Facebook
Google
GitHub
Linkedin
So I'll preface my remarks my pointing out that I'm not another lame hobbyist building an experimental ion thruster, although I could stray into that flaky realm after retirement.
Elon got rich building novel rockets and maybe I can, too. Probably shouldn't have waited until I was in my mid-70s to do it.
This is a genuine revenue project involving whacking piezoelectric transducers with 1kV pulses. These pulses are time-aligned with an FPGA to exert hundreds of megapascals of peak force on an object my boss insists cannot be named without costing me my job.
Functioning prototypes of this shock wave device were built repurposing high-voltage ferrite transformers from the era when LCD displays were backlit with CCFL lamps. The transformer's secondary has one (and only one!) Cockcroft-Walton stage,making it the voltage doubler configuration.
The desired power output on the 1kV output port is a mere 2 watts, less than 0.5 part per billion of Hoover Dam.
In fact, my C-W multiplier architecture was discussed ad nauseum by others in this forum 8 years ago: the full-wave version with an untapped transformer secondary:
https://forum.allaboutcircuits.com/...w-multiplier-w-o-center-tapped-driver.135119/
I use an off-the-shelf PWM IC (MIC3809) to drive the CCFL transformer's primary at 100kHz. The charge pumping capacitors on the secondary are a scant 33pF each because the switching frequency is so high. Any stray capacitance on the "moving" HV nodes is catastrophic to the circuit's performance. Even a 100x scope probe dramatically reduces output current.
My question: what kind of high voltage fast recovery diodes do I need for the C-W muliplier? Clearly they need to be rated at greater than 1kV to avoid avalanching. Seems that junction capacitance is also a vital specification given my 100x scope probe loading observation.
The circuit works with ordinary US1M 1kV fast recovery diodes but the conversion efficiency sucks.
Regards from Indianapolis US.
Elon got rich building novel rockets and maybe I can, too. Probably shouldn't have waited until I was in my mid-70s to do it.
This is a genuine revenue project involving whacking piezoelectric transducers with 1kV pulses. These pulses are time-aligned with an FPGA to exert hundreds of megapascals of peak force on an object my boss insists cannot be named without costing me my job.
Functioning prototypes of this shock wave device were built repurposing high-voltage ferrite transformers from the era when LCD displays were backlit with CCFL lamps. The transformer's secondary has one (and only one!) Cockcroft-Walton stage,making it the voltage doubler configuration.
The desired power output on the 1kV output port is a mere 2 watts, less than 0.5 part per billion of Hoover Dam.
In fact, my C-W multiplier architecture was discussed ad nauseum by others in this forum 8 years ago: the full-wave version with an untapped transformer secondary:
https://forum.allaboutcircuits.com/...w-multiplier-w-o-center-tapped-driver.135119/
I use an off-the-shelf PWM IC (MIC3809) to drive the CCFL transformer's primary at 100kHz. The charge pumping capacitors on the secondary are a scant 33pF each because the switching frequency is so high. Any stray capacitance on the "moving" HV nodes is catastrophic to the circuit's performance. Even a 100x scope probe dramatically reduces output current.
My question: what kind of high voltage fast recovery diodes do I need for the C-W muliplier? Clearly they need to be rated at greater than 1kV to avoid avalanching. Seems that junction capacitance is also a vital specification given my 100x scope probe loading observation.
The circuit works with ordinary US1M 1kV fast recovery diodes but the conversion efficiency sucks.
Regards from Indianapolis US.
Attachments
Last edited:
