Kapanadze Cousin - DALLY FREE ENERGY

[verpies]

So this means any MOSFET with internal parasitic diodes are causing this

Unfortunately yes. These parasitic diodes are one of the main disadvantages of MOSFETS compared to BJTs.

...meaning use your suggested method of extra Shottky diodes/snubbers, or as an alternative; swap to BJT transistors like in the diagram below?

Yes, except that the series Shottky diodes do not act as snubbers. They act only as reverse current blockers.
The parallel Zener diodes act as spike snubbers.

P.S.
I edited my two previous posts a little.

[Hoppy]

@ Verpies,

I would like your expert opinion please, as I'm going to wind a double 'E' core ferrite from an ATX PSU, so that both low and high frequency operation can be tried. If the inverter ran at 12-15KHz, do you think that 3 + 3 turns primary and 70 turns on each secondary as stated on schematic will produce around 200V after rectifification and smoothing? The schematic calls for 1.5mm wire for primary windings. Would you go for a Litz or multi-strand to reduce losses?

Hoppy

[verpies]

@ Verpies,
I would like your expert opinion please, as I'm going to wind a double 'E' core ferrite from an ATX PSU, so that both low and high frequency operation can be tried. If the inverter ran at 12-15KHz, do you think that 3 + 3 turns primary and 70 turns on each secondary as stated on schematic will produce around 200V after rectifification and smoothing? The schematic calls for 1.5mm wire for primary windings. Would you go for a Litz or multi-strand to reduce losses?

Yes, I think it will produce even more voltage if you supply the primaries with 12V.
I would go with stranded wire on the primaries because Litz is too expensive, but if you already have it then Litz is better of course.

The only problem I forsee is insufficient primary inductance at these pulse widths and 3 turn primaries...however if your ferrite has high permeability then you should be OK.  In any case do allow your primary current pulses to flatten out beyond Point C on the time axis, as shown on the yellow I_L trace in Reply #29.

[Hoppy]

Yes, I think it will produce even more voltage if you supply the primaries with 12V.
I would go with stranded wire on the primaries because Litz is too expensive, but if you already have it then Litz is better of course.

The only problem I forsee is insufficient primary inductance at these pulse widths and 3 turn primaries...however if your ferrite has high permeability then you should be OK.  In any case do allow your primary current pulses to flatten out beyond Point C on the time axis, as shown on the yellow I_L trace in Reply #29.

Thanks. I do have some short lengths of Litz, so will use these.

Regards
Hoppy

[dorcky]

Thanks. I do have some short lengths of Litz, so will use these.

Regards
Hoppy

Hoppy can you give more info about the Litz wire that you want to use?
Some picture or datasheet?

Thx!