Where's the spark ??

You need the capacitor to get a sufficient rate of current flow, through the ionized channel, to sustain the channel and to do the work. The required instantaneous current flow is very high -- far more than the internal resistance of your transformers would allow.

Somewhere around here I have an article I wrote on EDM for American Machinist, 25 years ago, that explains some of the things you're probably looking for. It's 16 pages of magazine text and it's dated, but the basics haven't changed. It was used by vo-techs and some EDM manufacturers as a basic text on the subject years ago.

If I can find it, I'll scan it and send it to you. It's still copyrighted by McGraw-Hill so I can't post it anywhere. Remind me in a few days if I haven't gotten to it.

I'm sure you will receive many replies full of technical jargon and unworkable calculus functions but the REAL answer is to be able to align the bisecting planes of the electrodes tangent to both the absolute and implied potential electrical tunneling effect with the true magnetic north pole of the earth not true north. Of course, this is reversed if you live in the southern hemisphere.

the

implied

Horsey Cocka Doo Doo....

John

If your set-up reaches a high enough field strength to ionise the air, the current will almost certainly flow through the wood, it's not a good insulator. Substitute plastic, preferably not black (some black plastics are loaded with carbon) for the wood. Add a capacitor because the transformers have way too much resistance and inductance to get a healthy spark. Martin

Dear John

I believe air at standard pressure and temp kicks over at ~ 28.5 kV/inch

So with 80 volts you need 80/28500 or 0.003 " . So you should have no problem at 0.001"

Paul

Spark erosion machining with 2 HP Printer wallwarts, I think some 'industry' purchasing departments will be kicking themselves.

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 ( snipped-for-privacy@misty.com)

And since he's using pine wood against "brash" shim stock, that needs to be dilithium phosphate.

-- Tweetldee Tweetldee at att dot net (Just subsitute the appropriate characters in the address)

capacitor.

the

implied

of

the

Probably not. My experience is you'll never get a fat spark from a TV anode power supply over 1". More like 1/8" or so. You may get some ionization, but not enough to create a full conduction channel in air. So, I think you need a LOT more voltage per inch, maybe about 70 KV/inch, to get full breakdown.

Anyway, sparks in air just make noise and light, but don't do much to metal. Running EDM processes under a liquid seems to greatly enhance the metal removal effect. You can use water or hydrocarbon fluids, although the latter do pose a fire hazard. I have had good results using about 30 V at an amp.

You need a capacitor to deliver a burst of current, and then to let the spark die out as the cap recharges from the power supply. A light bulb can do pretty well for that. See my el-cheapo EDM at

Jon

That's a function of current. A milliamp or two of "average current" will typically, in my experience, get the voltage drop down to somewhere around 15-20 KV per inch. 30 milliamps will get the voltage drop down to somewhere around 6-7 KV per inch. A few amps plus some fumes of metal vapor will get the voltage drop to a few hundred volts per inch.

Varies from approx. 1 to 3 KV per millimeter, or 25 to 75 KV per inch, depending on how the electrode geometry affects evenness of the electric field between the electrodes. Can be less volts/distance at higher voltages over 40 or 50 KV or so. I have heard 20 KV/inch before as some sort of a "1-size-fits-all".

You will usually need a couple hundred volts or so to spark through air no matter how small the gap is because of the cathode fall of a "glow discharge". Voltage drop in a spark gap will be less once the temperature of the air rises. That couple hundred volts "cathode fall" will "collapse" to a much lower voltage once an arc process on the cathode forms.

- Don Klipstein ( snipped-for-privacy@misty.com)

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

Nope. The breakdown for short gaps in air (Paschen's Law) is:

Volts = 30,000(cm) + 1,350

...in other words, you get no spark at all for voltages below 1.35KV, and then the voltage rises like this:

Gap (millimeters) Voltage 0 1400 0.2 1900 0.4 2600 0.6 3200 0.8 3800 1.0 4400

For even better equations see: HIGH VOLTAGE HANDBOOK

Nope. As I understand it, in EDM the electrodes actually touch together. Since they're not perfectly smooth, they touch in one or two tiny spots. The discharge then blasts a crater in both electrodes, the capacitor charges back up, and a spring pushes the electrodes back into contact for a repeat performance.

And isn't this always done immersed in an insulating liquid? I'd think you'd want to prevent the parts from welding themselves together. A liquid would cool and quench out the arc, and also prevent new arcs until the two electrodes were pushing together hard enough to squeeze out the liquid from between the highest solid points.

That's all very interesting, but tap busters (primitive EDM machines that use air dielectric, and that are used for removing broken taps and drill bits in workpieces) typically have an open-circuit voltage ranging from 90V to 120V. They'll spark without contact, although, given the servo mechanisms on those things (often a hand crank ), it's hard to tell when you've actually contacted the work.

FWIW, more sophisticated EDM machines, which use oil for dielectric (kerosene with flash suppressants) have an open-circuit voltage that ranges from around 100V to 350V. Those aren't the power-delivering cricuits. The power supplies are cascade arrangements with high-impedance, high-voltage circuits to polarize the gap; medium-impedence, medium-voltage (~90V) circuits to ionize the channel and to initiate discharge; and low-impedence, low-voltage (~15V) circuits to deliver the amperage that does the real work.

In those machines, the electrode never contacts the work. They have a servo mechanism that maintains a gap on the order of 0.0002 - 0.003 in.

Ed Huntress

I wonder where this 300V (or 1350V from Jim Lux' page) actually comes from?

Since the rest of the Paschen equation is simply the e-field needed for air breakdown, the mysterious threshold voltage sounds like some sort of microscopic "screening field" at the surface of the metal, with half the threshold mystery-voltage appearing at each surface (so 1350V would actually be 675V at each metal surface.) Since any polarized air molecules near the metal surface will be attracted to the metal by image effect, maybe the image effect somehow acts as a proportionally-growing voltage barrier, and the air out between the electrodes never even SEES any e-field until the 1350V threshold is exceeded? Imagine something like a Helmholtz Double Layer appearing against the electrodes, but in air rather than in electrolyte?

If my above speculation isn't right, then there must be some sort of mechanism which prevents the air in the gap from simply breaking down when the usual value for e-field is reached. Why won't the spark appear at a field strength of 30KV/cm? Or in other words, what strange thing happens in the region within 1350/3000/2 = .23mm from each metal surface which both prevents any air breakdown there, and also prevents the air farther away from that region from breaking down when 30KV/cm is reached?

(((((((((((((((((( ( ( ( ( (O) ) ) ) ) ))))))))))))))))))) William J. Beaty SCIENCE HOBBYIST website snipped-for-privacy@eskimo.com

Eh? Wood was being used for an insulator for voltages well above this, well before you were born! It won't be bothered by 80V. The ionization comes from concentrating the field between the screws.

No, although healthier transformers is probably a good idea anyway. You need an RC circuit because current kills the metal, and you get much higher peak currents when you discharge a capacitor in microseconds than continuously releasing the same energy (which BTW won't work for EDM... you'll be arcwelding instead ;).

Tim

-- "That's for the courts to decide." - Homer Simpson Website @

Nowhere. Tell me, how is stick welding performed? ;)

Tim

-- "That's for the courts to decide." - Homer Simpson Website @

Yes, defenitely EDM is normally not done in air but in oil (die-sinker) or (deionized) water (wire edm). Pay attention if you use oil, it could begin to burn (the spark must absolutely not happen near the oil-surface but deep submerged). I would suggest to use water to begin becaus it's harmless... It is possible to erode also with 80V, but you need to get closer... (~20 micrometers?). If you don't want to use more power-supplys you could also build a step-up converter. Since the spark-gap-voltage is about 20-25V (depends on electrode and machined material),the best would be to use a source lower than that voltage. Diode -----------Inductor--------->| ------------------- __|__ | __|__ | electrode _____DC source _|/switch _____ C * | (below 24V) |\ | _____Workpiece ---------------------------------------------------

The switch can be a NPN-bipolar transistor or a mosfet. If you searches the web for a step-up converte circuit you'll surely get a more compensible picture as my bad ascii-art...

Can anyone tell me what use you would put "spark erosion" to that couldn't be done by any other way? I'm puzzled about it's advantages.