Going back to my comments about adding a resistor in parallel with the two diodes in series with the gap. A perfect silicon diode would have close to 0.6 volts across is even with only a few microamps of current. So with no parallel resistor you could have about 1.2 volts due to leakage current through the dielectric. With a parallel resistor of say 100 ohms the voltage would be almost zero. When the dielectric breaks down the current would be many tens or hundreds of amps so the voltage across the two diodes would rise to around 1.5 volts. It is this change that you are trying to detect to step your logic onto the next step.

The diode you are using to set the difference between the lower and upper thresholds of the window detector may not have any current trough it (And therefore no voltage across it) This is because the input leakage current to the op amp/ comparator is very small and can be positive or negative. I think you are hoping for about 0.6 volts across this diode.

Hi Les, The reason I wasn't going with a resistor with the two gap diodes was to keep the full voltage (open circuit voltage) from being sent across the gap. Since the cap bank voltage will be somewhat smaller than open circuit voltage. Doing this will make the finish of the burn more consistent, or at least that's my thinking. I am adding a resistor to the other diode on that comparator.

Most of the DIY machines are just relaxation oscillator based. So some of the voltage/current from each discharge into the gap is used to ionize the dielectric. By keeping the Q1 on for the time it takes to ionize the gap, it should then give the cap bank it's full energy store into the burn. Again that's my thinking behind this.

I guess I should call this circuit a digital relaxation oscillator?

My circuit is quite simple when you break it down to blocks.

I wasn't trying to disparage your circuit, just trying to be funny. I really do appreciate all of your help with this. I just have a weird sense of humor.

 

I zoomed up your original schematics and converted to .PDF for you. Not super but maybe will help.
Hope so.
And as long as I was posting bad images, here is one of my '65 Corsa Ragtop 140 from 1970. Fun little ride.

View attachment 144475

Hi John. That car is the type of Corvair I really liked. If you squint your eyes a little between the style of those and a AMC Pacer, it is where most car styling is at today. I know you need to really squint to see it but I do see the correlation.

Thanks for the PDF, but the schematics will be change in the near future. And when I get to a final design I will do a PDF of it. What I posted was just something to get some feedback on what I'm trying to accomplish.

 

Now a couple of questions before I redraw this again.

What should the value of the top resistors be in the voltage dividers be. Isn't it the top resistor that is the most critical for proper use in a comparator?

And does the way I have the 555 circuit drawn look like it should? I've seen it drawn two ways for an adjustable frequency. One way was with the pot between pin 8 and pin 7 like it is now. And the other was with the pot between pin7 and pin 2. So which is correct? The one I used was from a Forrest Mims book.

 

I'm not sure what you mean by the top resistor on a comparator. I have done a very rough drawing to try to explain the operation of the comparator on the B input. (Gap voltage)

The triangular wave represents voltage B. (Note it will not change in this way. It is just drawn that way to show the state of the two comparator outputs for different valued of voltage B.) The top line marked 3 volts is the voltage the potentiometer (R12) slider. The lower line is this voltage minus the diode forward volt drop (About 0.6 volts) For the diode to produce this drop you would need a high value resistor (A few hundred K ohms) connected between it's cathode and the -12 volt rail. The two lower waveforms show the output of the comparators. So there are three states. Increase gap. No change ad reduce gap. I realised there is a flaw in the logic. We are only interested in ths gap voltage during the spark time. So without a modification during the time the capacitors are being charged the gap voltage will be zero. We need to prevent the feed trying to increase the gap during the charging time. For the 555 the position of the fixed and variable resistors need swapping over. The resistor between pin 7 and the capacitor is the main frequency control. The value of the resitor between pin 7 and the positive rail mainly influences the duty cycle but it does have some effect on the frequency.

Les.

 

@strantor, forgot to list my plasma cutter in my list. Thought it would be more useful than it really is. Also have a sandblast cabinet. And forgot about the Miller portable spot welder. Only paid $10 for it because that's what the guy was asking. My Mig is a Miller 235 with the spot weld add on. Both the plasma and the Mig were store demonstrators that they sold me for like 1/3 the new price when Miller came out with newer models. Like I said earlier, I keep my eyes open.

A while back there was some guys trying to make a DIY water jet using a pressure washer. Don't remember where I saw it though. I watched it for a while but have too many ongoing project to start another one.

Feel free to use this EDM circuit if you want and comment on what is being said. I plan on putting it out there into the DIY EDM community when it's working. Not for sale, but to just help people that want one. One that is more than just a toy.

I made my sanblasting cabinet out of scrap wood. Pics attached. I've documented my 3,000A+ resistive spot welder on this site before.

I would love to comment on your circuit and I would especially love to help improve it. But the sad fact is that I don't understand it. You're over my head here. I've designed a few circuits over the years and I learned just enough each time to build the circuit that I needed. My cup of tea is industrial controls. My purpose on this site is only fulfilled once in a blue moon. I'm mostly here for off-topic conversation with folks who are smarter than I am. On that note, I'll stop detracting from your thread with posts about what tools I have or don't have, and I'll just watch quietly while bigger minds figure out an EDM solution that I can implement.

 

I would love to comment on your circuit and I would especially love to help improve it. But the sad fact is that I don't understand it. You're over my head here. I've designed a few circuits over the years and I learned just enough each time to build the circuit that I needed.

That's pretty much my situation too. I know from all of the different plans out there, from paying close attention when I was running these machine at work and from online circuits plus a lot of just plain online reading what is needed. And for got I have both of the books about EDM principles. Real hard numbers about how the process actually works are held very close to the vest, because the market and price of the machines. This is my first try at designing anything electronic. Things like LTspice aren't much help because there isn't enough information available to set up a circuit in it. My whole logic circuit is so unconventional most guys that design logic circuits just flat out say it won't work. But I can see know places where I have broken any of the "rules" setup for logic circuits, just not doing it "normally".

The whole circuit as designed could be run slowly just using two toggle switches(Q1 and Q4), and a couple of voltmeters(the comparators). At it's most basic level that's all that's being done by the circuit.

 

I'm not sure what you mean by the top resistor on a comparator.

Not top resistor on the comparator , but the top resistor in the voltage divider , potential divider. To give the correct amperage value needed for the comparator to work correctly.

For the 555 the position of the fixed and variable resistors need swapping over. The resistor between pin 7 and the capacitor is the main frequency control. The value of the resitor between pin 7 and the positive rail mainly influences the duty cycle but it does have some effect on the frequency.

Thanks for that. I've seen circuits set up both ways and didn't know which was correct. Frequency control is what I'm after in this. It is basically the speed control for the stepper motor driver, that will power the ram.

I'm going to print out the rest of what you said to try and understand what your saying about the rams movement. Doesn't make sense to me right off the bat, but will think on it more.

 

Hi SB,
I assume you are talking about the potential dividers R9 / R10 and R11 / R12. It is the ratio of the values of the resitors in the potential dividers that is important more than the actual values. The purpose of the potential dividers is to scale down the voltage you want to detect to within the range the comparators can deal with. I assume you will want to charge the capacitor bank to around 150 volts. The comparators are powered by a +12 volt power supply so you need the output of the potential divider to be less than 12 volts. In practice you would probably aim for about 10 volts as some comparators don't work well close to the supply voltage rails. (I have not looked at the data sheet on the LM339) So to divide the 150 volts so it was represented by 10 volts at the comparator inputs the top resistor would need to be 14 times the value of the bottom resistor. (Comparators have a very high input resistance so you can ignore the loading effect on the potential divider. I would probably choose between about 1K and 10K for the lower resistor. Using high values can give problems with electrical noise causing false triggering of the comparators. Using lower values means more power dissipated in the resistors requiring higher power rating resistors. The effect of electrical noise can be reduced by putting a low value capacitor in parallel with the lower resistor. The current (Amperage as you call it.) though the gap is only limited by the wiring resistance, the effective series resistance of the capacitors, the ON resistance of the MOSFET and the gap resistance. The voltage across the gap will start at almost the capacitor charge voltage but drop during the duration of the spark. We would need to need to sample it at some time during the spark or average it over the duration of the spark to make the descision to increase the gap, decrease the gap or not change the gap.
Where does input "C" to block 4 of the diagram come from ?

Les.

 

Hi, Les,

I assume you are talking about the potential dividers R9 / R10 and R11 / R12. It is the ratio of the values of the resitors in the potential dividers that is important more than the actual values.

Correct on the divider resistors. I understand the concept of the divider but lack in knowledge on how to choose the values needed. I thought it needed to limit current inot the comparator. This cleared it up for me, thanks.

I assume you will want to charge the capacitor bank to around 150 volts.

Actually it will be a little under 100V. There is a schematic floating about online that was part of a college thesis, it's called "the Garden of EDM". He claimed you needed ~300V open circuit on the discharge. But someone forgot to tell all of the manufacturers of EDM machines. All of the machines I've run, and that is most of the old, no longer in business companies to the big recent ones, use ~100V for the open circuit. That actual cutting voltage, when current is flowing, is around the 30 - 40 volt area. The reason for the 100V and not higher is it keeps the gap distance, in all directions around the electrode to right around 0.001 inches. When you need a sharp corner in what your burning you end up with for most purposes a 0.002" radius. Acceptable for almost any purpose.

So to divide the 150 volts so it was represented by 10 volts at the comparator inputs the top resistor would need to be 14 times the value of the bottom resistor.Comparators have a very high input resistance so you can ignore the loading effect on the potential divider.

.

Actually I did it differently for a different reason. I'm more concerned with keeping the ~5V diference in the window in what I called block#3. So I divided the 5V by the diode drop to get ~0.12V being equal to real voltage of the discharge circuit. And will use it to get my voltage divider values.

The current (Amperage as you call it.) though the gap is only limited by the wiring resistance, the effective series resistance of the capacitors, the ON resistance of the MOSFET and the gap resistance. The voltage across the gap will start at almost the capacitor charge voltage but drop during the duration of the spark. We would need to need to sample it at some time during the spark or average it over the duration of the spark to make the descision to increase the gap, decrease the gap or not change the gap.

This was also told to me in one of my other old threads, and I disagree. Here's why, maybe my increased knowledge will help me get it across. If the logic was being used as a normal 'clocked' circuit I would agree. But, my circuit is only using the logic to make Q1 and Q4 into SPST toggle switches. To control the charge and discharge of the cap bank. Those switches are in return toggled by voltages in the power circuit. I only used the "clock" input on the flipflops to get around the fact that the compararors put out a continuous signal, so if I used the flipflop "set - reset" function they would lock up the circuit. so no toggling like is needed.

The ram/electrode needs to be able to move during a discharge, sometimes more than once during one discharge. So a sample and hold type thing is a no go. All of the circuits out there are doing it like Iam with reguard to ram movement.

Where does input "C" to block 4 of the diagram come from ?

"C" comes from the two diode drop after the spark gap block #10.

I have an answer to what you were saying in post #44 but am out of time for right now, will answer later.

 

Hi SB,
I just use 150 volts as an example for scaling the voltage at the electrodes down into a voltage to feed to the comparator circuit. The reason for limiting the voltage to around 100 volts makes sense.

The reasoning for scaling the 0.6 volts to represent the 5 volt window at the gap is a reasonable approach. The only thing is the 0.12 is not volts it is a ratio. The ratio of the input to the potential divider to the output. (Volts divided by volts has no units.) so a 40 volt gap voltage would be 40 volts x 0.12 = 4.8 volts at the input of the comparator. So R12 would be set to give 4.8 volts on the slider.
I too am running out of time and the next point reqires some thinking to answer.
I may well just edit this post rather than starting another one when I return.

Les.

 

Hi again Les.

For the diode to produce this drop you would need a high value resistor (A few hundred K ohms) connected between it's cathode and the -12 volt rail.

Do you really mean -12V? I'm only using a single ended supply here, +12V. Yes i'm putting in a resistor on the redraw.

I realised there is a flaw in the logic. We are only interested in ths gap voltage during the spark time. So without a modification during the time the capacitors are being charged the gap voltage will be zero. We need to prevent the feed trying to increase the gap during the charging time.

The biggest thing in getting a consistent burn in EDM is flush. In fact most people will tell you that 3 things are needed, flush, flush and flush. With that said if the electrode WOULD back up a slight amount during charging that would only allow better flushing, burning creates a ton of swarf, both from the metal being removed and the electrode wearing.

When I get this into testing I find that it retracts too much, I have a fix. A 4066 controlled by the the same signal that turns the discharge mosfet off. the 4066 would then be connected between the resistor you mentioned above and the common/ground. This would then turn off the reference voltage to that window comparator. With no voltages at all on that comparator, the motor can't move, so neither can the ram. Until the discharge mosfet turns on again. I've been planning this for a while and think I've found all of the bugs in it? But maybe not. It's the small stuff like putting that resistor in that I was missing.

I'm hoping to get some time in the next day or so to do the redraw.

 

Continuation of reply to post #49
Reply to point 4.
No matter how you control the circuit it can't disobey ohms law.
From what you say about the servo having to move during the discharge time my understanding up to now has been wrong. The servo on my setup would not be capable of starting and stopping within the discharge time. (From what I can gather is in the order of 100 uS) I am not sure a stepping motor is capable of 10000steps per second.

Point 5
I must have missed point "C" when I was last looking at the schematic. I must have seen it the first time I was following the logic as I commented about needing a resistor in parallel with diodes D4 & D5.

Other comments.
You could make the circuit triggering the flipflops simpler by differentiating the output of the comparitors and directly driving the set and reset pins. This is how you have done it from the start switch. (There should also be a resitor to ground from the left hand end of C2. This is so that it is discharged next time you come to use the start button.) I have also noticed that in two places you have connected the outputs of two gates (4081s) together. This should not be done except with open collector or open drain type gates. You should have combined the outputs using an OR gate.

Comments on post #51
The resistor can be connected to ground as you are not using a -12 volt rail.
You could use a 4066 to inhipit the servo during the charge time or you could just hold the reset pin (Pin 4) of the 555 low during the charge time.
Hopefully I have remembered to cover all the points.

Les.

 

I am not sure a stepping motor is capable of 10000steps per second.

Yes it is, depending on step resolution (I've run step motors at up to 100,000 steps per second) ... a step motor max rpm, though, is about 3,200.

 

Hi Les.

No matter how you control the circuit it can't disobey ohms law.

I'm not to sure what you mean by this, where have I done this?

(From what I can gather is in the order of 100 uS) I am not sure a stepping motor is capable of 10000steps per second.

The movement of the ram is only going to be around 0.0005 inches per step. And with that being that the metal removal rate is so small the stepper will not be stepping all that fast, maybe at most one or two steps per discharge if even that. The stepper driver will be set for 1/4 steps.

This is one of the big problems in Ben's build. His of all the plans I've bought and I spent a lot of $ on plans, is the only one not using stepper motors. I think because he couldn't figure out how they work. He like me is a machinist, not an electrical guy. He and I assume by the schematic you showed are using a brushed gear motor, that has, if I remember right, a 60RPM output. This in turn is driving a 1/4-20 thread lead screw. So the ram is moving very far every time it moves. This is why if you go to his forum, you will find guys that designed "over riding clutches" to go between the lead screw and the motor. Because the electrode was jamming it's self into the work, and each time it does it shorts out the burn then needs to retract. Since the ram moves so fast it retracts pretty far. Then it needs to advance again, that starts the cycle of advance/retract over again. By moving the ram at a rate much smaller than the gap distance, this advance and retract cycle is eliminated and a smoother more efficient burn can be done. That advance/retract cycle in the industry is usually called "dither", and is to be avoided, sice it isn't productive to what needs to be done. There used to be some Youtube videos that some of the early builders of Ben's design had up, and from some one that's actually done EDM, they aren't very pretty.

You could make the circuit triggering the flipflops simpler by differentiating the output of the comparitors and directly driving the set and reset pins. This is how you have done it from the start switch. I have also noticed that in two places you have connected the outputs of two gates (4081s) together. This should not be done except with open collector or open drain type gates. You should have combined the outputs using an OR gate.

I got (in my first thread on this circuit) that opinion and a few saying no. I like I have said many times, am no where near being an expert in this. So what I did was found a application online that lets you play with logic gates in simulation. And built my circuit block with it. I don't know how good it replicates actual logic gates, it's called, Digital Works. That's how I came to what is shown in the circuit.

I have striven in the whole circuit to make it simple. With as few IC's as possible. The main reason is the total of an EDM has so much electrical noise in it. Some guy's that know things like micro computers and how to use them have tried to use them, and the 'noise' just is too much, too many false signals they couldn't work around. I'm not that smart. I want something simple and that will get the job done. So this is where I'm coming from with this.

Didn't realize that was a no no with the 4081's. I really thought that the outputs would not be affected since they are if i remember correct, totem pole type outputs. And since they were wired to a "real input" (on the 4013) that will sink the input.

(There should also be a resitor to ground from the left hand end of C2. This is so that it is discharged next time you come to use the start button.)

I think, (maybe wrong) that AK addressed this early in the thread. But instead of a resistor he said a reverse bias diode. And for that matter won't the diodes in the input of the 40106 drain that over time? For most of the burns it will be tens of minutes between uses of the start button. Also since the last part of that start circuit is a "falling edge detector", won't making it low keep the output always high? Not something wanted.

 

Hi SB,
Point 1
I thought it was this statement of mine that you did not agree with,"The current (Amperage as you call it.) though the gap is only limited by the wiring resistance, the effective series resistance of the capacitors, the ON resistance of the MOSFET and the gap resistance. The voltage across the gap will start at almost the capacitor charge voltage but drop during the duration of the spark."

Point 2
I agree that stepper motors will enable smaller movements to be achieved. I used a geard DC motor as it was what I had at the time.
When I used to follow Ben's EDM forum I seem to remember some people did attempt to use stepper motors.


Point 3.
You make a very good point over the potential problems with electrical noise. Any problems with the circuit can be solved when they occure. The main thing is you have presented a new approach and described well how you want it to behave. (Which is a nice change to some of the questions posted on this forum where it takes many posts to even find out what the question realy is.)
The totem pole output is why you can't connect the outputs in parallel. You will get a situation when one gate is trying to pull the output high and the other is trying to pull it low.

Point 4
When the start switch is pressed the capacitor is in a discharged state so as the left hand side of the capacitor goes from low to high the right hand side initialy goes high but it is rapidly charged via the resistor to ground on it's right hand side. This generated a short positive pulse which will set (Or reset) the flipflop. If the capacitors input remains high there is no further change on its output. If the input is pulled low it's output tries to go negative of ground but can only go to about - 0.6 volts because of the diode. (This does not create a postive pulse.) As the start switch will then be open. (I am assuming it is a push button.) there is nothing connected to the left hand side of the capacitor so it will remain in it's charged state. So if the button is pressed again nothing will happen. (Unless it had been discharged.) In practice there is a good chance that it will self dischage due to leakage but you could not rely on this.

Les.

 

Hi Les. Doing this online with a time delay is hard. Would be better if in person.

Your point 1. I agree and don't think I am.

Point 2. Some did and wern't too successful, due to trying to adapt to his circuit board and the way he did things. Kind of works with the motor driver chip he used in the original version but is hard to adapt it to a stepper. And when you get one of his books the circuit is just a suggestion, he then wants you to buy a board from him with a little booklet that is different than the book circuit.

Point 3. That's one of the things none of the books on logic say. There will be OR gates in the next version. Thank you so much.

Point 4. I show a toggle switch as the start switch. In reality it will be a push button and relay that uses on set of contacts to latch it on, like in a motor contactor circuit. This is needed to also shut down the burn when it gets to the correct depth, through a NC micro switch. The schematic as shown was meant to just show the bare necessities, the working parts of the logic and power output. To make sure what I'm trying to do looks workable.

The start circuit is really doing two things. 1. It resets the flipflops to a known start condition. 2. After that when the reset times out, on the falling edge of #1, it then makes a short + pulse to start the circuit. Or at least that's the intention.

 

Sometimes the time difference has helped on this topic as it gives me more time to think about the way to answer various points. I have found myself running two instances of the thread. One typing a response and one going over the schematic and your description of the sequence of events. I did plan to try to do a sort of timeline diagram of the states of the flipflops and mosfet switches showing what event stepped the logic onto the next state. I only got as far as creating a text file putting your description into my own words. It was in the process of doing this that I noticed the gates with the outputs in parallel. (I have attached this file. You may spot something that I have misinterpreted.) My feeling at the moment is that this design is a cross between an RC design and a pulse design as the pulse time is defined only by the time it takes for the capacitor to discharge. My impression of pulse designs was that the controlling the pulse width and spacing improved the quality of the erosion process. When I build experimental circuits I normally use Veroboard. (Stripboard) for the logic part. I also make it large enough to add more components as the design evolves. (One problem I create for myself is building it before drawing schematics and then having to go back to check the schematics I draw match the circuit.)

Les.

 

Hi Les. You don't know how much I appreciate you working with me through this. Trying to explain what an EDM is an does and still getting help has been hard. EDM flies in the face of most other electronic principles, when talking to most people versed in electronics. Thank you!

One thing that would help me though is if you would consider using the "quote" function when replying instead of the point numbers. It sometimes kind of confusing to me because I'm not sure what your meaning by point#. I hope you don't mind.

I only got as far as creating a text file putting your description into my own words

I will print it out and read it. I do better with reading things on paper for somee reason. Makes me think more about what's being said I guess. I will comment if your interpretation is different than mine.

My feeling at the moment is that this design is a cross between an RC design and a pulse design as the pulse time is defined only by the time it takes for the capacitor to discharge. My impression of pulse designs was that the controlling the pulse width and spacing improved the quality of the erosion process.

It is loosely based on two different machines. An Eltee Pulsatron and a basic Charmilles. While I never saw a schematic of either machine, I paid very close attention to how they were sparking compared to other machines I ran. With that I have tried to figure out a way to do it. And in my mind this should work. Once I prototype it I can go from there to refine things.

In an earlier post I called it a, "digital relaxation oscillator". That's as close to an explanation as I can come up with. I really can't take credit for all of the design am just implementing many things I seen in my research into one design. A RC machine and even most pulse circuits use some of the stored energy to ignite/ionize the gap. By leaving the cap charge voltage on until ionization takes place then shutting the charge voltage off, more of the cap energy is available to the burn.

What your saying about the pulse design is true. But that also has problems. When seeing a scope reading of Ben's pulse circuit it has many "missing pulses", missing because they didn't ionize the gap. when those pulses aren't used the burn suffers. With the pulse width and timing being controlled by a 'clock', it doesn't take into count the gap conditions. I've run old machines that had this problem, the so called "second generation" machines. What I'm doing is trying to implement a "third generation" type, like the Eltee.

When I build experimental circuits I normally use Veroboard. (Stripboard) for the logic part. I also make it large enough to add more components as the design evolves.

That is my intention too. When I drew up the set of drawings that have been used so far, I didn't know about the confusion having two drawings with the same numbers would cause, and will change that. Did it though because they are two completely separate boards.

 

Hi SB,

Hi Les. You don't know how much I appreciate you working with me through this. Trying to explain what an EDM is an does and still getting help has been hard. EDM flies in the face of most other electronic principles, when talking to most people versed in electronics. Thank you!

One thing that would help me though is if you would consider using the "quote" function when replying instead of the point numbers. It sometimes kind of confusing to me because I'm not sure what your meaning by point#. I hope you don't mind.

I actually found the ETO forum as a result of searching for information on pulse type EDM. Someone on there suggested that I have a look at the AAC forum. I still find EDM interesting but I think now it is unlikely I will build a pulse EDM. I now realise the mechanical parts need to be very rigid and there would be a lot of work building a suitable filtering system for the dielectric. So following your build is interesting.

I will print it out and read it. I do better with reading things on paper for somee reason. Makes me think more about what's being said I guess. I will comment if your interpretation is different than mine.

I also find it better to print things out to save having to jump between windows.

It is loosely based on two different machines. An Eltee Pulsatron and a basic Charmilles. While I never saw a schematic of either machine, I paid very close attention to how they were sparking compared to other machines I ran. With that I have tried to figure out a way to do it. And in my mind this should work. Once I prototype it I can go from there to refine things.

In an earlier post I called it a, "digital relaxation oscillator". That's as close to an explanation as I can come up with. I really can't take credit for all of the design am just implementing many things I seen in my research into one design. A RC machine and even most pulse circuits use some of the stored energy to ignite/ionize the gap. By leaving the cap charge voltage on until ionization takes place then shutting the charge voltage off, more of the cap energy is available to the burn.

What your saying about the pulse design is true. But that also has problems. When seeing a scope reading of Ben's pulse circuit it has many "missing pulses", missing because they didn't ionize the gap. when those pulses aren't used the burn suffers. With the pulse width and timing being controlled by a 'clock', it doesn't take into count the gap conditions. I've run old machines that had this problem, the so called "second generation" machines. What I'm doing is trying to implement a "third generation" type, like the Eltee.

I have never seen any comercial EDM machines and don't know the names any of the manufacturers that make them. when I was seaching for information on EDM I was hoping to find posts that included Ben Fleming's pulse EDM schematics to get a general idea how the gap current was controlled and the sort of pulse timing that was required.

That is my intention too. When I drew up the set of drawings that have been used so far, I didn't know about the confusion having two drawings with the same numbers would cause, and will change that. Did it though because they are two completely separate boards.

It makes sense to separate the logic parts from the high power parts to reduce the effects of electrical noise. It might even be worth considering using optical isolators betwee the logic and the MOSFET driver chips if you have problems that you think could be caused by electrical noise.
I could not find any information on using the "quote" function on this forum. I had to go to the ETO forum to find it. (I hope it works the same way.)

Les.

 

@LesJones , the quote thing, it just seems to work better when talking about something like this. but it doesn't work like ETO. You highlight what you want to quote and there's a 'drop down box' that then comes on the screen, you then can choose quote or reply. Quote lets you put more than one thing in a post. It also automatically give an alert to the other person.

Looked at your txt file and you got it right, except the point #4 which we already talked about.

The pulse EDM doesn't need to be any more rigid than a RC type since it is doing no contact machining. The filter system is simple too, many like Ben use a car oil filter. But I'm going to use a whole house water filter, they are made just like commercial EDM filters.