So I need ~1MHz @ ~10amps into <1ohm

[Speakerguy]

Removed! Sorry

 

[Hero999]

What's the voltage rating on the capacitor?

10A though a 1nF will give over 2.2kV!

Why are you doing this?

There are far more effective ways to generate high voltages than this.

 

[Speakerguy]

Sorry, removed!

 

[mneary]

Tap your inductor at a reasonable drive point so you can deliver some power.

Use a resonant drive for the MOSFET gates because of the high capacitance of MOSFET gates (up to 3 nF). You're probably better off with a BJT anyway.

 

[Speakerguy]

Removed again...

 

[Nigel Goodwin]

Why would BJT's be better? I am trying for power efficiency with this (focusing on using switches and reactive components obviously) and I thought Vce sat on most high current parts would be high. But again, open to ANY ideas. I have to 'make it to work' any way possible, even if the result doesn't meet expectations. This is a "damn the torpedoes" and run with it experiment.

I think RF power devices are usually bipolar rather than FET?, it's easier to make a bipolar switch fast than an FET due to it's high gate capacitance.

How fast can I make a 555 timer go, realistically? I feel silly for asking this question at ~725kHz or so but it would be a real simple solution that I could make easily tuneable with the right choice of an R with a series multiturn low R potentiometer. Also are there any other faster timer chips that would work? The 555 is ancient and I'm sure there is something better but am not familiar with newer parts.

It's more common to use CMOS logic gates as RC oscillators for higher frequencies, 555's don't really go very high.

 

[Speakerguy]

And again....

 

[Optikon]

6.3KVDC, I also have 20kV 5.2nF ceramics around too.

What problems do you see with this approach? Efficiency?

Goal is to have the cap discharge through an arc gap once breakdown voltage between electrodes is achieved, then recharge via the oscillation ckt. I'm looking for maybe 1 millijoule per pulse through air or thereabouts for 100-200ns.

I have used high voltage supplies from Ultravolt, Emco etc with very expensive HV switches (Behlke) to get what I need, but am looking for something way cheaper/simpler. Also have looked into transformers, ignition coil variants, etc. This is just one of several ideas, open to any others you might have!

Use a transformer to get to your HV.. all this resonant schtuff will give you more headaches than you care for. How are you going to keep things resonant? You dont even know the ESR of your capacitor. The conditions for resonance will likely be a moving target.

 

[Hero999]

A TV flyback is more better at producing high voltages.

What do you want the high voltage for?

Is it for igniting something or just playing around?

 

[mneary]

A bjt is favored for RF because of the high gate capacitance of the big MOSFETs. To switch a MOSFET fast enough to be efficient at 1 MHz you may need 10A of gate current. If you're looking for 10A output, you've gained nothing.

The high gate current isn't a problem at a few KHz, because you don't need it very often.

 

[Hero999]

Does this need to be really high powered?

You could build quasi-resonant converter.

The idea is you switch the MOSFET when there's no current flowing so it doesn't dissipate any energy. There are lots of papers that can help you design one whcih can be found from Google.

If you're interested I'll PM you some information you I can't post but I'll have to wait until Tuesday as it's on my work's comptuter.

 

[Speakerguy]

And yet again...

 

[Hero999]

Re: what this is for, it's for generating a non-thermal plasma. A spark between two electrodes that is so short in duration that it doesn't heat up the ionized charge carriers, thus wasting no energy heating up the plasma. You get a 100-200 nanosecond zap, then nothing. Pulse power is 10,000W for the duration of the pulse, but my duty cycle is so low that total power consumed by the system should be less than 100w.

So you need a high energy 200ns pulse.

Have you considered something simple like an automitive ignition coil and capacitor? They give pretty high energy short duration energy pulsed.

You could also charge up a capacitor using a high voltage power supple then discharge is using a spark gap switch.

 

[Speakerguy]

Removed.....

 

[Speakerguy]

Removed....

 

[Nigel Goodwin]

Anyway, problem now is that my FET driver isn't working. Still trying to debug that.

As suggested previously, a FET is probably harder than a bipolar - you will need considerable charge and discharge capability to switch it fast enough.

 

[Speakerguy]

Removed....

 

[Hero999]

I still don't understand what you're doing.

I thougt you wanted a short brief high voltage pulse not a continious wave.

For a continious high voltage you're much better off using a high frequency neon sign transformer or TV flyback rather than messing around with 555 timers and capacitors.

For a short high voltage pulse it's best to charge a capacitor to an imtermediate voltage (about 300V) then discharge it though a transformer. A higher primary voltage on the primary will help to give a faster di/di and thus a faster rise and fall time on the secondary.

 

[Speakerguy]

Removed....

 

[Hero999]

Lol did you actually carry out my suggestion (see quote below) or did you think of it for yourself?

You could also charge up a capacitor using a high voltage power supply then discharge is using a spark gap switch.