That's how it is supposed to work. We get to learn about EDM and you get to learn about electronics.
Two diodes makes the circuit easier in a way. Now a single diode can be used for the voltage reference for the comparator - about .7 volts and 2 diodes for the current sense - about 1.4 volts.

Not knowing anything about a spice program, (LTSpice?) I'm assuming that "U1" would be the spark gap?

And D2, won't it allow a path from the gap sense to the the 339 power input?

 

@Max, thank you for the Elox material will read them this evening.

In that case, you can possibly use an inductive pick up similar to the old timing lights used to set engine timing with. It could be as simple as a wire wrapped around a torroid place over the ground wire.
I might have been a little too quick to dismiss this circuit. I would need to see the actual paper and the author's reasoning. Still, I think there would be a better way to accomplish what you're trying to do.

Wouldn't that just give a pulse for each time the gap arcs over?

I'll print out and copy the page from the PDF, and post it. Haven't figured out how to do it direct from a PDF. I'll also show my latest rough layout and what I'm trying to do with it step by step.

 

So the electrode moves harder into a workpiece if voltage is too high (no load), and backs away from the work piece if voltage is too low (shorted), and approaches again at a slower speed? Would that be the correct logic?

Pretty much it.
Max.

 

[

@Max, thank you for the Elox material will read them this evening.

.

I may put the whole manual into a PDF, as Ronv said it is all 80's logic components that need a touch of 21st century..
Max.

 

[

I may put the whole manual into a PDF, as Ronv said it is all 80's logic components that need a touch of 21st century..
Max.

Not to hijack the thread, but I just ran into a 1980s logic board. A maintenance guy was walking around with a big (8 x 14") pcb. I stopped him to see it was actually two PCBs stacked back-to-back with some type of connector. One side was filled with 74 series logic chips and the other side had a z-80 chip with a bunch of connectors on it. I asked where this dinosaur came from and what he planned to do with it. It was the board for the Lennox heater in a multi zone office area built in the 1980s. It got wet from the leaking roof, He was going to dry it off and put it back in the box on the wall. I still have heat, it must have worked.

 

Not to hijack the thread, but I just ran into a 1980s logic board. A maintenance guy was walking around with a big (8 x 14") pcb. I stopped him to see it was actually two PCBs stacked back-to-back with some type of connector. One side was filled with 74 series logic chips and the other side had a z-80 chip with a bunch of connectors on it. I asked where this dinosaur came from and what he planned to do with it. It was the board for the Lennox heater in a multi zone office area built in the 1980s. It got wet from the leaking roof, He was going to dry it off and put it back in the box on the wall. I still have heat, it must have worked.

Farewell, Zilog, fare thee well... :,(

With nostalgic regards
HP

 

Not knowing anything about a spice program, (LTSpice?) I'm assuming that "U1" would be the spark gap?

And D2, won't it allow a path from the gap sense to the the 339 power input?

Yes, U1 is the spark gap. I used a neon bulb to simulate it.

The 2 diodes protect the input of the comparator from voltage spikes. They won't be able to go more than 0.6 volts above +12 or - 0.6 volts below ground. With so much voltage and current flying around there is bound to be a lot of noise.

 

A most interesting thread. I have never played with EDM.

I believe I would try a rf detector.

For a commercial solution, one might invest some time and money and see if the spark can be analyzed.

With our sampling speeds and software.....who knows....you might get enough info to modulate the gap and the flush, keeping a continuous spark until the whole workpiece is done.

 

OK, hope site lets me download. Here is the page I got the two diode drop idea from. The full PDF of that, it's only four pages and tells about the different EDM circuits leading up to this. And a sketch showing what I'm wanting/trying to do. With an explanation and wave form of the gap. I really appreciate the help everyone is giving!!

 

Yes, U1 is the spark gap. I used a neon bulb to simulate it.

The 2 diodes protect the input of the comparator from voltage spikes. They won't be able to go more than 0.6 volts above +12 or - 0.6 volts below ground. With so much voltage and current flying around there is bound to be a lot of noise.

Thanks again Ron!! All this help makes me think I might even get this working.

 

A most interesting thread. I have never played with EDM.

I believe I would try a rf detector.

For a commercial solution, one might invest some time and money and see if the spark can be analyzed.

With our sampling speeds and software.....who knows....you might get enough info to modulate the gap and the flush, keeping a continuous spark until the whole workpiece is done.

The process of EDM does put out RF, think Marconi transmitter. You can't listen to AM radio near it.

The commercial machines have had much data harvested. When the company I worked for bought new machines the manufacturers put data recorders on them for the first month or so. But this is just a hobby/DIY project. The real world machines and industry is so competitive there is very little information out there. Even the stuff like Max shared is hard to come by.

The latest machines are CNC, the operator set a start-up parameter and the machine takes over and changes things for the best burn. This process has come a long way from the late 1940's. Two Russians using light bulbs and capacitors to make an RC relaxation oscillator circuit to cut metal. What I'm trying for would be called a simple "third generation" type machine.

 

Not to hijack the thread, but I just ran into a 1980s logic board. A maintenance guy was walking around with a big (8 x 14") pcb. I stopped him to see it was actually two PCBs stacked back-to-back with some type of connector. One side was filled with 74 series logic chips and the other side had a z-80 chip with a bunch of connectors on it. I asked where this dinosaur came from and what he planned to do with it. It was the board for the Lennox heater in a multi zone office area built in the 1980s. It got wet from the leaking roof, He was going to dry it off and put it back in the box on the wall. I still have heat, it must have worked.

I can remember them scrapping boards that size done with wire wrap connections at work. Wasn't interested in trying electronics then, so didn't grab any of the boards. Wish now I had, just for the sockets and boards.

 

What limits the current? Is it the oil or does the circuit do it?
In other words, how do you know 15 amps will be enough to still get 25 or 30 volts across the gap?

I wasn't thinking when I answered this before. The current limit is also controlled by the energy stored in the discharge caps. They do this so each discharge is as close in energy to the next one, to give a better finish in the burn/cut.

 

What can you tell us about the transformer for the power supply.

 

The one for the 95VDC is from an old servo motor drive. It's listed as 70V/16A, but actual output measures 68.5V on my DMM. The one for the 113VDC is a toroidal with a 80V AC output of unknown origin. I'm using full wave bridge rectification so multiplied by 1.41 to get my voltages. But didn't subtract the voltage drops of the bridge diodes.

The 68.5V one weighs ~15 or more pounds and the other one is ~6 pounds. Both with 120/240 primaries.

 

Wouldn't that just give a pulse for each time the gap arcs over?

Thought that's what you wanted:

I want to use the voltage on the ground side of the gap to start a timer, to control the length of the sparking time. If I measure the circuit voltage at the other side of the electrode, That won't tell me when the sparking starts. Just the voltage level.

Looking at your schematic, the diodes should work for you. I didn't understand what you were trying to do when I said they wouldn't work. I'm doing a poor job of understanding your objectives, and others seem to be doing much better. I'll just read now.

 

Shortbus,
Do you have a scope? I think you will need one.

 

I actually have two, but don't know much about using them. Went to an estate sale a few years ago and bought scopes, capacitance and resistance boxes(don't know what they are really called, they allow different values of caps and resistors to be selected and tried in a circuit), a couple of frequency generator/readers analog bench VOM's and other assorted test instruments all for $200.00! They set the price and I accepted! The old guy that owned them had died and his family was trying to clean out the house.

Again thank you for your interest and help!

 

Thought that's what you wanted:
Looking at your schematic, the diodes should work for you. I didn't understand what you were trying to do when I said they wouldn't work. I'm doing a poor job of understanding your objectives, and others seem to be doing much better. I'll just read now.

If I was using a micro controller (which I'm not) a pulse would probably do. But a steady voltage to a comparator will be used to keep SW3 in my schematic turned off during the on time sequence of events, to keep the spark ignition voltage from reverse biasing the capacitor discharge into the gap. Or at least that's what my thinking is.

Please feel free to keep reading and asking questions about this. I'm just winging it here, and learning as I go! Like many newbies I have ideas and concepts but not enough knowledge to implement. I bought my first computer to learn how to do this project, and even the small amount of computer and electronics skills have been a struggle. But with the help of people on this site mostly, I think I've come a long way from where I started.

 

Incidentally what are you using for your servo system and electrode material, copper or graphite?
Max.