Welcome! Log In Create A New Profile



Posted by Joseph 
November 02, 2013 05:29PM
Does anyone have a EDM schematic, of a proven build, that they can share? We want to do some experimenting with EDM and creating nozzles.

Thanks in advance!

November 02, 2013 05:37PM
This video instructional looks simple:

Make a doorbell EDM machine to burn small holes into the nozzle.
You could even burn the ID profile! Use a dielectric fluid, and a plastic guide for the electrode.
Use a sewing needle as the electrode.

Originally posted in this thread:
November 02, 2013 06:38PM
Have you seen [www.homebuiltedmmachines.com]

One of the local guys is building one of these
November 03, 2013 08:58AM
* What is the rate of erosion, (nonsystematic)error of this? To be honest i think it might be slow for use as for instance an alternative toolhead.(But fast enough to make holes for hot ends) there is already a method, but more is better.
* Rate of erosion of tool. Graphite is cheap and probably doesnt erode quickly?
* Do there need to be measures to remove material? Something like air jets sound more convenient than oil?
* What sort of power supply works well? Measures to avoid damaging electronics needed? (5V@1A doesnt sound menacing other than starting fires, but they're sudden pulses, inductance might work up the voltage)
* Number of windings depending radius, expected current, weight, ferromagnetic material/magnet strength needed.
* Jeri Ellsworth is awesome.
November 03, 2013 10:04AM
A lot of folks have made DIY EDM machines, particularly from the CNC crowd. Around 2001 (before there was a reprap smiling smiley ) I made a contraption to do what I call "chisel EDM" which is sort of in between sinker EDM and wire EDM. The idea is that you use a metal blade as the cutting tool, and then cut out patterns in metal sheet. It relates to self replication because you can use, say, an aluminum blade to cut out another blade from aluminum sheet.

Here are some pictures of what's left of the project. It was no work of art, and it has been sitting in a box for more than 10 years and some of the parts have been removed for other projects. But the bottom line is: it worked! EDM would be a super add-on to a RepRap machine. It is totally, totally, feasible.

I used either vegetable oil or distilled water as the dielectric fluid. The power supply went up to 40V, and there was a switchable capacitor box to adjust capacitance. A little (terribly designed) op amp circuit measured tool-to-work voltage and controlled the vertical position of the tool (this was done with a continuous rotation RC servo and a 6-32 lead screw). Power resistors were used for current limiting.

When properly tuned, this setup would cut nicely through 0.010 inch aluminum in a few minutes.

Light bulbs are often used in EDM setups as the current limiter. They work great for this. Also, if I do this project again (someday maybe) I would scrap the op-amp circuit and just use an Arduino.

Control boxes. Power resistor block at top. From left to right: switchable power supply, capacitor box, toolhead controller, XY controller.

Mostly complete machine. This is missing the servo-leadscrew mechanism that would raise and lower the toolhead. Also missing the toolhead and toolhead mount.

Toolhead closeup. This is the blade or "EDM chisel". The servo here is not the servo used to raise and lower the blade. This servo is for rotating the blade to get cuts at different angles.

Example cuts. Again, this was not a work of art smiling smiley. You can see that when properly tuned it is possible to get a straight clean cut. I think the holes are attempts from using a wire as a cutting tool.

November 03, 2013 10:12AM
Joseph, this Hardware Hacker has an ultra-simple (and dangerous!) EDM schematic using light bulbs and mains electricity: Hardware Hacker 63. If you're making a small setup, you can probably use a safer power supply than straight mains and this sort of circuit should work OK.
November 03, 2013 10:58AM
This diagram shows how all the pieces tied together. I made the supply voltage, threshold voltages, main capacitance, and current limiting resistor all adjustable so that the optimum settings could be discovered.

If you are doing something as small as a nozzle, you should be able to get away with low voltages. You probably could do the whole thing with a slightly modified printer. Hold the tool fixed (where the extruder goes) and use the Z platform to move the work up/down for arc control. A simple voltage divider piped into a spare pin on your printer controller can read Vsense (some people might be worried that such a noisy signal from the arc might damage the pin - you can add a filter if you're worried).

The control algorithm is simple and you could probably add it to any of the popular firmwares.

November 03, 2013 11:05AM
Also, for cutting holes in nozzles you could probably use "lead refills" for mechanical pencils as electrodes. They're graphite, they're easy to get, you can get them in 0.3mm, 0.5mm, 0.7mm diameters, etc.

(Sorry for all these posts. Can you tell I get excited about this topic? smiling smiley )
November 03, 2013 11:37AM
I read an article about this recently in which a cylinder (for a steam engine) was machined from a casting - obviously industrial scale. The article seemed to assume the readers were familiar with the technology. I had assumed that it would require a complex dangerous high voltage power supply.

But it seems from the earlier posts here that it works fine at a small scale with relatively low voltages. Is that correct?

What is the purpose of the fluid?

November 03, 2013 12:09PM
... it works fine at a small scale with relatively low voltages. Is that correct?
In my experience, yes. My little setup only went up to 40V, and it worked OK at 20V too. For something as small as a nozzle hole you might get away with 10V or less.

What is the purpose of the fluid?

The fluid washes away the tiny bits of metal and debris that are blasted free by the tiny arcs. It probably helps to cool things down too. Also (especially if you're doing something weird like cutting aluminum with aluminum) it helps prevent the tool from spot-welding onto the work.

You can blast through a chunk of metal without using fluid, but the quality of the cut will be poor compared to one made using fluid.
November 03, 2013 12:46PM
Dielectric Fluid

Sinker Dielectric Fundamentals:
• The dielectric oil acts as a medium through which
controlled electrical discharges occur.
• The dielectric oil acts as a quenching medium to cool
and solidify the gaseous EDM debris resulting from
the discharge.
• The dielectric oil acts as a medium used to carry away
the solidified EDM debris from the discharge gap to the
filter system.
• The dielectric oil acts as a heat transfer medium to absorb
and carry away the heat generated by the discharges from
both the electrode and the workpiece.

Electrical discharge machining:
When the distance between the two electrodes is reduced, the intensity of the electric field in the volume between the electrodes becomes greater than the strength of the dielectric (at least in some point(s)), which breaks, allowing current to flow between the two electrodes. This phenomenon is the same as the breakdown of a capacitor (condenser) (see also breakdown voltage). As a result, material is removed from both the electrodes. Once the current flow stops (or it is stopped – depending on the type of generator), new liquid dielectric is usually conveyed into the inter-electrode volume enabling the solid particles (debris) to be carried away and the insulating properties of the dielectric to be restored. Adding new liquid dielectric in the inter-electrode volume is commonly referred to as flushing. Also, after a current flow, a difference of potential between the two electrodes is restored to what it was before the breakdown, so that a new liquid dielectric breakdown can occur.

Effect of Composition of Powder Mixed Dielectric Fluid on Performance of Electric Discharge Machining (2011):
material removal rate (MRR).
tool wear rate (TWR).
Jeswani [3]
studied the effects of 4 g of fine graphite powder per
litre of dielectric fluid (kerosene) and found out that this
addition increased the MRR by 60% and TWR by 15 %
in EDM. The machining conditions used were for
roughing only and not for finishing operations.
November 04, 2013 12:05AM
Thanks all.
Something to keep in mind for the future.

Because I hadn't imagined it would work at a hobby scale it hadn't occurred to me to try Wikipedia.

November 04, 2013 12:45PM
I am afraid I still didn't get it fully.

Does that mean I just put a needle into my nozzle, clamp it to 24 V and "spark away" those parts of a metal block that I would have needed a CNC for otherwise?

Is it possible to build a printed lathe for that purpose?


Excalibur Hotend

November 04, 2013 03:51PM

Does that mean I just put a needle into my nozzle, clamp it to 24 V and "spark away" those parts of a metal block that I would have needed a CNC for otherwise?
Well, the material removal rate of EDM is quite a bit smaller than that of milling, but basically: yes, that's it.

I just wouldn't clamp it into the nozzle to avoid ruining it. And you need an interrupted voltage, else your arc is uninterrupted and won't break away anything. As always, the details are a science on its own.

Generation 7 Electronics Electrochemical Machining RepRap DIY
Teacup FirmwarePCB Milling
November 05, 2013 06:12AM
I wish to thank the RepRap community for their very informative answers to my question. Especially want to thank MattMoses. Your enthusiasm is contagious. smiling smiley

As we progress with our experiments we will post any info on this thread.

Again, y'all are just great. Thank you.

November 05, 2013 08:26AM
You're welcome! smiling smiley I hope the information is useful to you and others.

I made one other picture to help answer theodleif's question. This is how I would outfit a reprap for doing some EDM experiments:

Also, here are the range of values I used in my old setup:
Supply voltage: 10V, 20V, or 40V
Current limiting resistance: approximately 100 ohm, 50 ohm, 33 ohm, 25 ohm, 20 ohm, 17 ohm (six 100 ohm, 10 watt power resistors that could be hooked up in parallel)
Main capacitance: six capacitors which could be switched in in parallel: 4700 uF, 210 uF, 100 uF, 20 uF, 20 uF, 20 uF
November 05, 2013 08:34AM
Also, to answer theodleif's other question (although I expect he probably already knows the answer winking smiley )...

Is it possible to build a printed lathe for that purpose?

Of course! I think that is one of the applications that Sublime had in mind when he made his printable lathe.

You can also drill the tiny nozzle hole by hand if you practice the technique: example here.
November 05, 2013 09:50AM
Does the up/down motor yield useful control or is the gravity+spring+controlling time of application idea basically good enough? The former has better control i expect..

I read those articles A2 posted. Dont see how anything in there can do better than vegetable oil, or how it would be better, to be honest. Also no idea how bits of carbon in the oil helps. Maybe it smoothes out the electric fields by conducting a bit or something.(so the break due to the material not resisting the electric field is smoother too) Anyway i expect that 'just vegetable oil' is the answer for now. Also looked for estimates for erosion rate a bit, havent found much yet..

I also put up a wiki page.(looks like images in the forum can immediately be used in the wiki)
November 05, 2013 12:17PM

I also put up a wiki page

Excellent. Thank you.


Does the up/down motor yield useful control or is the gravity+spring+controlling time of application idea basically good enough?

I guess you have two choices here. One is to sink very slowly to make sure you never exceed the material removal rate.

Professional machines measure the current actually flowing. The closer the gap between electrode and workpiece is, the more current flows. The goal is to keep this current constant, which is achieved by sinking the electrode according to the current measured. A simple regulating circuit: current too low -> do a motor step downwards. Current too high: don't move.

You end up with sort of an automatic rapid feedrate. As long as the electrode is far away from the workpiece, (zero current) you constantly step and slow down as soon as the first sparks happen.

Professional machines also interrupt the voltage actively, with controlled on/off times. To me this sounds like the PWM-able heater outputs of RepRap electronics are just made for this.

Generation 7 Electronics Electrochemical Machining RepRap DIY
Teacup FirmwarePCB Milling
November 06, 2013 01:01AM
A little while ago I started building an EDM machine based on discussions that I had a Model Engineering show in Warwickshire. Although the EDM machine is in abeyance until I have become tired of 3D printing, I did take lots of notes and waded through many web pages trying to get to the best simple control that I could. The logic for most of them seemed to be:

1) Advance until current flows
2) Back off if current becomes excessive (Short circuit)
3) Back off anyway after a time - a few milliseconds. This keeps dielectric fluid moving in and out of the eroded cavity, flushing away particulates and cooling the tool.
4) While not down far enough, goto 1

Something equivalent to the modified doorbell circuit would probably work very well, although it would need to be adjusted for the job that it is doing and would have no flexibility outside that.

November 17, 2013 10:19AM
I put some more stuff on the wiki page, for those that are interested.
November 17, 2013 02:23PM
Would you advise EDM for making tiny PCBs?


Excalibur Hotend

November 17, 2013 06:58PM
Would you advise EDM for making tiny PCBs?

It could probably be done, but I doubt it would work better than the more established methods.

For doing really tiny PCB work with a "kitchen setup" it seems like the most promising approach would be a direct-print etch resist, kind of like this project or this project or this project, and a spray etcher for a fast, even etch (see homebuilt spray etchers here and here).

VDX has done some pretty impressive patterns in etch resist using lasers (see here and here).

And Traumflug did some experiments with electro *chemical* machining (ECM) for making PCBs here with mixed results. I've done some additional work with ECM and got it to work OK, but I don't think it will ever compete with lasers or direct-print for making PCBs.
November 17, 2013 11:58PM
... I've posted some results here in the forums too - [forums.reprap.org]

The thinnest engraved lines with the 445nm-'blue' laserdiodes are between 0.03 and 0.05mm, with the IR-diodes around 0.1mm ...

November 28, 2013 07:13AM
EDM is probably good for materials that other machining has trouble with, and maybe has a 'cheap-initial cost' way to slowly do regular matching?

I see you have much improved the wiki page smiling smiley Added a bit in text about how the doorbell thing works. I think at least for nozzles that approach is much simpler. It may be good for as moving tool on the reprap tool, the firmware would have to go forward when it stops jumping back.

That'd seem less complicated, and less wear to me than measuring the current, and moving up using the machine itself.(you could tie it to the filament and pull the filament back i suppose, i have undertested designs for doing that with a pen plotter, i think moving the threaded rods is more prone to wear?)

I 'deconstructed' an ancient digital (photo)camera and found a tiny solenoid in it, tried to make the mechanism for EDM with it(just a stationary one), but failed so far, wires broke at some point, it was too wobbly.. I'll keep trying on a bit, maybe just scavenge the thing for the wires.(the housing may be ferromagnetic and serve a function..)
November 29, 2013 08:36AM
That'd seem less complicated, and less wear to me than measuring the current, and moving up using the machine itself.
Yes. I think the main disadvantage of the solenoid method is that it is harder to control the voltage at which the tool retracts. In some cases, lack of this control may result in poor quality cuts. But if one is careful it may not matter. If I work on this again, I would make a special add-on attachment for EDM that contained the device for retraction. This could be a motor+leadscrew or a voicecoil actuator or a solenoid... And I would use a dedicated Arduino to control the tool retraction, so that the firmware wouldn't be too loaded down with the high speed tool retracts...

I 'deconstructed' an ancient digital (photo)camera and found a tiny solenoid in it
You might be able to use an ordinary relay - carefully remove the cover, and attach the tool to the moving part of the relay. You wouldn't use the contacts of the relay, just the solenoid and the little hinged mechanism that moves:

January 17, 2014 09:08AM
Must have a relay around somewhere.. But it isnt where it should be.. How would you attach it, hot glue i suppose?

Havent done that much, but not nothing, have a little setup.. The orange kapton tape constrains the coil and i use rivets both as the moving end and as the core for the coil.

It buzzes, both at 5V or 12V but it soon finds a way to break the loop permanently. The loop is closed by that wire touching the bottom rivit, and it has other ways to move.

I'll convert that to a newer design with the same idea, but with more decent holder and slightly more decent sliding thingy. However still feel the thing wobbles a bit and it rotates freely, that might make attaching the in a way that doesnt move too much tricky..

The idea with the lever arm might have the same problem if it can move side to side too much. A printed one could do, thin bits of PLA do seem to be sufficiently flexible, dont know how that scales to well days@10Hz is millions of times bended back and forth.. perhaps go for a pivot. The sliding thing also posits the issue of finding a slider that everyone can get/find, whereas a pivot/flexible tongue only requires some bit of ferromagnetic material/differently permeable/magnet there.(big differences between materials, and there is difference in how much it retains magnetization)

Electrically it is connected simply from the nut at the bottom through the gap, through the coil and then through a LED&resistor which i short out to 'run' it, and shows if it connects through when not running.

Should try make a model, look up plasma physics on how the spark works again.(suppose it separates charge and then conducts with some voltage difference or something..) I think the behavior might change a little by inhibiting the coil a bit with some capacitance there, for instance.(But probably thats not useful)
January 18, 2014 12:05PM
A simpler way to build this so the motion is accurate is to use two identical flat spring steel strips parallel to each other, one at the top one at the bottom. Fasten the fixed ends with two screws to a plastic insulator block. Use a threaded rod through a hole in the other end of each strip with double nuts to fasten it in place. A loosely fitted coil goes around the rod at the top and sucks the rod in when current flows. The rod will stay perfectly vertical and for small vertical displacements it will move side to side very little. If you use 3/8-24 threaded rod a standard drill chuck can be used to hold tools. You also could put the coil just above the chuck and let it pull up on that.
January 23, 2014 12:52PM
I like it, Jasper! You'll be drilling holes in no time! smiling smiley

use two identical flat spring steel strips parallel to each other, one at the top one at the bottom

I don't quite follow all of this description without seeing a picture, but it is probably similar to this:

This would be a way to do it so you could use an existing solenoid. The whole flexure mechanism could be printed and it would probably work fine. You could also probably make the flexures out of a plasitc soda bottle or an aluminum drink can (don't cut yourself slicing these open!). For something light-duty like this, the flexures don't have to be spring steel. (For anyone interested there are a few experiments with flexure mechanisms documented on the wiki here and here.)
March 11, 2014 07:06PM
Sorry for being so exceedingly slow moving! Been very distracted by the cryptocurrency revolution.(ethereum in particular, but also stuff like namecoin)

Based on garyhlucas, thats what i made, but printed PLA as you said, guessed it would work myself though :p

The design could be better by having the strips come from both sides, or maybe even four sides. It would have less ability to move side to side, more layers might also be something to consider. (maybe just two directions is good too..)

The first issue to fix is that my electromagnet simply seems to simply too weak. It can pull the nut up, but only when you put it in a too small distance it too easily gets out of. I'll (finally)wind a new one tomorrow.

As with 99% of the cases, i put it up at github. If anyone wants to give it a shot, dont worry about catching up with this snail :p I dont mind if others end up getting further/moving faster than me.

To be clear, what i was thinking was, this could help me with this stupid idea. Probably wont work, but useful/cool if it does. And secondly, the idea is to EDM with a reprap. Basically slice with 100% infill, really low movement, and 'slice top down', figure out a way to stop the machine whenever the EDM keeps discharging.(probably simply have the actuation touch two metal bits which then short and have a pin measure that, probably with some side electronics) We're closer to replabs if repraps can easily be retrofitted to do something like EDM. If it can work in air, that would be great.. If in oil, it would be less convenient, but probably still workable.
Sorry, only registered users may post in this forum.

Click here to login