DIY mostly printed desktop Wire EDM machine

[G-EDM]

Hi, maybe some people are interested in DIY Sinker and Wire EDM. I'm working on the G-EDM which includes a mostly printed CNC router, a DIY pulsegenerator and a custom motion control firmware that is able to move lines back in the line history to retract on short circuits.

It runs on an ESP32 and samples a current feedback with I2S ADC at close to 1MSps. The feedback is optically isolated through a linear optocoupler. PWM to power the drive the gate driver circuit is also isolated. Works pretty well so far.

The wire module had a redesign I am currently printing and building.

A little preview for those interested can be found on Youtube:

[G-EDM]

Small update:




Images show 30mm mild steel and 2mm stainless steel.

[moffy]

A beautiful looking project, wire EDM is relatively simple in principle, a wire an electric arc and a cutting medium, but there is an awful lot of detail to be overcome to get it working successfully, congrats.

[G-EDM]

Thanks.

If it comes down to all the details I honestly would not know where to start. There where just so many of them. 
All in all software was the biggest pain to do.

[wobbly]

That is very impressive.  Looks like a bought one as Dave would say . Great work

[Infraviolet]

You mention sinker as well as wire, does that mean your EDM system can cut all the sorts of forms a 3 axis CNC mill can? Not just the types of "laser cutter"/"water jet" style fully-through-the-material that wire EDM can? I've seen videos of a wire EDM project with separate 2 axis positioners above and below the workpiece, so it can cut slanted edges as well as vertically through, but that's stilll a lot less versatile than sinker can be. Nice work G-EDM, look forward to seeing more of it.

[G-EDM]

The software has 3 modes.

2D wire
1D sinker
2D engraving (Z axis floating above the work and XY moving. That one is currently disabled. So much has changed and it needs a full review first)

In sinker mode it can move in one direction only. But it can use every axis in both directions. But only one per run.
Multidimensional sinker is not available but I could add it once there is some time again.

There is no probing routine currently in the process and thus the wear ratio remains unknown making it hard to do mil like operations. But this is something that will be added for sure in sinker mode. Cutting a little bit and then move somewhere to reprobe the surface and get the electrode wear to calculate the true cutting depth.

[moffy]

Do you use distilled water for the cutting medium? From what other EDM DIY pulser circuits I've seen, the peak voltage is around 90V-100V is that what you use?

[G-EDM]

Normal tap water. Voltage is 60v but I may test up to 90v in the future.

Edit: The pulseboard is designed for >200v (trace clearance and caps etc.). The limiting factor is the current sensor and the zener diodes.

The DPM8605 that is used for current control can deliver max 60v but there is a DPH series that can go up to 96v.

The INA sensor can take 110v max.

[lugaw]

Can we look at the pulse board schematic?

[G-EDM]

The pulseboard is a basic low side switch switching a DPM8605 programmable step down converter.

https://github.com/G-EDM/GEDM-EVO2-CNC/tree/main/files/schematics-and-gerbers

[G-EDM]

Fresh batch of PCBs arrived today.

[SteveThackery]

Normal tap water.

I didn't know you could use tap water. The one at college used paraffin.

How much current are we talking about for the erosion?

[G-EDM]

Tap water works fine for most cases. Aluminum can make some problems with tap water and rust will come quickly for steel.

I only have 0.2mm wire and you can't really push it much. It is very sensitive. The max current I actually got a stable process without any broken wires was avg 0.8A with peaks to 1.4-1.6A.

There is much more going on here then just the current. Arcing is one of the major problems with thin wires. An arc is a short circuit against a tiny spot on the wire. Doesn't take long until the heat makes the wire snap. This hot arcs turn the glow of the burn red. if you see a red glow there is a problem. Only the blue glow is wanted. The software has some logic to shut down arcing early.

Another thing is the wire touching the work. If there is a high current situation going on AND the wire touches the work the physical stress can damage the hot wire too or deform it.

And then there is the removal rate of the cutting debris. The higher the current the more particles will be in the cut. Without high pressure flushing this can turn into a problem.

Industrial machines seem to go up to 1.6A for .2mm wire. But this is a mostly 3D printed machine with hotend nozzles as wire guides and a low pressure pump for flushing. So working with .8A avg is actually pretty nice for a .2mm wire.

Made a crappy short clip in the dark to give an impression of how a .8A burn looks like:



PS: While building the tank I poured solvent everywhere and the bottom plate had some stains. Used fire polishing to get rid of them. Wasn't a good idea. The heat removed the stains but created internal stress and as soon as a tiny drop of solvent touches those spots it starts to crack everywhere. Couldn't stand those cracks anymore and painted the tank white.

[G-EDM]

Made some more parts. A little gift and something I don't know what it is but it is tiny.

[G-EDM]

New controlbox.

[G-EDM]

Short test with the DPH running at 84v: