Controlled Saturable Reactors in Power Supplies

The trick would be to saturate a coil (in the microhenries and 170 amperes range) and to transmitt its energy (1/2Li^2) to the resonnant capacitor, in the range 0,35µF to 0,5µF.

I think that by saturating the coil you will no longer have an optimum transfer of energy, the core of the coil will get a bit hot. And even a 1000 microhenry coil, driven with 170A current cannot handle to much energy (some 14W).

Oops!

Thank you for the GA, but going over what I have said, I noticed a small omission. The energy is for one cycle of the switching supply which is in the microsecond range. Therefore, the energy handled in one second is waaaay larger. Bad boy, bad boy (this is me, of coarse)

I'm surprised/flattered that someone who's designed such units (running at many hundreds of amp) would even ask our opinion.

Although I've played around with DC/DC converters I don't even understand what you're asking. Can you explain a bit more?

What sort of topology? What sort of control scheme? What does "second arch ..." mean?

I believe he is referring to the output of a transformer into a rectified and inductive circuit where the second arch is the power being drawn from the stored inductor energy.

Power outputs can be increased at higher frequencies with smaller inductive components but this throws more strain on the H bridge switching units.

Excuse me for being late in answering your question.
A resonant power supply uses soft switching, at zero voltage and/or zero current, in a resonant circuit including the primary of the power transformer, a power capacitor (polypropylene) and a stray coil which helps to tune the resonant frequency.

The advantage is that the stress and dissipated power is much lower than using "hard" switching (at high voltage and / or current levels), specially when using IGBTs which have a long tail (turn off energy).For example, at 20kW power level, the tail loss in hard switching could rise 20 millijoules,and at 50 kHz, the dissipated power would reach 1kW, really much too high.

Using soft switching suppresses such losses, but there are two drawbacks:

1)the energy stored in the resonant circuit is 2 to 3 times that matching the mean power, so there are high currents ( 170 amperes at 20kW level and 550 Volts)

2)the rectifiers at the secondary(ies) of the transformer rectify only the first half cycle of the decaying sinusoïd.Above 12 kilovolts, one stacks secondaries, because it is the only mean to achieve high power (up to 100 kW).

During the second half sinusoïd, the stored energy has decreased, because power has been extracted by the rectifiers and load.The peak voltage is now too low for the rectifiers to conduce once again and extract more power.
My idea is to use a saturable stray reactor, which would be turned off during the second half cycle, to increase the peak voltage(by equipartition L --> C),and thus enable further rectification and an increased power output.
I wonder wether anyone has tried this