Homebrew EDM power supply question

[rickturner646]

I'm starting to plan the build of a homebrew wire EDM machine, so have been looking at the sub-assemblies and components needed. One of these is the power supply, and since I've not been able to find one that does what I need (40-60v at 6A, with a chopper for spark generation) I sat down and drew up a suitable circuit from scratch yesterday.

About 3am this morning I had an idea that is either dumb (probably!) or might just be A Good Idea, so thought that I'd put it up here and invite comments.

I wondered if it would be possible to use (or mis-use!) a small MIG or TIG welder as the EDM power supply. They have the right sort of voltage (typically around 45v), can supply way more current than would be needed, and come pre-built with current limiters, thermal protection, heatsinks, fans, cases, ..... They're also no more expensive than a home built PSU - the transformer alone is about half the price of a small MIG welder.

The only thing that I wonder about is the duty cycle - an EDM unit is continuous operation for fairly long periods (tens of minutes) but then again, the current it will pull is very low compared to welding (about 0.5 - 1.0 amps, versus ~100A for a weld) so heating up should not be a problem.

So, if anybody has any comments, please lets hear them.

Regards

Rick

[Eric Allan]

Hi, I fix machines like these and the reason a a TIG or MIG wont do is the frequency and duty cycle, not the long running power on time but the actual cutting duty cycle.
Usually in megahertz for example 12usec on and 50usec off. this cannot be done with a welder.
My experience is mainly normal EDM's but I do have some thoughts re supply of erosion power.
Reply EricAllan@msn.com.au

[rickturner646]

I started to experiment with a welder as power supply, but have had to park it for a bit as my 'real' day job got really busy suddenly. The results I got were mixed - it would cut, but with variable quality - so I need to mess around some more to find out why. It might well be a duty cycle issue as you suggest, but I'm not sure yet - might also be to do with power density in the spark (due to a combination of current and cutting speed).

Rick

[Jet A]

To drag an old thread from the grave......

I have been working on a small sinker EDM for various reasons, but i believe the biggest hurdle for most of us hobby guys is the power supply.

I believe your small welder power supply theory to be valid. Of course there will need to be some external controls to handle the pulse generation and so on. My understanding of the erosion process is that a constant arcing will not cause an uniform erosion. Once an arc is sustained, the gap can be increased(not what we want). in a perfect world we would like perfect erosion across an entire surface. This is where the pulsing comes back into play. The pulsing allows the arc to find the closest surface to the electrode (path of least resistance).

Back to the power supply. I have been working on an external control board for my Chicago electric (harbor freight) tig welder. this little welder stack some impressive specifications as a power supply.(as a welder is another story)

Input-220v single phase 20amp

output 65v no load 35volts @ 130amps.

duty cycle 40% at 130 amps. 100% at 80 amps

OVERHEATING PROTECTION, OVER-VOLTAGE, OVER-CURRENT, and has all of the necessary internal handling of a dead short of the electrode.

This would be more than enough to sustain a constant process on a small edm. The dial on the front give you the ability to adjust current as need be for the exact material and size of electrode.

The external board- this external board will include, voltage control, pulse generation, current sensing & voltage sensing. For my particular application, it will also include a micro-controller capable of 80mhz processing w/ 8 cores "COGS". This will allow me to close loop mange the power supply, pulse generation and control my dcservo motor with a PID loop.

Pulse generation - This will consist of a high power MOSFET or IGBT, a drive for the transistor and the microcontroller. The micro will deliver pulses ranging from 1 - 20mhz, will be externally controlled via potentiometer for setup and then given a range at a latter date once it is fined tune. it seems like the 1khz to 10 khz range is popular. This will also give me control over voltage to the electrode. Using timing and duration i should be able to experiment enough to maintain a frequency and also control the voltage.

Current sense - will be accomplished with a shunt resistor or a chip produced by ALLEGRO. Feeding back analog voltage to the micro or possible digital signal through an a/d controller. This will allow me to monitor the current.

Voltage sense - This will be done via a voltage divider, and possible some form of isolation -optical

Motor controller - The motor is a small hobby motor with a USDIgital through shaft encoder. I will drive the motor from an H-Bridge controlled by the Micro. The encoder will be picked up for verification of movement and eventually PID that will set the feedrate.

How they all work together. A logical table must be devised to tell the micro what to do with the information. Current high and voltage low tells the micro to back away from the work (electrical short). High voltage and low current does the opposite (open circuit). the two within a predefined range will tell the motor to keep advancing slower or faster depending on the PID controls.

I want to reiterate that all of the above information is based on my own independent research and countless hours of forums. this ranges from basic electronics to micro programming. I have not tested any of these, but will soon have the external board complete to begin bench testing. I would be happy to post any results if there is any interest. I would like to hear any thoughts on the process and possibly some points that i may have missed or edm concepts that i do not understand.

[Drassk]

I'm eager to hear your results. I'd like to put together an EDM 'head' to stick in my mill to burn out taps etc, but it seems like a waste to make a non-pulsed version when the pulsed ones work so much better. Problem is that nobody's posted a tested pulsed design online that I've found.

I've got a 220A MIG sitting around 360 days a year, so using that would be handy!

[Perfo]

Did anyone progress this ? I started thinking about EDM for very small hole drilling only yesterday and read some of the various stuff on this here web thing.
Doesn't some MIG/TIG machines already have a mark space ratio adjustment? Mine doesn't as it's a cheapy but I too was thinking I've already got this MIG thing so would be good to use it if possible. I also thought the micro program should try and do a learning cycle IE not just bump in to the work pieces and back off then bump in to it again but start slow and incrementally up the speed until it bumps then back off one increment and hopefully unless something changes that should be the setting for the duration of the cut/cuts. Has anyone got any experience drilling small holes (0.1 mm ish) with EDM ?

[bozidar22]

check out the book by Ben fleming, The EDM how-to book.

it is fairly cheap to make and he sells the PCB for the controls.

EDMHomeBuilders : EDM HomeBuilders is were you can get the information .

I saw his EDM working at last years cnc work shop,impressed me so much I bought a copy and a PCB.

he is also working on a pulse type EDM.

the Digital machinest magazine is going to have a series on buliding a wire EDM.

[Perfo]

Thanks for that.
I'm not saying the book isn't worth every cent of $50 but I'm afraid that's too expensive for me. Especially as this is only one small project of many, many that I don't even really get to use. Saying that, if I was actually contemplating buying an EDM machine then of course $50 is peanuts. I like to share my knowledge and ideas with others for interest and community spiritedness but don't knock anyone that is trying to make a living out of it.

[bozidar22]

did you look into the EDM using a bridge rectifier, capasitors and a light bulb to limit current draw. can't get much cheaper than that.


that was my first EDM

[Perfo]

Thanks for the reply,
No I haven't seen that design does it set up an R/C oscillator with the capacitors? Could you point me in the direction of any more info on it?
I've seen the design using a solenoid out of a door bell and actually striking the work piece with the electrode thus energising the solenoid that then retracts the electrode breaking the circuit and electrode goes forward again and it cycles like that ad infinitum. This would probably be ok for one or two holes in thin stuff. I'm doing experiments with a thing called a Babington burner and they use something like a 0.35 mm hole for air. I can and have drilled these holes in brass with a Dremel type thing and that works ok but I'd like to be able to drill smaller ones ie 0.1 mm and do patterns of holes etc in harder stuff like stainless and thus I thought maybe EDM was the fellow to go for.

[bozidar22]

the desighn came from the home shop machinest magazine. the controls were completly manual. you have to control the feed by turning a knob. crude but works for burning out broken taps.


there was a guy at the cabin fever expo with set up using a car battery,pvc pipe wrapped with wire. the electrode bounced up and down like a solenoid as it burned.

PM-EDM (1 of 2) - YouTube

PM-EDM (1 of 2) - YouTube

[Perfo]

That coil thing is the sort of thing I had in mind. Very simple and good for the odd hole, not sure how my 0.1mm electrode would put up with the bashing in to the work piece though. I had thought of maybe using a hypodermic needle or even the tungsten filament out of a light bulb. I couldn't find any details about the capacitor and diode one you mentioned.

On a completely different note in terms of complexity. It seems to me that for a better machine, be it wire fed or plunge type we could do with a way of measuring the gap between electrode and work piece. Just thinking off the top of my head I'm thinking either capacitance or magnetic field could give us a pointer. If an AC signal was superimposed on the electrode then in between the pulse on cycles this AC voltage will electrically couple to the work piece and the strength of the signal measured. The potential of the returned signal may give an accurate enough measurement of the gap and used as feed back to the motor drives.