Spark Eroder pulse generator

[sihaam]

Hi all. I am building a small spark erosion machine using a 50V DC x 5 amp power supply as the output source for the spark eroder. However, I do not need this "straight" DC. I need a circuit which can vary this 50VDC into a variable pulsed square wave. All help here will be appreciated

 

[MikeMl]

HV FET in-series between the supply and the spark-gap to create the pulsed output?

I'm thinking there is more to this, e.g. you need a "ballast" (constant-current circuit like a welder, or carbon-arc projector) to get things started.

 

[sihaam]

Spark erosion is the process whereby you use set "sizes" of sparks to cut away metal. In spark erosion, you need to "chip" the workpiece with sparks of a uniform amperage, eg 3amps x 50VDC. Every time a spark is generated, it "eats away" a precise amount of the workpiece. The time space between the sparks (the on time must be able to be varied from 1 millisec to 500 millisecs). Similarly, the off time must also be able to be varied. So I assume that the circuit will have two knobs, one to control the ON time and the other to control the off time. I am no an expert in electronics, but everyone in spark erosion tells me that the wave must be square. I assume the square means that the entire period of the "on" pulse will be eg 3 amps x 50V and when off, it is abruptly zero. In this manner we will be generating a set size of the spark. I need a simple circuit to take the 50VDC and make it into controlled pulse lengths. No special power supply is needed. Standard regulated 5 amp x 50VDC power supplies are commonly used. It is definitely not like a welder. In a welder the arc is continuous. In a spark eroder, the arc is simply a controlled burst of a spark jumping from the cathode onto the work piece. See attached pic to get a better idea.
View attachment 68949

 

[jpanhalt]

EDM might be a more familiar name for this technique.

https://www.engineersedge.com/edm.shtml

It is still not clear what you want help on. Can you provide more specifications?

John

 

[shortbus=]

HV FET in-series between the supply and the spark-gap to create the pulsed output?

I'm thinking there is more to this, e.g. you need a "ballast" (constant-current circuit like a welder, or carbon-arc projector) to get things started.

Most of them use a capacitor bank to store a set amount of energy then discharge it with the mosfet/IGBT. The work is done in a dielectric liquid tank. The voltage breaks down the liquid then the current transfers to erode the metal.

John is correct, it's known as EDM.

 

[MikeMl]

Most of them use a capacitor bank to store a set amount of energy then discharge it with the mosfet/IGBT.

So you need two switches (FETs); one which turns on only while charging the capacitor, and another to discharge the cap into the arc (after the first is opened, to prevent shorting the supply)?

Or is the supply just current-limited, and then you have just a single FET to discharge the cap into the arc?

 

[dougy83]

You can use a 555 timer for controlling the pulse ON and OFF times, perhaps similar to the attached circuit.

 

[sihaam]

Yes, I am quite familiar with the technique, whether some call it EDM or spark erosion. I did not want to get very technical concerning the die-electric liquid etc. That is all beside the point. Also the power supply is perfectly adequate as our standard formula requires around 1 amp per 1mm thick material. Cutting 3mm will require 3 amps and so on. Reason why some machines seem to be so large- they cut deep, sometimes more than 300mm , hence 300 amp machines. Any circuit which will allow me to use this 50V DC in pulse forms will be acceptable. Using the 555 timer suggested bu Dougy83 will work too, but will this circuit be able to handle the 5 amp x 50VDC? I am not clear on the outputs of this circuit. Could someone add onto the circuit the necessary components to complete it so that I can end up with a cathode -50V DC and an andode which could then be directly used by me?

 

[dougy83]

No, the circuit I drew is just a controller; it would be used to control the power section.

You stated that the current should be constant; do you want this to be adjustable as well, e.g. using another 'knob'?

As your ON time can be quite long and potentially the load will only dropping 10-30V (correct me if I'm wrong), if a linear current source is used, the pass transistor will be dissipating a lot of heat and will need a large heatsink. The alternative is to use a SMPS (or at least a chopper), which will not suffer from the same issue, but will have a some ripple current. Which is acceptable to you? Either is quite simple.

Also, can you confirm that the circuit has to be powered from the 50V supply, or is there a lower supply available?

One last thing... does it matter to you whether the anode or cathode is switched?

 

[camerart]

Hi,

I think the mechanics must be accurate.

You will need a precise movement and I would suggest linear bearings, inside a boot perhaps.

The feed speed must be adjustable so that the spark stays constant, not too fast not too slow.

 

[sihaam]

Yes Dougy83, if the current can also be varied via an extra knob, it will be great as it will then allow me to use thicker/thinner material. The power supply has a very good cooling fan system and 500 millsecs will rarely be used. The optimum ON time is around 20 millisecs, so both alternatives are acceptable. Yes, the circuit should be powered by the same source. However, if an external p/supply is necessary, I could obtain that. The anode and cathode can be switched, but we use the positive connection on the workpiece always, otherwise we will be eating out more of the tool than the workpiece . Lastly, in reply to Camerart, I am well aware of the mechanics of the moving parts, we use stepper motors, controllers on a precision XY table via a PC

 

[dougy83]

The heatsink is for the pass transistor in the attached circuit - it may not get that hot for low duty cycles, but you can watch that (if it does, then put a bigger heatsink on it).

The values are ballpark. You'll probably find that a 10uF cap on the 555 timer is more suitable. Try with a 1k resistor in series with the opamp output too.

 

[sihaam]

Hello Dougy and thank you for your circuit. I will certainly be using it and post the results here. Just a bit on the circuit, as I said that I am no expert. Below the 50VDC input on the left are those two 15V diodes? And is it the +50V or the -?. And I suppose the points marked "12" will all be connected together? And finally, does the 10K pot control the current? Thanks once again

 

[dougy83]

The diodes marked '15V' are 15V zener diodes, and each will drop the 50V down to a level acceptable to the regulator, which further drops the voltage down to 12V for the control circuit.

I have marked the positive input rail with '50V' and the negative rail with a GND symbolView attachment 68974

Yes the 10K pot controls the current.

 

[sihaam]

Thank you again Dougy83 for the Circuit. Everything seems fine so far. However, is there a way to add a circuit which will light up a led when the anode and cathode makes a short? Also, if possible could there be an analog output when the anode and cathode comes close to each other while dipped in some dielectric solution? Perhaps a voltage from 0-10v? So the closer the get, the voltage can either rise or fall?

 

[dougy83]

The short circuit detect should be pretty simple - just detect if the cathode voltage is too high when the 555 output is low. A pulse extender (e.g. another 555) would be used to drive the LED to stop it pulsing (unless you wanted it to).

The proximity sensing could possibly be done using an RF oscillator on the anode and a receiver on the cathode, with chokes isolating the RF from the rest of the circuit. It would probably require a bit a tweaking to get it to work due to the number of interfering objects affecting the coupling.

 

[sihaam]

Good, but then even a better way will be to totally end the power to the cathode when a short is detected. The LED can then come on. Because a short is a fatal error, I would like the circuit to cut off power to the cathode (or anode, whiever is easier). When the short is "un-shorted", the cathode should regain power. The LED could simultaneously light up upon the short (pulsing LED is no problem). Could you perhaps give me a little sketch as to the required circuit? This is a fascinating project and I am certain that many readers here would love to try their hand at spark erosion.

 

[dougy83]

Do you have a component limit or anything? Depending on what features you want to put on it, you'll need to add more parts. Also, it's best to decide what features you want and requirements up front, or you'll end up with a circuit full of tacked-on sub-circuits.

As for tacking on... If you want to just tack an LED onto the previous circuit, then you can use the attached circuit. It just lights the LED with the cathode voltage is too high during a pulse. It doesn't switch off the current or anything.

 

[dougy83]

As mentioned above, if the current rise time requirements are not extreme, you can just use a SMPS or chopper to regulate the current. The great thing about this is it doesn't matter if there's a short-circuit - the SMPS will just continue pumping the 3A through the contacts, but as there's no voltage, there'll be no output power (and minimal power dissipation in the electronics). With the linear current source, if there's a short circuit, the electronics dissipate the full power (e.g. 3A * 50V = 150W) and get hot (or die).

 

[sihaam]

I have already designed the circuit and will stick with it instead of the SMPS. I dont mind tacking the led on, but where is the attached "tacked circuit" I dont see it. I have an idea, how about using an opto coupler instead of the LED. The ON will simply disengage the cathode (or anode whichever is easier). The prime purpose is to prevent the thin anode from destruction. A "short" lasting more than 50millisecs will roast the anode .