I'm sorry.

Did I confuse you? I've labeled the phases so now it should be perfectly clear what is charging, and when.

DUTY CYCLE

How wrong is having a 58% duty cycle? Its a hard thing to decide. 58% is a hard ratio to make with flip-flops. Is it worth it adding extra inductance to phase 2 to even out the effects of a reduced capacitance, half, due to c1, c2 in series? Such that the duty cycle becomes exactly 50%, so that a good flip flop will drive them. It turns out this calls for 2 equal inductors.

OPEN LOOP BUCK CONVERTER

Why not have a fixed duty cycle? Whats the worst that could happen.

Transformers

are not to be used as the prime point here is using an old fashioned transformer-buck or flyback converter is old-fashioned, and old fashioned circuits are nothing new and certainly no better than the status quo.

ON THE METHOD OF REDUCING THE OCCURRENCE OF CONCURRENTLY SWITCHED-ON MOSFETs

How long must one wait after turning off one MOSFET before turning on another such that they are never on at the same time? I'm inclined to think a whole microsecond (1 µs) to be on the safe side. That's is enough leeway so that any buffers, inverters, drivers, gate delay, gate capacitance should still fall within that timeframe.

1 microsecond places a genuine upper limit to the frequency of the converter. 100 khz is likely the maximum practical frequency for such a converter when this delay is incorporated.

Efficiently improving efficiency.

High voltage is the arch-enemy of high current. (for a given power level). Therefore to decrease current and increase efficiency at a given power level the capacitors must not be fully discharged. There must be residual voltage remaining at both phases. The energy in J [dis]charging a cap c from v1 to v2 is given by $1/2$c (v1^2-v2^2).

Charging a LC circuit from v1 to v2 requires a maximum current proportional to v1 - v2. But there is more energy transferred at higher voltages, proportional to v^2.

But this won't work.

Asking too much?

It may have been asked before and the time has come to ask again, the question in question being:

What kind of 33µF 320V capacitor enjoys the fine feeling of nearly 40A RMS ripple current?