high voltage high frequency H-Bridge

[George]

I'm wanting to make a high frequency supply square wave - around 450Vac at 1MHz - 40mA, The voltage is pretty easy, it's the switching I'm trying to get my head around. I'm using a CCFL board to generate the high voltage bit - it does it at 50kHz. So I was then thinking of turning that into DC and chopping it up with an H-Bridge I've found a transistor that might do the job a BSP135 its a small signal transistor 600V 100mA and rise/fall times around 10-20ns.

However I'm not sure how to power the high side of the H Bridge - I would use opto couplers but they dont respond fast enough.

I was going to use the clkout on my pic and some ttl logic devices to trigger the bridge. But getting that high voltage is difficult.

The other way I've thought of is rewinding the transformer to a center tap and run a +ve and -ve voltage through a half bridge using a PNP on top and NPN on bottom and a couple of 200V zeners between clkout (based around 0v) and the base legs. It would be problematic as it would not allow me to vary the supply voltage as I'd like, and would need fine tuning before it worked (could boost the voltage of the CLKOUT a certain amount with a comparator to help a limited amount).

Any other suggestions or ideas would be greatly appreciated.

Thanks

[Archangel]

Hi George,
I do not think you can get a square wave out of a transformer, can you ? I would think it would be too reactive. I think televisions used to switch high voltage DC in the Horizontal circuits and may well be a source for transistors to do that, still don't know how square the output will be. Are you building a Tesla coil here ?

[George]

Yah, most small high speed ferrite transformers fall over at about 300kHz, this is why I was going to first boost the voltage, convert to high voltage DC and switch it back into a square wave, I'm just not sure on how to switch it with both high voltage and high frequency. I'm pretty sure i'd be fine if it were one or the other but combined they make it much more difficult for my simple brain.

It's for an experimental rf heating device.

Thanks

[Archangel]

It's for an experimental rf heating device.

Mmmmmm . . . 450v x .040 amp = 18W not going to heat very much with that . . . Why does it have to be a square wave ? Seems like a class c rf amp would do what you want if sinewave driven. . .

[George]

well - yah, it doesn't have to be much power, I was only wanting about 10W of heat at this stage. One of the most important things is it's able to be made cheaply and reliably. I't doesn't have to be a square wave, a square wave is just preferable as it's kind of like a higher duty cycle. I have thought about RF amps but cant seem to find any small cheap ones that can do that voltage.

So yah - ideally i'd like to transistor (bipolar or FET) switch it but it's just getting the Gate/base voltage up there.

thanks

[Melanie]

Most MF/HF Amps are designed to work into a 50 ohm load... so you're looking at a 4kW device to give you the voltage you want.

40mA 450V at 1MHz is still 20W of anybody's money... easilly doable with a single 807 or 6146 valve... but the plate Transformer will set you back $100-$200.

Look on eBay for old Heathkit AM Ham HF Transmitters (from memory they all had American Indian native names I think). Nobody plays with AM anymore since they discovered SSB goes further. Those old 'Boat-Anchors' go for a song... and are a goldmine of parts you just can't buy anymore.

Oh... if you achieve it with a square-wave input, let me know... I want to set a stop watch to see how long it takes for the FCC (or NZ's equivallent) to break-down your door...

[Dave]

Melanie,
Quote" Nobody plays with AM anymore since they discovered SSB goes further."

You should live in the US.... Much better audio....

Dave Purola,
N8NTA