I’m currently trying to make an electrical discharge machining (EDM) machine that uses electrical discharge to erode away metal.

Right now, I’m just trying to test it out with a 200V 100uF capacitor, and a heating element acting as a resistor (measured at about 11 ohms) to limit the current (as I’ll literally be short-circuiting).

I'm using a flyback as the power supply and varying the voltage going to the ZVS driver to vary the output from the transformer.

I’m not getting any sparks at the receiving end when I put the electrode and workpiece close together in that plastic container full of distilled water.

The capacitor I think is still getting charged though (I disconnected it and when discharged made a lot of sparks).

Here is a video of what I'm trying to achieve

enter image description here

enter image description here

This is the circuit diagram I was going off of, just without the IGBT and pulse generator because this was just a test (planning to add later.)

Can someone help me troubleshoot the problem (I can provide more info) and what to change?

  • 1
    \$\begingroup\$ Well you want the current to be high - why the resistor? \$\endgroup\$
    – mmmm
    Aug 17, 2021 at 9:24
  • \$\begingroup\$ To limit the maximum short circuit current, just what I saw from other designs online so decided to add it. Although, the high voltage is the one to ignite the spark and it's not even doing that. The current is just what removes the material. \$\endgroup\$ Aug 17, 2021 at 9:27
  • 1
    \$\begingroup\$ A 'flyback'! You've connected to the EHT output. That is meant to give you 10kV at very low current. Not what you want to charge the capacitor. You want something more akin to a camera flash unit. Try getting a disposable film camera - these (used to) have a flash circuit in them. \$\endgroup\$
    – Kartman
    Aug 17, 2021 at 9:45

2 Answers 2


You have to pulse the discharge through the workpiece. As you have it, it will discharge once, then never be able to charge the capacitor again. You have only (at most) the current through the heating element.

Here's how it is supposed to work:

  1. Capacitor charges through the resistor.
  2. The electrode is moved to nearly touching the workpiece.
  3. The electrode and the workpiece are connected to the capacitor.
  4. The capacitor discharges a high current through the electrode and the workpiece, eroding the workpiece.
  5. The capacitor is disconnected from the electrode and the workpiece.
  6. The capacitor is charged through the resistor.
  7. The electrode is moved a tiny bit closer to the workpiece.
  8. Repeat from step 3 until the task is complete.

  • You can work from a low current supply if you give the capacitor enough time to charge between discharges.
  • A higher voltage means you can have a larger gap between the electrode and the workpiece.
  • If that high voltage flyback produces more than 200VDC, then you have destroyed your capacitor.
  • \$\begingroup\$ Thank you that makes much more sense now. Excuse my lack of knowledge but why does the cap only discharge only once and not recharge after being discharged? \$\endgroup\$ Aug 17, 2021 at 9:55
  • \$\begingroup\$ Your setup allows current to flow for as long as the electrode is close to the workpiece. Your power supply delivers a low current that can charge the capacitor but isn't high enough to "blast" bits off the workpiece. When the capacitor discharges, it delivers all of its energy in a short, high current pulse that can remove material from the workpiece. \$\endgroup\$
    – JRE
    Aug 17, 2021 at 9:58
  • \$\begingroup\$ Yes the cap is what allows for the actual eroding part. I'm still not sure why then the power supply will "never be able to charge the capacitor again". Also a piece of information that may or may not be useful is that the mosfets on zvs driver get fairly hot when connecting the flyback to the rest of the setup. The current also spikes. Not sure what this means. \$\endgroup\$ Aug 17, 2021 at 10:04
  • \$\begingroup\$ You put the electrode close to the workpiece. The capacitor discharges. The current from the power supply continues to flow through the electrode to the workpiece. Rather than charging the capacitor, the low current continues to flow through the electrode to the workpiece. You must disconnect the electrode and the workpiece from the capacitor to allow the capacitor to charge. \$\endgroup\$
    – JRE
    Aug 17, 2021 at 10:06
  • \$\begingroup\$ Ah that's much clearer! But even when I discharge the cap once I still don't get any sparks on the first discharge. Shouldn't it spark at least once? I get that it won't continue to charge the cap(after the first discharge), but shouldn't it charge it at least one time at the start? \$\endgroup\$ Aug 17, 2021 at 10:08

The flyback is woefully inadequate for this application. It outputs way too much voltage and way too little current.

You'd get much better results from a Variac + isolation transformer + diode bridge. Of course this becomes very dangerous.

The application, is 10-30A sparks, from 100-300VDC.

There must exist something to interrupt the spark current. When an arc occurs, the impedance of the arc goes from very high to very low, and will consume all of the available power, preventing another recharge from ever happening - indefinitely. Real machines pulse this current 50,000 times or more per second.

This high frequency may require the use of litz wire to the work to reduce losses. The whole EDM process is a huge generator of electrical interference - radio waves - don't be surprised if nearby electronics start misbehaving. Hopefully nobody nearby has a pacemaker.

The ionic conductivity of the water also matters. Cannot use distilled water, as it's conductivity is too low, requiring the electrode/work to be impossibly close, welding the tool and work frequently. And cannot use tap water, as the conductivity is too high, leading to electrolysis and loss of cutting resolution. EDM machines using water employ a conductivity tester and ionic filter to regulate the ion concentration. Others use a special oil. Both require exotic filters to remove particulate and control ion concentration.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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