Don Klipstein wrote:
In article , John Albers wrote:

I am thinking about building an EDM (electric discharge machine) which
remove tiny bits of metal by sparks. I have found plans and
schematics on the web and in various magazines like Home Machinsts
Workshop. I could just blindly copy their designs, but I wanted to
understand how they work first. Most are build around a relaxation
oscillator circuit which charges a capacitor and the capacitor
discarges across a spark gap. Most of the EDM machines that I have
plans for work with voltages around 80-100 volts. Expreimenting helps
me understand things better, so I tried the following experiment:

I took a couple of old HP deskjet 40V DC wall warts and wired them in
series to produce 80V DC. I checked this with a DVOM and I am getting
80V DC output. According to various sources the break down voltage of
air is around 20V per .001 inch. I have a block of wood (pine) with a
notch in it holding two machine screws facing each other to create a
spark gap. Using a piece of .001 inch brash shim stock, I adjusted the
gap to be .001 inch. I am thinking that if 80V is supposed to jump a
.004 inch gap, then 80V should have no problem jumping a .001 inch
gap.

I am not getting a spark. Should I be getting a spark with this set
up?


I believe it takes around 200, maybe 300 volts to break through air no
matter how little, and where the voltage and/or electrode geometry do not
favor corona and where the electric field within the spark gap is evenly
distributed, it takes about 75 volts beyond that 200-300 volt figure per
.001 inch (3 volts per micrometer), give or take a little depending on
temperature and air pressure.
I suspect EDM requires actual contact or some sort of assistance
("ignition pulses", UV, photoelectric effect, whatever) to spark at 80
volts.

- Don Klipstein )

Interesting Don, it usually takes some minimum in oil also, I did not
know there was a lower I would say Weibull limit (called a location
parameter and usually 0 for many things!) here but it sounds correct.
There must be some empirical evidence here but this sounds like the
problem why it won't work for 80 volts no matter what.

Needs more of those 40 volts sources, say 10.
Paul