McGyver, you are correct that many areas can be simplified if productivity is not the concern. For example, the aforementioned tension method of jockey pulley and counter weight is much simpler than an electronic brake motor, and very effective. Also in that category would be submerged cutting and high pressure flush.
Moving the table accurately is about 1/3 of the final result. You will also need to guide the wire, (more on that later), and control the overburn.
The links you gave are good. I have run Agie’s and they make an excellent machine. The Virtual Machine Shop looks like a potential black hole time drain; I had to tear myself away from a lot of interesting reading. I did find two points in the edm chapters I would take issue with. One is they say copper wire is common; in fact brass is almost universally predominant, although there are some other materials being used in special applications. The other is their recommendation to use kerosene as a dielectric fluid in a sinker edm. This is never done anymore, kerosene is too volatile. Most machines use synthetic oil with a much higher flash point. I sent Ron Smith an e-mail regarding this; it might be nice if they could adopt a “wika” concept for open editing.
Now back to your questions. The state of the art is .0001mm steps, (4 millionths), or 10 millionths inch resolution. The general concepts for linear guides, rotary encoders, stepper motors, etc, is the same as the more conventional machines people are building, just to a higher order of magnitude. Your part will always be some amount less accurate than you can position your table, so this is the first order limitation. You don’t need much speed or thrust, so you can use whatever techniques (gear reduction, etc) that help towards that end.
As to the forward motion, the spark generation happens thousands of times per second, depending on the settings for off time and on time. At the same time, the gap condition (open gap voltage) is monitored and adjusted thousands of times per second. This is an interesting debate ongoing between machines with ballscrews that sense the gap more often, but have to deal with lost mechanical motion and limitations to acceleration, versus machines with linear motors that sample the gap less often, but can respond and accelerate the table quicker. Also toss glass scales into the mix. Some machines use closed loop feedback; others rely on the rotary encoders.
For a Diy system, you could use a forward feed low enough that you would not have to sense or respond to anything. Early machines were like this, and the next step in evolution was to stop the machine in response to a short circuit.
I think the next installment would focus on the wire drive/guide system, and the impact of dielectric fluid (deionized water) delivery/control.
Roy Solomon  |