# Making small sparks

#### shortbus

Joined Sep 30, 2009
10,049
Onward with my project. This is for an EDM, electrical discharge machine. Instead of a normal electronic circuit that avoids making sparks, the whole idea is to make shot time controlled sparks. The sparks are to happen between a shaped electrode and a conductive metal work piece, while immersed in a light viscosity oil bath. To make the shape of the electrode into the work piece by eroding the metal with the sparks, the oil then carries the eroded metal away. This is a common machine shop way of machining hardened metal or difficult shapes.

The first type of this machine used a RC(resistor capacitor) type circuit. It then changed to a "pulsed" circuit over time, into today's CNC high tech machines. My desire is to come up with a simple pulsed EDM circuit. This has been a very long time project that I would love to complete be for kicking the bucket, kind of a bucket list project. I've had the mechanical parts of this done for a long time, my mechanicals have even been pictured in a book by another guy, Ben Flemming, that sells plans for making DIY EDM's. But I don't care for his way of doing the spark generator. Even he agrees it's not optimal but it is easy.

I've attached my latest idea for my spark generator. I don't show the timing circuit, because thanks to this, AAC, site I've got that covered now. Thank you to all of you that helped for that. Now maybe with your help I can get the power side of the circuit figured out.

Please take a look at what I've come up with. The low side mosfet will turn on and charge the inductor, then be turned off. Due to the lower resistance of the spark gap, compared to the load resistor, when the mosfet turns off, the kickback will cause a spark in the gap. Or at least this is my thinking. Does it look feasible? Please give your thoughts, thank you.

#### #12

Joined Nov 30, 2010
18,224
The math for turning off an inductor suggests that the resulting voltage can go to infinity if there is nothing to absorb the energy. Look at a spark plug oscilloscope. It doesn't tell how good the inductor is, it tells the spark plug gap!

Therefore, my advice is to include something to protect the mosfet in case somebody lights up the machine with no load in place.

#### Alec_t

Joined Sep 17, 2013
14,409
Methinks you'll need to drive a step-up transformer. Try googling for 'electric fence circuit'.

#### #12

Joined Nov 30, 2010
18,224
Why would you need a step-up transformer when a simple inductor can deliver thousands of volts?
Because you can't buy a mosfet better than 1000 volts rating.
Is 1000 volts enough for sparking parts?
You know.

Joined Jul 18, 2013
28,829
I have a schematic for an old Elox commercial Sinker EDM, they were once big in EDM, they used 8 parallel power transistors for pulsed arc operation.
I believe around 1kh, they also used simple Cmos electronics and only use 70vdc for the arc.
Are you making a wire or a sinker?
You do know the +ve electrode is the work and the die the -ve?
Max.

Last edited:
#12

#### shortbus

Joined Sep 30, 2009
10,049
Max, it's a sinker. Yeah, ran EDM machines for ~13years. When I made the sketch I wasn't going to do the long explanation so just drew it to show a spark gap. The open circuit voltage will be ~95VDC, the current carrying voltage step is between 35 and 25 VDC, these numbers are from the many books I have on EDM.

The plans that Flemming sells world wide just have the work connected to V+ and the mosfet(s) as a low side switch. The gap wave form doesn't always make plasma to break down the dielectric oil though. many missed pulses. Some of the patent drawings show a inductor, and one of the members of a DIY forum said he thought just a switched inductor would do it. That forum has since pretty much died out.

We had several old Elox's at work, but the power supplies were junk, they were only used when nothing else was open for a job.

Last edited:

#### shortbus

Joined Sep 30, 2009
10,049
#12 and Alec, the open circuit voltage is only 95VDC, an EDM uses fairly low volts but high amps. The gap is only around 0.002 inches. The spark will/should quench the voltage surge, it needs a higher than open circuit voltage to make the spike at the beginning of the gap wave form, to break down the oil/dielectric and make a 'plasma' that then allows the current to rise between the electrode and work. This melts the metal and then the oil rushes back in and takes away the small amount of cooled metal from the gap. The melted metal actually makes very small spheres when every thing is working correctly.

#### cmartinez

Joined Jan 17, 2007
8,285
I'm very much interested in this thread... I've always wanted to build a plasma machine, and I have the plans for a plunger type somewhere lying around. But I'm finding your mosfet-inductor circuit more appealing

#### shortbus

Joined Sep 30, 2009
10,049
cmartinez, who's plans do you have? To my way of thinking the ones from Langlois out of the Home Shop Machinist magazine are the best. This project is the reason I got involved with electronics and joined AAC. Wish I knew more about it though. Making the sparks seems counterintuitive to most people though. The inductor wouldn't need to big of a value, just enough to kick start the spark in the gap.

#### cmartinez

Joined Jan 17, 2007
8,285
cmartinez, who's plans do you have? To my way of thinking the ones from Langlois out of the Home Shop Machinist magazine are the best. This project is the reason I got involved with electronics and joined AAC. Wish I knew more about it though. Making the sparks seems counterintuitive to most people though. The inductor wouldn't need to big of a value, just enough to kick start the spark in the gap.
Haven't found the plans yet, they're probably at my shop, which I won't get to until Tuesday. Anyway, I do remember they involved a voltage multiplier using the traditional zig-zag array of diodes on an input AC waveform, and after that it rectified the voltage to DC... I'll get back to you as soon as I find them, promise.
EDIT: BTW, if you want to tag someone, you need to add the @ symbol before that, like this: @shortbus

Joined Jul 18, 2013
28,829
I picked up the book by R. Langlois many years ago and had intended to build a higher strength one but never got around to it, his uses the capacitive bank discharge method, advance the electrode, as soon as the voltage collapses, retract until the voltage builds and advance over again.
I have rebuilt Tap Disintegrators but they use a slightly different principle, of blasting the broken tap etc, by the use of very low AC voltage at high current.
Max.

#### shortbus

Joined Sep 30, 2009
10,049
The Langlois circuit is one of the RC based ones, a guy named Jim Glass even built a wire EDM based on that power supply. As did Langlois in a later HSM article, but it didn't make it into the book. The RC charges the cap through a higher value (~15 ohm) until the voltage gets high enough to start the plasma and current flow in the lower resistance( ~1.5 ohm) gap.

You have the ram a little wrong. The electrode only should retract when it gets a short circuit. And advance when it gets an open circuit.

In the back part of the Langlois book he gives a prototype circuit of a pulse EDM using mosfets that many have said is unworkable. I don't have enough knowledge to argue one way or another.

#### shortbus

Joined Sep 30, 2009
10,049
cmartinez, you and I may be talking about two different types of machines? The use of a Cockroft-Walton circuit sounds more like a Plasma cutter, instead of an EDM. The plasma cutter is like an electronic oxy-acetylene torch. The EDM is more like a mill or drill press.

Joined Jul 18, 2013
28,829
You have the ram a little wrong. The electrode only should retract when it gets a short circuit. And advance when it gets an open circuit.

.
I think basically that is what I said or at least meant, "Advance the electrode, as soon as the voltage collapses (shorts), then retract until the voltage builds and advance over again"
His LM339 op amp monitors the cap voltage and implements a fwd and rev of the stepper motor, .
This is the way many of the old Elox worked using a Hyd servo valve for ram control.
Max.

#### cmartinez

Joined Jan 17, 2007
8,285
cmartinez, you and I may be talking about two different types of machines? The use of a Cockroft-Walton circuit sounds more like a Plasma cutter, instead of an EDM. The plasma cutter is like an electronic oxy-acetylene torch. The EDM is more like a mill or drill press.
Nope... I know exactly what a plasma machine is, and I've been working with them for years. The machine I'm talking about is an EDM, but it's not a wire EDM, it's a carbon-electrode plunger-type EDM, used for making orifices. It works through the setup that Max just mentioned.

Joined Jul 18, 2013
28,829
Commonly know as a 'Sinker' EDM, BTW I have a Ecopy of the DIY book by Ben Fleming.
Max.

#### cmartinez

Joined Jan 17, 2007
8,285
Commonly know as a 'Sinker' EDM, BTW I have a Ecopy of the DIY book by Ben Fleming.
Max.
We might very well be talking about the same book... But I just have to find mine first!
Can you post an image of the high voltage generating circuit?

Joined Jul 18, 2013
28,829
pdf.
Max.

#### Attachments

• 604.1 KB Views: 149

#### cmartinez

Joined Jan 17, 2007
8,285
pdf.
Max.
No... that's definitely not the same circuit I have. Which is a good thing because now we can compare the two. I'll see if I can post mine tomorrow. Thanks.