/diy/ - Do It Yourself

>>2761337
That's what I was wondering. Though I don't think the high voltage power supplies are the toughest part as he was saying, rather the electron emitting cathode and the beam deflection and focusing lenses seem like they'd be the toughest.

>>2761345
For very small things, or just as a proof of concept? I bet you could print solids to a similar tolerance to wire EDM, and with tighter radii, so making millimetre-scale or even micrometer-scale mechanicals might be a unique benefit of such a printer. Though growing monocrystalline silicon layer by layer via UV masking for MEMS seems to already do that pretty easily.

>>2761372
Rack Robotics' design is bad. See their github's issues page.

>>2761460
Resistive ballasts are a common way for making an EDM machine, and using an aluminium PCB as a resistor isn't a terrible solution (I'd just use nichrome in a bucket of water), but the real problem with it is it necessitates a medium voltage power supply (~72VDC). They're expensive. By using a system that stores current in a big inductor instead of in capacitors, the voltage for jumping to the workpiece is given from the kickback of the inductor. You could use an old ATX PSU or two for the base voltage, maybe a MOT primary for the inductor. The circuit to keep current flowing in the inductor is just a low-voltage half-bridge and a high-voltage single FET, pic related.
I based the design of this:
>https://tenebryo.github.io/posts/2021-04-07-edm-power-supply.html
But he goes the extra mile for bipolar pulsing, and is also a maniac for running it off rectified mains. Wouldn't go near that while it was cutting.