A little background: I'm a mechanical engineering graduate student with a focus in materials, so electrical engineering is not my core background. I'm conducting research on plasma surface treatment of metals and there's a lot of variables I'm trying to control reliably, but when it comes to the electrical aspect of the project I'm struggling a bit. Here's a rough schematic of the electrical aspect of my text fixture:

enter image description here

Starting with the left box, from left to right:
DB1: wall power, 110V, 60 Hz, 20 amp breaker
Bridge rectifier, full wave, single phase: 50A max forward current, 1000V reverse voltage
C1: 500uF capacitor
L2: Self wound inductor/choke 50 turns high voltage side, 50 turns low voltage side, 55mm OD x 35mm ID x 14mm H, wound with 12 gauge magnet wire (I haven't measured the inductance, sorry!)
D1: 6amp 20kV diode
V2: spark gap, basically two 12 gauge copper wire "rails", sandwiched between insulating sheets (PTFE), and two grade 52 neodymium magnets. High voltage ignites a plasma between the rails, and the lorentz force accelerates the plasma, out of the gap and onto test samples.

R1: Power resistor bank to avoid tripping the breaker

Right now the box on the right is a cheapo AA battery powered High voltage boost converter (~20kV).

The circuit on the left provides a (relatively) high current-low voltage DC power supply, and the circuit on the right strikes the spark gap with high voltage and ignites the plasma.

The issue I'm running into is that I really need to push the current level of the plasma jet to continue pursuing my research. The diode was $40, and I've burnt out 3 so far due to overcurrent(1 single, and then 2 in parallel). I'd like to get the current to around 50 amps, but the high voltage diodes are extremely cost prohibitive.

So here is my question: Is there an equivalent or similar circuit, free of high current high voltage diodes, which can ignite a plasma with a high voltage/low current source, then immediately dump current into that plasma from a low voltage/high current source. The boost converter results in a pulsed system, it would be a bonus to have a semi-steady state system.

Thank you!

  • \$\begingroup\$ Have you considered large diameter Graphite rods instead. Carbon Arc sky beams use this. \$\endgroup\$ Commented Mar 9, 2018 at 7:27
  • 1
    \$\begingroup\$ I'd start by understanding the plasma (at least 2 spatial and one time dimension of PDEs coupled to at least 6-dim ODEs, along with radiation transport and atomic interactions.) Global rate equations assuming spatially averaged densities for the charged particles and neutral atoms and molecules in excited and metastable states may work? A text by Lieberman and Lichtenberg, Principles of Plasma Discharges and Materials Processing, treat the basics: discharges, global models, collisions, and DC and RF discharges. I think you need to describe the physics first, design later? Or, disagree? \$\endgroup\$
    – jonk
    Commented Mar 9, 2018 at 7:35
  • \$\begingroup\$ This looks like a great resource, I'll get to reading it! \$\endgroup\$ Commented Mar 9, 2018 at 7:43
  • \$\begingroup\$ Maybe use a high frequency high voltage source like the little modules advertised on eBay, and place the diode in front of the inductor instead of after it? \$\endgroup\$ Commented Mar 9, 2018 at 8:43
  • \$\begingroup\$ Part numbers are preferred in situations like this. What is D1? \$\endgroup\$
    – Andy aka
    Commented Mar 9, 2018 at 9:06

1 Answer 1


Take a leaf from the cinema projection trade (xenon lamps) and use a series injection transformer to eliminate the diode, they are standard spare parts from the suppliers of parts to that trade (There are a few left), see also the series injection choke based igniters often used in old school ion lasers.

50A on the main current side is routine, and they easily strike an arc in a high pressure lamp.

You may incidentally wish to investigate coax arcs, very funky physics at the point the arc collapses, to the point that if you try it in a deterium atmosphere you actually get a detectable number of neutrons produced.

  • \$\begingroup\$ This sounds like a great recommendation! Thank you for your help. \$\endgroup\$ Commented Mar 16, 2018 at 21:15

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.