Fusor Forums

Does a Spellman power supply schematics suffice ? I know that Cliff Scapellati, a member of this community, works with Spellman. If it also works then I will contact him to ,hopefully, sends me schematics that helps. =) By the way, what is the price range of the IGBTs series 3 or the FETs ?

Thanks,

Jack Puntawong

Jack, there are several topologies that you can choose from, each with strengths and weaknesses which you will have to decide on to accomplish the task to your budget/skill/purpose. Some research looking for terms such as 'bridge topologies' should get you on your self-learning way. There are also 'resonant' and 'soft switching' variants of the bridge circuits too.

There is an unusual full-bridge half-cycle topology that I came up with that uses diodes for some of the switching elements (rather than transistors) and that therefore makes construction very simple... but I later found out that it had already been invented (not entirely unexpected! It's tricky to come up with things completely new these days!!). It avoids flux-walking completely (where any asymmetry in the supplied current loads up a DC bias in the core), which can be a significant problem especially in home-brew supplies due to the poor waveforms that can come out of amateur circuits.

I'll post on it sometime over the next week so folks can see it. I can't remember what it was called when I found it in an old book [naturally!!] so can't go looking for a link on it!

You can easily solve DC biasing the core by putting a capacitor in series with the primary. Depending on whether or not you want to build a resonant converter, the capacitor will have different values.

Yes, but I think that is a solution good for lower powers and where efficiency is not important, probably fine for this sort of application. I think you would still run higher up one side of the BH curve than the other if you have an asymmetric supply waveform, which can get quite critical if you're pushing the thermal limits of the core. You'll need some fairly specialised high value capacitors too, to run order of amps at 40 kHz, something like a very low ESR 20uF polypropylene type I'd expect would be a minimum. Readily 'do-able', but all adds to cost and complexity, does it not?