As someone who has seen a lot of EDM in action...
his whole edm thing creates a
very interesting paradox. I don't think it could work because as soon
as the circuit trace becomes isolated and loses it's electrical
connection, the EDM could no longer burn it.
You would need to burn your PCB in a dielectric liquid like a "sinker EDM" machine. The dielectric extinguishes the arc and carries away liberated material and keeps stuff nice and cool and clean. Normally sinker EDMs go up and down and up and down and up and down. Humm... would it be possible to use one of those iridium spark plugs with the little bent over "finger" clipped off as your electrode? Perhaps you could use an actual PCB drill bit (or router bit) bit as the electrode? They are very cheap and HARD and mfg to tight tolerances. If you change the diameter of the PCB drill bit, change the diameter of your cut. Perhaps use a traditional PCB routing technique in dielecric fluid, but electrify the router bit and route the board as if you were actually cutting it, but use the electrical arc to actually do the cutting. This would solve your rolled over copper foil from the mechanical cutting process
A way out there solution could be.....you could machine a negative of your PCB into a block of carbon then use an edm sinker to burn the whole board at once. That sounds like a pain, though. Perhaps you could print or transfer a conductive material onto a plate and then use that to burn your image as an entire plate onto the PCB.
This whole edm thing creates a very interesting paradox. I don't think it could work because as soon as the circuit trace becomes isolated and loses it's electrical connection, the EDM could no longer burn it.
Just some random thoughts thinking out of the box
-neal
________________________________
From: Reginald Beardsley <
pulaskite@...>
To:
Homebrew_PCBs@yahoogroups.com Sent: Wednesday, February 15, 2012 10:41 AM
Subject: Re: [Homebrew_PCBs] Re: PCB EDM
Andrew,
First thank you for taking so much of your time to reply.
What I have in mind is feeding a fine (e.g. 0.003") steel wire end on into the PCB w/ spark running until sparking stops & the wire hits the PCB substrate. Retract wire, move to next point and repeat. There is certainly an issue w/ loss of electrical continuity that has to be overcome when an area is being isolated from the board electrical attachment point. I'd expect to get small whiskers extending out from the isolated area which could become a serious QC problem.
The energy required to melt a copper circle .003" diameter and .0013-.0026" thick is very small, so even at very low efficiencies, it's still not a lot of power. I'm expecting 1-2 A peak pulses, but have not yet calculated what the fine steel wire will tolerate w/o adverse effects from heating. Having pricked my finger far too many times w/ a .01" steel string end 1/2" long, I'm not worried about the stiffness of the wire as it will be supported for all but a very short (i.e. < 0.1") distance. In the absence of actual data, I've assumed that surface tension would keep the wire tip approximately hemispherical, so by controlling the voltage I can limit the cutting area to the wire diameter.
The reason I'm not attracted to the milling cutter approach is two fold. Copper is difficult to machine because it is ductile. Normally it is cut w/ a diamond at very high speed. At slower speeds burr formation is a problem. There may be specially designed cutters that will work for PCB milling w/o being too fragile, but having cut a 3" slot in 1/4" steel w/ an ordinary 1/32" end mill, I don't relish working w/ even smaller sizes.
The genesis of this effort is learning to program the MSP430 on a TI LaunchPad. Not much sense to it if you don't build something. When it became apparent that the really interesting parts were only made in packages w/ insanely fine pitches I concluded that my original notion of milling one off boards was not as attractive as I first thought. Which leaves conventional photo or transfer processes as the best option if my EDM idea doesn't work.
I have a large number of things I want to do w/ the MSP430 which involve controlling stepper motors and electronic switches, so even if the project fails, I won't be wasting my time. Also I'm a research scientist and very used to failure. Rule of thumb I learned in grad school was 3 attempts for ordinary stuff and 4 if it was important. And some things you just never make work.
Thanks again for your time.
Have Fun!
Reg
[Non-text portions of this message have been removed]
[Non-text portions of this message have been removed]