LTSpice model for a SiC MOSFET

The grand-daughter of an old friend works for Microsoft and has found that asking one AI program to create prompts for a second AI program saves her a lot of time. She's married to a Ph.D. chemist who works on lithium batteries - as an undergraduate I seem to have chosen my friends remarkably wisely.

I'm not as well-placed as she is, and I'll wait until the old friend starts using it - with two sons who work for Microsoft he's well placed to get access when it is worth having.

Actually 1kV at 10uA in. The one 1mA was a constraint on the switching transistor based on the current that might circulate in the resonant tank.

Misunderstanding the constraints can lead people to propose inappropriate solutions.

Maybe use an electrostatic motor driving a small generator!

Good grief! A 10-megohm quarter watt resistor with a zener diode is the obvious answer.

How can anyone misunderstand something they have never been told?

A 3.3v real zener will probably leak more than 10uA. Low voltage zeners are awful.

And he wants 3 mA out... I think.

Bill was the confused person.

The problem gets much more interesting with 10 uA in and 99% efficiency requirement.

A Baxandall isn't going pull microamps at 1 KV.

Maybe DON'T FEED THE TROLL?

The guy I was talking to had 10uA at 1kV, and wanted a 3.3V output. 1mW at 3.3V is 3mA. The 3ma is the upper limit the circuit could supply.

You managed it.

John Larkin does like to think that.

There never was an efficiency requirement. 50% would have been nice.

Keeping a resonant tank resonating at a couple of kV is going to use up power in the winding resistance. I think I can get 245H out of RM14 core. If we can keep the parallel capacitance down to 25pF that would resonate at 2kHz and the reactive impedance would be 3 Mohm, giving a peak current 0f 1mA. The resistance of the winding is going to be about

A properly designed Baxandall might well pull less than 1uA if lightly loaded.

Geez, check your work before you post it. 3.3v * 3 mA is 10 mW.

I have a few kids that I'm teaching to design electronics. They already have EE degrees.

One of the fundamental concepts: check your work.

Build it.

It is an interesting problem, designing a logic supply that runs off a

Actually, the uncertainty opens up more circuit possibilities to consider. We can thank Sloman for being confused.

Finally, a purpose ...

You are probably best with bipolar. If you use a push-pull current fed half bridge the transistor breakdown rating will be 500*PI (1.57 kV). Since the off transistor will have a reverse bias of a few volts on it's base it's breakdown voltage will (typically) be a few hundred volts greater than it's Vceo spec. The 2sc4634/4636 would suffice.

Not until I've simulated it. That's a lot cheaper.

I don't think that John Larkin needs my help to get confused.

The first line of my original post was.

"I'm looking at a problem where somebody wants to step down a 1kV low current source to 3.3V. "

That's been the purpose all along.

Your (still confused and unresolved) specifications opened up a wide range of possible implementations. Thanks.

Since I've now at least got a PSpice model for the IXTH1N450HV 4.5KV FET, I can - with luck - get on with being confused by an actual (if simulated circuit). Typical Ciss is depressingly high at 1700pF.

The higher current (but lower voltage) Infineon IMWH170R450M1 has a typical Ciss of 506pF and is cheaper than $US45 IXYS part.

You may find the specification to be confused and unresolved. I haven't spelt it out in detail because - as the subject line spells out - that's not what I'm here for. As far as I can see I've been consistent about what I believe my contact is seeking, and what you are complaining about is the way you have misconstrued passing comments.

What are the requirements? 1 mA, 3 mA, 1 mW, 10 mW, some microamps, where? There have been all sorts of numbers.

Does cost matter?

The basic idea is that my contact has 1kV source capable of delivering about 10uA - 10mW - and wants to use it to power some logic from a 3.3V power rail. He'd like better than 50% efficiency, so presumaby he wants something approaching 1.5mA. About the only other number I recall posting was the idea that a switching MOSFET might need to handle up to

I've now had an interesting an potentially patentable idea that I'm obviously not going to post on a public forum. Like most patentable ideas it is unlikely to actually work, let alone be be practicable.

He does want it to be inexpensive. There doesn't seem to be any specialised market where a few customers can afford to pay a lot of money for a few devices, or if there is he hasn't mentioned it to me (and probably wouldn't) if there was.

I can imagine some cheap approaches if something like 25% efficiency were acceptable.

I finally got a Pspice model for the IXTH02N450HV out of IXYS and - with a bit of help from John May - got the part working in a simulation.

I did managed to get 1mA out at 3.3V but I was drawing 12.5uA out of the

There a 40kHz ripple on the current drawn from the 1kV suuply at about