Wire EDM machine

[Neale]

This thread is under the "Other Machines..." heading as it doesn't have its own category - yet?

I know that some forum members will be aware of the Society of Model and Experimental Engineers (SMEE), a old-established society with mainly UK but also international membership. A small group of members of the SMEE Digital Group have been collaborating on a wire EDM(*) machine for the last couple of years. For those who haven't come across wire EDM, think of a hot-wire cutter as used for foam, but this time using a spark between the wire and the (metal) workpiece. Enough sparks for a long enough time, and the workpiece gets cut away, leaving a cut around 0.1mm wider than the wire used for "cutting" - and this machine uses 0.25mm wire, so that's a pretty fine cut. Move the wire under CNC control, and you can cut intricate and accurate shapes out of difficult-to-cut materials, including hardened metals that are next to impossible to cut otherwise. Wire EDM sits alongside laser and water-jet cutting as a go-to technology for tricky jobs. If the principle sounds easy (which it is), the practice is a bit more tricky. The wire has to keep moving as it gets eaten away as well, while maintaining tension; the spark voltage has to be accurately controlled by adjusting the position of the wire and thus the spark gap - too close and you need to back off, too far away and there's no spark; the whole thing is working in an incredibly noisy electrical environment as all the electronics is running a few centimetres from this spark transmitter.

If you do a bit of googling on "wire EDM", you will find plenty of big - and I mean pretty big - commercial machines and companies offering wire EDM services. Look hard and you will find a few, a very few, "home-built" wire EDM machines. We now have a working wire EDM machine that is totally self-contained and small enough to go into the back of a car, and as a small demonstration of its capabilities, here are a couple of sample pieces I cut on it earlier this evening:

The letter E is about 8mm high, the star about 4-5mm across. In fact, I had difficulty in finding the star in the tank (the cutting operation happens submerged in distilled water with a jet directed at the cutting area to flush away cutting debris).

I know that there are people who have made claims to fabulous machines on this forum but have offered little proof of their existence. Well, you will be able to see this machine in the flesh at the Bristol Model Engineering Exhibition next week, where with a bit of luck and a following wind, I shall have it running on the SMEE stand for all three days of the show. Forum members are cordially invited to come along, and if you can't find me, just ask anyone on the stand. If you are also a wire EDM user, then you are even more welcome to come along and compare our efforts with your commercial machine!


(*) Electrical Discharge Machining

[m_c]

WEDM is on my desired toy list.

Are you aware of the plasmaboog yahoo group, or http://www.mechatronicprojects.com/blog/ ?

[Neale]

Thanks - I had found EDM Yahoo group, but mostly seemed to be concerned with building from a kit of parts available in US. Hadn't seen that blog before - looks as if it started relatively recently.

I'm currently custodian of our team's machine as I'm taking it to Bristol next week so I shall try to get some pictures, and if the technology isn't too much for me, a clip or two of it cutting.

[m_c]

The mechatronics site is fairly new, but I think the author has been working on the machine for a couple years.

Benny Croonen (IIRC) who started/runs the plasmaboog group has been working on one for a good few years, with the group being a spillover from the EDM group run by Ben Fleming (I've got his original EDM book/plans).

I've liked reading the periodic posts regarding the problems faced. They're quite a complex machine that needs everything to work nicely together to get a good end result.

Is there any more info about your machine published anywhere?

[Neale]

No, nothing published as yet. You are also right about the "complex machine that needs everything to work nicely together to get a good end result" - or even to work at all. The wire feed and tensioner (DC motor with PWM control for wire feed plus stepper for tensioning) is one sub-assembly and includes a solenoid-operated guillotine mechanism for chopping used wire into short lengths as it comes out the end. Then there's the water flow and filtration, the XY movement with ballscrews and profile rails (tank moves on one axis and wire guide assembly on the other) plus stepper motors, and then the electronics. This machine has three PIC embedded microcontrollers in it - one for the keypad and screen, one for wire control, and one for spark generation which also controls the XY motion as these are inter-related. If anything is below par, then nothing works. In testing, a few weeks were wasted chasing electronic faults when it turned out that the deionised water wasn't quite pure enough. It's that bad. Electrical noise has been a perennial problem which is why the main control board is now up to version 9b, there are three separate power supplies, and it's wall-to-wall opto-isolators and so on.

Cutting area is something like 100mm square (although I need to check this). It uses a variable spark voltage of something like 35-60V although it generally seems happy at about 50V. Sparks are generated from a bank of capacitors switchable from 10uF to 100uF or so (again, need to check but it's something like that). It uses MOSFETs for switching - turn on one to charge capacitor bank, when charged that MOSFET turns off and another turns on to cause the spark. When the capacitor voltage has dropped below some threshold, the spark is turned off and the cycle restarts. The motion control is linked to this; if no spark occurs the wire is advanced and if the voltage drops too low the wire is stopped or even backed up a little. It's a very simple idea but has taken a lot of work to make it function. Wire speed is separately controllable. All the motion control software was custom-written; no use of modified Mach3, LCNC, Kmotion, grbl, or anything else. This does mean that the software has complete control of wire motion so it is possible to backtrack the entire cutting path if needed, for example if a cut ends in the middle of a workpiece rather than cutting its way out again.

We do not use any fancy commercial EDM components apart from the reel of EDM wire. So, no commercial wire guides, wire contacts, or things like that. Keeps the price down but whether this is going to hang together in the long term remains to be seen. It certainly works at the moment. There is already a long list of "wouldn't it be great if we could..." ideas!

[Neale]

Quick video showing the main features of the wire EDM machine. Video clip of operation currently in the editing suite!

[magicniner]

If it can't take G-Code from a CAM package it will remain happily niche.
I suppose it could be seen as an advantage that it will have no mass appeal to the new "maker" generation :D

[Neale]

If it can't take G-Code from a CAM package it will remain happily niche.
I suppose it could be seen as an advantage that it will have no mass appeal to the new "maker" generation :D

I may have misunderstood, but I'm not quite sure what you mean by this.

Wire EDM is certainly not niche as a technology. Even though to date it has been relatively little used commercially because of cost and complexity compared to more traditional methods of machining, this is starting to change. For £12K or so delivered, you could order a wire EDM machine from suppliers on AliExpress. I was talking to someone recently from a university department where they have recently installed a wire EDM machine and it is starting to take over from more conventional milling machines as the "go-to" choice.

The machine I'm talking about here is certainly niche. The fact that you probably won't find descriptions of more than, say, a half-dozen or so home-built/DIY machines like it world-wide rather proves that fact. However, it is precisely because it is a bit unusual that I am describing it here. I thought that readers of this forum might have at least a passing interest in something just a little bit different. This is not a build log, and I'm not expecting anyone to copy it. It's not a machine intended for the commercial market. The work needed to commercialise something like this is beyond our little team, and really doesn't interest us anyway.

It is also niche in the sense that it was not intended for "production" use, although people seeing it are not slow to come up with suggestions for how it could do some useful little jobs, certainly on the "model engineering" scale. It was intended to be relatively portable and, at a pinch, can be loaded and unloaded into the back of a car by one man so that it can be taken to exhibitions and so on for demonstration purposes. It spent three days at the recent Bristol exhibition and the plan is to take it to the Midlands Model Engineering exhibition in October. Of course, these exhibitions themselves are rather niche (although a surprising number of the general public do attend) so that just underlines the niche-ness, perhaps.

This machine fits into the CAD-to-cutting cycle in the same way as the ordinary CNC router or mill. We have generated our cutting files from Vectric software, producing gcode using a standard CAM post-processor. We take step-by-step instructions to drive the motion of the machine. We differ only slightly from the usual Mach3/LinuxCNC type gcode processing and step generation in that we have chosen for practical reasons to use what in computing terms would be a compiler, converting gcode into steps as a separate distinct phase, rather than the more common approach of using an interpreter running in more-or-less real-time to do exactly the same thing.

It is very easy to look at something like this and say, "well, all you do is connect a pulse generator to a bit of moving wire and shove a workpiece around with a bit of gcode and some steppers. How difficult can that be?" Well, we do just connect and shove. It's just that it's not quite as easy as it seems.

The controls on this machine - the first successful version of a number of prototypes - are simple and limited in scope and function. Too limited at the moment. However, they are the minimum needed to get the thing to cut. With a bit more user interface development and what my computing industry colleagues always referred to as SMOP - a simple matter of programming - this can and probably will be enhanced. The really difficult bits have been done.

Interestingly, there were two broad classes of visitor/spectator that I spoke to at Bristol. There were members of the public as well as model engineers with no knowledge of EDM at all, and I think that most of them went away with at least the idea of how it worked without ever wanting to replicate the technology or even have it in their workshop. And there were the guys who really know EDM in all its forms (which is not surprising being a stone's throw from Rolls-Royce who use this stuff extensively). They were, I think, both interested and impressed - a typical comment was, "I wouldn't have believed you could build a working wire EDM machine that small!" and they were generally kind enough to pass on some of their own knowledge. Well, to be honest, there was a third class as well. That was the guy who took one look (after his small son pointed it out) who said, "I've just bought a real one of those for my workshop" and walked off. Pillock.

Well, after all that, I'm prepared to make a bet with any forum member here. What's the chances that Boyan will have one of these things up and running by Christmas?

Only joking, Boyan - I think that you have rather more real work to do than play with toys like this!

But if anyone is interested to know a bit more, just ask.

[Neale]

If anyone wants to see metal being cut with a bit of sparking from a bit of wire, there's a clip here. That shows the machine that I've been talking about, although there are plenty of Youtube videos of "real" machines in operation. I have to say that some of those are pretty damn impressive. All the same, to have something using the same underlying technology actually cutting in your own workshop is quite something. YMMV.

[Clive S]

Neale What a fantastic project. About 12 - 15 years ago I happened to be in a small engineering place and I saw an EDM machine cutting a gear cog about 250mm thick and about 400mm dia. out of bronze or similar, I was mesmerised by it I think it took about 28 hours to cut it. There were three EDM machines in there.