high voltage pulse

[Jamez]

I am trying to build a power supply which will have an input of somewhere between 9VDC and 24VDC, and have an output of around 1800VDC pulses @ >20Khz (so as to not be audible) and a low duty cycle maybe 10-20%, so the output will basically look like a series of 1800VDC spikes, and should not go negative by very much, 180V negative would probably tolerable.. The load it will be driving will look like a capacitor in series with a resistor, 600pf and 1000ohms respectively. Any ideas how to do this? Flyback? Voltage multiplier? etc... is it even possible, with this type of load to charge and discharge it that fast? Any help is greatly appreciated.

James

 

[hacksaw]

your design requirements are incomplete, but suggest that you try looking on the web. a remarkable amount of chip based HV supplies are available.

 

[electricuwe]

That's quite a lot of power for a voltage of 9 V. The peak power will be 1800 V * 1.8 A = 3240 W and the average power will be (1800 V)^2*600pF*20 kHz= 40 Watt at least.

Not impossible, but not an easy task and for sure not done with a "chip based HV power supply"

Good Luck

 

[Jamez]

the power supply is for a high pressure plasma discharge lamp. It needs to have an input power of less than 50 Watts.
I have done plenty of research on the internet on chip based HV supplies. There is plenty out there in the way of Flyback driver ICs, high voltage IGBT's and what not...
right now I'm still in the feasability stage of the project and basically just want a heads up and what type of topology would be best. I realize my design requirments are incomplete, and vague, but I don't know exactly what duty, freq., power, etc. will be needed.

Thanks
James

 

[lcsjk]

This may be difficult but not impossible. Yes, you should be able to drive at that speed. Your minimum duty cycle pulse is about 5 us and your load time constant is 600nano-sec. So working within the 15-20% duty cycle is OK. I would consider two approaches. Use a switching power supply with a step-up transformer to give approx 1800volts and then use an N-channel mosFET to switch the HV to the output. The mosfet gate will need to be transformer coupled and also zener diode protected for max Vds (both ways).

A second option will be to use a step up transformer with an output rectifier to power the load and then switch the input. I am not sure this can even be made to work at your freq and Duty-cycle. THe transformer inductance may limit your on/off times too much.

You might also contact an applications engineer at National or LInear Technology.

 

[Jamez]

lcsjk,
Thanks for the input, that exactly what I wanted to hear. switching the output is one approach I was thinking of, I have some high voltage (2500V) IGBT's from IXYS that I was planning on using. Another approach I was considering was using an avalanche diode in series with the lamp, when it reached breakdown voltage it will dump the current into the lamp, and make some sort of a relaxation osscilator, don't know if that would work, also don't know what frequency it would do.. With an IGBT I can control the frequency/duty cycle, etc.....