SWITCHER DESIGN Q.

Using the flyback boost at 80A is problematic. The difficulty is to catch the 160A of current in the secondary path. The stray inductance can hardly be done less then ~20nH. So, for the reasonable voltage peaking the FET turn off time should be in the hundreds of nanoseconds

-> low efficiency! Also, the ripple current of 160A represents the significant problem for the reservoir capacitors.

My recommendation is the forward convertor with the push-pull primary.

Having said that, I've seen the 12V boost convertors of the 1kW class. They kinda worked for some time, with the poor efficiency, avalanching FETs in the primary, and blowing the capacitors in the secondary.

Take the 12 gauge shotgun case and wind ~ 10...15 turns on it.

Vladimir Vassilevsky DSP and Mixed Signal Design Consultant

What we did in ham radio was find the largest available core that still had reasonable pricing, then buy enough of these and stack them. Just make sure they can't rattle. Back in those days anything above 2" O.D., while available, carried boutique prices. So we stacked 2" toroids.

Vladimir's sugestion of going with a bridge architecture is a valid one, boost is iffy at that power level.

--
Regards, Joerg

http://www.analogconsultants.com/

"gmail" domain blocked because of excessive spam.
Use another domain or send PM.

--
http://www.google.com/search?hl=en&q=high+power+magnetic+cores&btnG=Google+Search&aq=f&oq=&aqi=


http://www.mag-inc.com/pdf/ps-01.pdf
JF

Consider polyphase or just in-phase parallel boost, with multiple fets, inductors, and diodes. Things are a lot more managable at, say,

20 amps, and more standard parts are likely to be available. Thermals can be better, too, with heat sources spread out.

John

who invented the polyphase buck switcher ca 1981, for a 120 amp CAMAC crate power supply. I expect somebody else invented it first.

Multiphase boost works too, however it is not a problem to make just a conventional push-pull for the power level of mere 1..2kW. It is going to be simple, cheap, pretty compact and efficient.

Vladimir Vassilevsky DSP and Mixed Signal Design Consultant

You can use just one tapped winding:

GND ----/ o---------+------->!-------------+----24V ) ! ) ! ) ! 12V------+ ! ) ! ) ! ) ! GND ---/ o----------+------->!-------------

Only 1/2 the power actually goes through the core so a much smaller core can be used. The load side rectifiers snub the swing so leakage inductance doesn't matter as much.

At 1% duty, it is not a 1KW component.

At the same time, you don't want it to be a fuse, either.

Design for average power that reflects the higher stresses anticipated of a pulsed design.

RL

That used to be easy, maybe not now. Just ask a manufacturer / vendor for samples, and tell them 12 v in, and 24 V out, F, (maybe L,) and power, ask to compare a few shapes. Do not tell them intermittent operation.