipating power), and then during the coil discharge time period only 1 passive diode is needed to be turned on to charge the DC cap? All PFC would still be controlled by a PFC IC (PWM ect), only an "either or" type of cct would be added between the PFC IC switching output pin, and the JFETs (tradionally a single N-Channel FET), perhaps an inverted "either or" cct? 

During coil charge up no signal would be given to either JFET so they'd conduct simulataneously and allow current to short through the coils. Then when it was time to charge the cap, only 1 JFET (high side) would be signaled to turn off, which would direct current towards the DC side and allow the coil to discharge through only a single diode on its way to charge up the capacitor on the DC side? This way during coil charge up there's no diodes conducting, and during coil discharge only 1 diode dissipates energy?

Would 2 coils in series be beneficial, ie would it be easier/cheaper/more efficient to source 2 inductors of the same rated current rated but lower rated inductance? Is the relationship between coil Ω's and their inductance linear? If not which is more efficient, 2 smaller in series, or 1 larger one?

The reason I'm thinking JFETs is due to my assumption that no power is dissipated at the gate in order to flow current, but actually the time period for coil charging is probably less than 50% duty cycle, so perhaps N-Channel FETs would be more efficient? Do they make GaN JFETs?

Anyways, just trying to reduce diode losses and curious if it was more efficient than a typical bridgeless rectifier? Maybe 2 coils is a waste and 1 coil is same as 2, or cheaper/more efficient?

Thanks

[T3sl4co1l]

You're looking for "bridgeless PFC":
https://www.ti.com/lit/an/slyt750/slyt750.pdf

It's a fine method but it requires an actual controller, not just comparators and pulse generators.

The two inductors are in series, but they are better kept balanced, for EMI purposes.

Tim

[TheDood]

You're looking for "bridgeless PFC":
https://www.ti.com/lit/an/slyt750/slyt750.pdf

It's a fine method but it requires an actual controller, not just comparators and pulse generators.

The two inductors are in series, but they are better kept balanced, for EMI purposes.

Tim

Thanks Tim,

Yes that was the inspiration, and thanks I know Ill need to use a controller (per your guidance). The topology posted reduces an additional conducting diode from the totem pole PFC I was looking at. There's only 1 conducting diode per half sine wave and 0 conducting diodes during coil charge (or that was the intention).

Previously I wanted to see if I could use OPAmps to create a crude controller by themselves. I was curious if I understood the principles of PFC and wanted to try to implement dynamic signaling in LTSpice to test out my comprehension.  After simulating and calculating what the power consumed by the comparators would be, I've given up for a dedicated controller when it comes to practicality.

I was wondering if the above posted operation, in principle, is sound, or possible? Can (2) N-Channel JFETs operate like I'm intending, in terms of high side and low side? Can I eliminate a conducting diode? In theory, I was thinking Id have to put some sort of inverting opamp config between the controller output pin and JFET gates (+ a hide side/low side switch selector), but otherwise I could use a typical single FET PFC controller?

EDIT:
Well after looking at your link its pretty much the exact same thing as the first example except using JFETs. Perhaps it's not any more efficient.

I think I can use an Arduino programmed MC to sense storage cap V, line V and line current, and use those measurements to manage the switching FETs. I know I can buy a dedicated controller but I'm a tinkerer and curious lol I still will probably buy a few controller chips to test out and compare with.

[TheDood]

Also, where do I put the GND node on a cct like that? If I put under cap will it solve weird when both FETS are conducting the AC signal?