Видео

Thank you for your feedback, The power generator used on the EDM was a transistor type based on an audio amp design but modified for a square wave pulse input which was provided by a “bought off the shelf” square wave frequency generator so no driver ic is required. The pulse frequency generator and the transistor type power generator are both shown at the start of the video. The voltage divider is fitted after the output current limiting resistors so that once the spark erosion process starts, due to Ohm’s law, the voltage at the electrode and in the divider will drop and this voltage drop is used to automatically control the D.C. motor controller. The motor controllers direction control works by using an input voltage which is either higher or lower than a fixed internal reference voltage, so the input voltage from the voltage divider can control the burn position of the electrode automatically. No other parts are used or are necessary although if you add a potentiometer to the voltage divider, you can manually adjust the output voltage, and set the electrode plasma burn voltage level which is handy for different spark conditions. The output current is set by switching in power resistors in parallel. My videos are not fake or click bait and I never personally ask anyone to like, subscribe or contribute financially in any way so I have nothing to gain from any of my content. I've always said that building an EDM is easy to do (there's nothing "magical" about them) and this type of “Easy-Peasy” EDM machine works simply by Ohms law and is very easy and cheap to build; also I’m surprised that no one else, now or in the past, has thought of using this way to control an EDM before?

@@edmguru9056 This sounds great...!!! I´ve read your answers to several comments, and I couldn´t figure out how it all worked, but I was really interested, And I finally can follow you... And the motor control sounds really unique !!! I´ll write again, when I understand it better...

Thank you for your feedback and subscription. You have some useful and interesting ideas there for an EDM but the point of this video was to test a new design of power generator under extreme conditions and to push it to the max to see what would fail first as other designs of power generator you’ll see online tend to have performance issues. This DC motor version of my EDM was actually scrapped in favour of my new stepper motor version, which has a superior control design, and it was just hastily rebuilt and thrown together to test the new generator, and with no heat sink fitted to the output power resistors, only two fans and I was surprised that no power resistors failed. On the stepper motor EDM, I don’t have any problems here as I use homemade, and fan cooled, power output resistors made from nichrome wire which can handle a lot more current. Also, as a point of interest, I’m sure you’re aware that with the EDM process not all output pulses are of sufficient quality to ionise the dielectric fluid, usually through bad flushing etc, so are missing at the electrode so to speak. With this in mind I have designed a new isopulse power generator which basically looks at the rising edge voltage and the rising edge current values of the output pulse, and if there’s a delay, a second, but at a higher voltage (in order to restart the ionisation) pulse, is injected at the output, to fill in the missing part of the original output pulse, this should increase the stock removal rate and decrease the total erosion time taken. When I get the time I’ll make and test it, using the old DC motor EDM again, mainly because I can’t be bothered to strip down my main EDM power generator as its wall mounted and the power transformer weighs 17Kg on its own.

Thank you for your feedback. There was no problem with flex during the erosion process and I agree the video is too long but if I didn't show it all people would say that it's fake so I left it long to prove the point, also the head (Z axis) can be tilted over as it's mounted to a small milling machine which has the provision to do so on its Z axis.

Thank you for your feedback. Yes the anti arc circuit works well with no problems so far. I've done a separate video about it on here that's worth checking out, if you're interested, about how it works.

@@edmguru9056 working on it Im at the second last one you posted now ... just about to watch it as I noticed a comment reply LOL I do hope to see this with at least 3 axis ... after that perhaps a rough costing of what a kit version would be worth ... leave the tub out since a sink is easy to find for anyone ... also perhaps add a filter to the bath and before it hits the filter add an electromagnet belt that pulls ferrous bits away from the filter and keeps it cleaner longer

Thank you for your feedback. I agree, inline dielectric filtration is important for an EDM.@@kaboom-zf2bl

Thank you for your feedback. The EDM power limiting resistors are on a separate fan cooled board in the video and the two resistors on the new power generator board are required for the “balance” of the charge pump circuit. Only the centrally mounted IRFP250N mosfet on the power generator board provides the output voltage (100V max, depending on the pulse duty cycle percentage) via three independently switched relays to the power limiting resistors, which are then connected to the EDM electrode, and the second IRFP250N mosfet is just used to provide the pulse to the gate of the first main power output mosfet. This second mosfet could be of any other suitable type even a standard transistor, it only has to be able to handle the voltage pulse (approx 120V) to the gate of the main power mosfet, which is high side switched. I used it in this new design simply because I had one spare and for no other reason, but I know this has caused a bit of confusion over how this power generator works because it looks like some kind of push pull dual mosfet design but it’s really a simple, and somewhat unique, new design of a basic bootstrap circuit. The output power limiting resistors values are calculated at the EDM working voltage which is approx 50V so for a 3 Amp output I would use a power resistor value of 16 Ohms at 200W (two 32 Ohm 100W power resistors in parallel) and all power resistors used on an EDM should be mounted on a heat sink and fan cooled.

Thank you for your reply, this comment space might not be the right place for asking so many questions. You might have seen the video of the guy with Ben's system. He got it working with huge heat sinks for his resistors. My question is what happens if his electrode size would double, will the amp's also double, with his limited power? @@edmguru9056

@@marcelgombault3657 Thank you for your feedback. The EDM output amps are fixed by the power mosfets and the power limiting resistors, so the electrode size or volume would only changes the amount of time that it takes to remove the material, the larger the electrode, the longer it takes to erode a cavity at a fixed output amps. My other EDM that uses a stepper motor on the Z axis, has an output of 30 amps, which is done by using two power generators in parallel, one at 10 amps output and the other at 20 amps output. This is how a commercial EDM works by simply using multiple power generators or power mosfets in parallel and this way you are limited only by the power rating of your power transformer, which in my case is 2KVA. It’s quite large and heavy weighing approx 17KG on its own, so by using the right power transformer in the first place, you can do some serious spark erosion work in a shorter time.

Thank to your feedback, i would like to build a EDM machine using three 30 A mosfets limiting each to 10 A and three of the power generators with the following specs, tell me what you think of them. An AC motor controller is known as a device that controls the speed of the AC motor. An AC controller can also be referred to as a variable frequency drive, adjustable speed drive, frequency converter, and more. The AC motor receives power, which is converted by the AC motor controller into an adjustable frequency. Want to run a DC high voltage motor from an AC 220Volts? This AC to DC controller can be used it adjusts the AC volts from 0-220Vac and converts it to DC : 50-210V linear underload Maximum current: 20A Continuous current: 10A Speed range: 0% - 100% Overload protection: Fuse Speed regulation: potentiometer (270 degrees with switch) Connection mode: fence-type terminal Housing size: 85mm*58mm*38mm Not sure if we should not continue this by mail not the place to send pictures or schematics.@@edmguru9056

Thank you for your feedback. It’s certainly an interesting idea you have there of controlling an EDM Z axis by using a VFD. It’s not something I’ve done and I don’t have much experience with this type of motor drive, but it maybe possible so long as the VFD can change the motors direction rapidly? Commercial EDM manufacturers tend to use servo motors on their machines, which are similar, but use an encoder I believe to control the rpm or torque of the motor. I don’t know why you would want to use a high voltage motor on the Z axis unless it’s the high output torque that’s required for your needs, but to me it would be easier to use a large stepper motor to do the job as a stepper motor controller is designed to stop and rapidly reverse the motor direction using the EN and DIR inputs, also the motor will run smooth with torque at a low speed, due to micro stepping, which a VFD might not be capable of?@@marcelgombault3657

EDM before computers .... was CNC milling ... turbine blades cooling holes came from NASA and the aerospace industry who had computers ... this is where EDM started ... his second video has a more detailed explanation and a non computer manual controlled system ... the first video is also manual just installed in an old pc case for ease of setup as for a default design ... it is literally OHM's Law ... being applied to do milling work by electronic erosion of the anode(the part you want) by the cathode (the part tool you are using to cut with)

Thank you for your feedback. Normally you would try to aim for approx 25 Amps power output for 1 inch of surface area if using a copper electrode at positive polarity. Positive polarity at the electrode will have lower electrode wear than if set at negative. (Negative polarity removes material much faster than positive but it has a high electrode wear rate and is usually only used on wire EDM machines or for eroding tungsten carbide but that needs a tungsten copper alloy electrode to keep the wear rate down). This is not a fixed rule and can be exceeded slightly and more so if it’s a roughing electrode as wear doesn’t matter so much. For a small EDM machine of say 10 amps output then keep the pulse frequency lower than 10 KHz for low wear. 10 KHz will remove material faster than say 5 KHz but 5 KHz will have a lower electrode wear rate. It’s difficult to explain but basically the higher the pulse frequency the more the electrode wear and keeping the output Amps low helps to decrease this wear rate so final finishing is done at low amps, high frequency and you only remove the last 0.02mm to keep electrode wear low. The most important thing is the quality of your output square wave, it must be as near perfect a square wave as possible or you’ll never achieve low wear rates. Use an oscilloscope to check your output wave form and if necessary redesign your power generator if it’s at fault.

@@edmguru9056 excellent description ... the missing bit ... it is a compromise between speed of cut and smoothness or precision of cut ... low n slow for very precise and smooth ... or high and fast for fast removal ... I would be interested to see this setup working on aircraft grade titanium .... as that is where I plan to use it ... need small turbines ... just havent seen a decent setup with low electrode problems until now ...

Thank you for your feedback. Dither is one of the setup functions that you’ll find on all commercial EDM machines. It’s just like the spark setting but the output amps are very low so no actual eroding takes place and this feature is used in the initial electrode setup to find the lowest point or to set a distance from a datum, to set the depth of burn, or even if eroding a hole or keyway in a shaft to centralise the electrode in say a v block so that the hole or keyway is set through the centre of the shafts diameter etc. The up/ down setting has no servo feedback so it’s used for clocking the work piece etc and care has to be taken not to ram the electrode or dial gauge into the work and cause damage.

Thank you for your feedback. I’ve worked on EDM machines of all makes since the 70’s so I understand what you’re saying and my machines are based on simple old school technology. Most of the time the output amps was low at only 3A so it’s not really going to show any smoke on the video and the only reason that I quote duty cycle times in my videos is because the square wave frequency generators that I use are bought online and only have frequency pulse and duty cycle settings although yes I agree individual pulse on and off time duration would be better and this could be done by making a custom pulse frequency circuit but for home use this machine is more than capable.

Thank you for your feedback. I don’t do social media as such but I have been on other website forums such as “All About Circuits” which have friendly, informative and technically minded members who can help to some extent with EDM. You can either start a new thread on there or if that’s not practical then you can send me a personal message.

Thank you for your feedback. Yes it’s not bad at all and since then I’ve rebalanced the bootstrap circuit and the 150 KHz waveform now has a better, more squared shape.

Do you use any snubbers? @@edmguru9056

would 400 to 600 bucks be considered worth it if the control box came preassembled ... and it had 3 axis functionality ... even if only static x y orientation ... of curse self sourcing the metal sink for the tub would not be a part of the price since sinks can be bought cheap everywhere

@@kaboom-zf2bl I really would love to see the total cost going a little lot lower, as ireally have NOT much monry to spend for anything other than food and cheap electronics to make me feel good. What is the benefit of preassembled hardware, does it get cheaper that way? Just create 3 packets in different price regions and with a different Level of "flawless machines out of the Box". I think after that we have a better basis to make a decision

Thank you for your feedback. The Easy-Peasy machine build is very easy to do and works well. The new design of a mosfet power generator is also an easy build which only leaves the current limiting power resistor layout, which is basically just resistors switched on in parallel, and the Z axis control which is taken care of automatically by the bought online DC motor speed controllers circuitry, so I might draw some sort of schematic up at some stage. The stepper motor type of EDM is a more advanced type of machine and as a result has complicated wiring which would take more time to build and for me to draw up but I’ll consider it.

Thank you for your feedback. The dielectric used is paraffin (kerosene) and in my opinion is the best fluid to use for a plunge EDM also it is widely available in DIY stores here in the UK as fuel for greenhouse heaters. There is no filtering of the fluid used on that machine but an inline fuel filter would probably be ok. I find that when using paraffin as a dielectric fluid that any particles or soot seem to settle out in just a few hours.

Thanks for unswer. Also, you've promised 10A test, but only 3A in video. Resistors blown or mosfets?) What parameters need to be changed for speed up the penetration process?

@@vasilhryshko208 There are two tests in video. First test is the detail test at 3A. The second test is at max power for the transformer used which is 10A. The test starts at 3A but then switches up to 10A after a few seconds. To speed up the penetration you simply need more amps so therefore a larger transformer.

Thank you for your feedback. You’re right the video is too long but you can just skip through the boring bits. The N-mos transistors aren’t in parallel, only the one at the back next to the large capacitor is the high side switching one, the other one on the right side is just providing the pulse to the other mosfets gate and it can be any suitable type of mosfet or transistor I only used it here because I bought four IRFP250N mosfets as I expected to blow some on the way but everything went ok. The blue relays just switch in three separate outputs from the power generator to the power resistors. The output Amps is therefore just a resistors in parallel calculation. If you’re asking which performs better, the transistor version or the new mosfet version, I’m undecided. The transistor version uses more power and is not as efficient and also suffers from “ringing” on the wave form at 150KHz but it might have a better surface finish and the new mosfet version starts to loose its wave form shape at 150KHz which could lead to higher wear but in the test it seems ok. The new mosfet version is the smallest (half the size) and cheap and easy to make so you decide.

​@@edmguru9056, Do you use ignition voltage. What sort of transformer you r using, steel or switching. Have you ever consider using cheap Chinese MMA welding machine as a main transformer (55VDC)

I don’t use a higher ignition voltage circuit in any of my power generator designs. The transformer can be seen briefly several times in the video and it’s a standard toroidal type with a primary and a secondary winding also I haven’t considered using a welder type transformer. My understanding is that they are of a switching type so the output would be pulses of a high frequency usually between 20 KHz and 100 KHz. This would be a problem as the output from the power generator would be the desired user set pulse rate comprised of a higher pulse rate stream which would alter the desired surface finish and electrode wear rate.

Thank you for your feedback. The LM358 opamp is used as a comparator and it is set to detect the very low voltage pulses at the electrode, which can damage to your work, and then trigger the NE555 timer which in turn stops the pulse frequency to the power generator by effectively grounding or turning off the pulse. A short delay time cycle has to be used because the pulse has to be restarted in order for the EDM to work again until another very low voltage at the electrode is detected when the cycle starts again. If the output frequency pulse is not restarted then the power generator will remain in an off state which is no good as this will just cause the Z axis to retract fully up. This is basically how an EDM works, by sensing the output voltage at the electrode via a voltage divider and comparing it to a reference voltage, above the reference voltage the Z axis goes down, below the reference voltage the Z axis will go up. I hope this helps.

So in previous videos post you have made on RUclips do I understand that you had NO control circuit? The movement of the electrode was just a timed cycle. If a short occurred it remained shorted until that timed cycle was completed the electrode retracted thus breaking the short?

@@Uflyppc Thank you for your feedback. No there is a control circuit to maintain the correct positioning of the Z axis to maintain the burn. The Anti-Arc board is a separate add on feature designed to prevent the very low voltage pulses which could lead to an arc which would damage your work and the timed cycle that the Anti-arc board uses is there to allow the 555 timer to reset otherwise the output will remain in a low or off state. Originally the control and the Anti-Arc circuits were combined into one which made it more difficult to regulate as making any voltage adjustments to one circuit altered the other circuits performance, so now they operate independent to each other.

Thank you for your feedback. My idea was to stop all the very low voltage pulses to the electrode before they can cause an arc and damage the work piece or electrode. It can be done by using a 555 timers output to stop the pulse at the frequency generator output stage before it switches the 100V power output generator on. This seems to work ok but I was just asking if this is the right way to do it, and does anyone have a better way to do it? As you may know arcing at the fine finishing stage on an EDM can damage your finished product and if you can stop it, it will save you having to do a lot of rework or possibly start again from new. I don’t normally get any arcs on my EDM even when at the fine finishing stage but some people have a problem with it but it’s usually just down to bad flushing of the electrode.

Your machine can achieve best surface quality among DIY. Mine has been concentrated only on how the electrode wear. On my electronic board the output from a comparator(pulled up) is fed into a push-pull common emitter output stage then to ON Mosfets. You can interrupt(off) the MOSFETs by grounding input of the push-pull(output of the comparator)

Ok thank you. I’ve just designed a simple new mosfet power generator and quickly rebuilt with “rats nest” wiring and no RF shielding to make it difficult, the original EASY-PEASY machine to test it on. The results surprised me so if you’re interested keep a look out for it when I post it online tomorrow.

When the output from the timer turns off, transistor 2 is no longer turned fully on and returns to normal operation. This sends the square wave pulses to the power generator which in turn outputs the high voltage pulses to the electrode until a low signal is detected and sent again to the op amp, and then the cycle starts again.

Thank you for your feedback. Yes you’re right, the aim is to stop all the very low and potentially arc forming pulses before they can damage both the electrode and the work piece

That is good because the arcing can cause big problems. I see you have a comparator LM358 on the board so I am assuming you are using that in some way as a control the z axis motor circuit (with and upper V-hi and lower V-lo voltage) window. If you turn the power off to the gap how does the comparator get an input signal to tell the motor to move up or down ??

The LM358 is an op amp not a comparator, it can be used as one but a comparator can’t be used as an op amp and the reason I used one here is that they’re more stable than comparators (but not as fast). The output from any design of feedback control to a stepper motor or dc motor is set so that a high voltage signal = Z axis down and low voltage signal = Z axis up so if you’re using a stepper motor to drive the Z axis then this feedback signal would go to the direction terminal on the stepper motor controller which changes the stepper motor direction on a high or low voltage signal. Also don't forget that the pulse off time is very small and that it self resets back to on until the next bad or very low voltage pulse which starts the cycle again.

Thank you for your feedback. Time permitting I hope to make two new update videos soon, one on the anti arc control which is now independent of the z axis controller and one on a new mosfet designed power generator which looks like it has a good output square wave form, but so far its performance is untested.

Thank you for your feedback. If you could describe exactly what the problems are with your machines I might be able to help. Most spark erosion z axis control feedback circuits are based on having a simple voltage divider taken after the output power resistors (current limiting resistors) and a window comparator to set the reference voltage to the z axis controller, which is either a stepper motor controller or a dc motor H bridge type speed controller, to determine the z axis up/ down direction or pause position. Some mosfet type power generators use a mosfet feedback current control by using the mosfets Rds on resistance and a comparator to control the output current. This design type cuts manufacturing costs, is not reliably accurate and also is no good for use in a pulse EDM designed to give low electrode wear rates etc, for fine finish work, in my opinion.

@@edmguru9056 thanks for the reply I have 4 EDMs down 2 Creators 1 Strathclyde with DC servos and a Japax witch is a dual stepper motor drive i think they are using one of the steppers as kind of encoder all of the have spark generating power supplies are working. All are 80s tech with lots of dual sided rack mounted boards making wire chasing difficult. I have no schematics for them or support as all of those companies are tits up. That makes chasing down an issue tougher. There are no tell tale sooty marks or board browning. The Japax is a little different, as it has a box inside a box for the steeper driver but i could not find any issues with the drive parts. It is covered IC's way more than I can keep in my 2 bit brain

I would have thought that all of those machines use dc servo motors to drive the z axis which could be the problem as they are driven by a logic/ pc based feedback system (hence all the ICs) and are more prone to failure than using stepper motor drivers. It should be possible to have four working EDM machines again if you could change the drive motors for large stepper motors and new stepper motor drivers and stepper motor controllers, as everything else on the machines is working, but this would require cnc build knowledge maybe from someone local to you to save on costs? The downside is that you would loose some of the fancy features that these Chinese machines have if you rebuild the power generator control side, but at the end of the day they all work simply by Ohm’s law in that the output from the power generator is passed through current limiting high power resistors (selected output amps) which when eroding causes a voltage drop on the output to the electrode and by using a simple voltage divider at this point to lower the voltage to a lower useable level, you can control a stepper motor controllers (comparator IC) voltage reference point for z axis direction and therefore the stepper motor driver etc. Sorry I can’t be of more help but personally, if these machines are needed, I would rebuild all four because as they stand now they have little more than scrap value. @@raymacklures5899

@@edmguru9056 It is my goal to rebuild them, as covid really sucked the funds out of the company and $50K on a new one is out of the question and 2nd hand ones are rarer than rocking horse shit here in OZ. Your right they are just $60/ton chunks on cast iron it is only the Z axis I need to control I have played with arduino and have made a machine that indexes to 3 positions with a phnematic cylinder holding the part down and a fixed nail gun fires, a nail of sorts into that position then the part is removed and replaced by hand. Having said that I did not write the code a guy I know did, I can read it enough to modifi the values but my brain and time has not evolved enough to write it. I do a lot of Cnc machining, I use a Cam package to do so. I can read G code enough to recognize potential issues and change them But if you asked me to write a multi pass rectangle then put the kettle on..

Thank you for your feedback. The term “commercial grade” in the title sentence refers to the quality of the end result compared to other homebuilt machines I’ve seen online and not to “for profit” as you say but if the new design mosfet power generator, once tested, works even better than my amp based transistor version then I might consider making it a “for profit” venture, but then again I doubt that there is much of a market for an EDM considering how much CNC machining has evolved for home use over the years.

Thank you for your feedback. Wire EDM use a different design of power generator to mine, which is of the pulsed resistor capacitor type and is very fast at cutting steel but with a very high electrode wear rate, which is not a problem in wire EDM as the electrode wire is constantly being replaced. Pulse EDM (my machine) is designed mainly to have a low electrode wear rate for detailed work but the trade off is a slower metal removal rate. I don’t have any circuit boards available as I’ve just designed a new mosfet power generator which I will be testing soon but if you Google image “pulse EDM RC circuit” there are plenty of circuit designs available there.

Thank you for your feedback. The output current is set by the output power resistors (200W min) resistance value in Ohms and is simply a resistor in parallel calculation. Each resistor that I used was 16 Ohms so the output Amps using one resistor is (calculation is at the working voltage of 50V and not the open voltage of 100V) approx 3 Amps (50/ 16 = 3.125). By using two resistors in parallel gives an output resistance of 8 Ohm so current is now approx 6 Amps, using three gives 5.33 Ohms output resistance so output Amps is now approx 9 Amps and so on up to the limitations of your output transistors/ mosfets or your power generators transformer etc. The design of the audio Amp used as a power generator basically can be an A class type but the output will be inverted so use a second transistor on the output to flip it back in phase with the input frequency generator and use a final power NPN transistor set as an emitter follower (output is from the emitter not the collector) as the output to the power resistors, this makes it act like a switch and not amplify which is what is required for an EDM power generator. Also you have to remember that it is a square wave that is being used and not a sine wave (audio wave), so remove the input capacitor shown in most designs or it won’t work. I never bothered with a schematic for the power generator as it’s just a basic amplifier design which is easy to build if you know how amplifiers work also I have just designed a new power generator using power mosfets instead of transistors which hopefully will be more efficient and give even lower electrode wear rates but I haven’t built or tested it yet.

@@edmguru9056 isnt that extremely inefficent? i mean youre just burning power in the resistors, wouldnt it be simpler, and more efficent to just use a buck switching stage instead, with appropriate current limiting. or a switching converter like the common welders.

@@contomo5710 Thank you for your feedback. Yes you’re right it is inefficient to control the output amps like this but it’s the easiest and cheapest way to measure the voltage feedback for the z axis controller, simply by using the voltage drop when the current is shunted through the power resistors, also any other type of control method must be kept stable and work linear to the output voltage to maintain a stable burn rate, which could be a problem unless maybe some type of cpu and coding etc is used, but all that is beyond me I’m afraid.

Thank you for your feedback. I have no plans to sell the power generator at the moment as I’m working on a new mosfet power amp design but you could just follow this basic design and build one. The square wave is set by a square wave pulse frequency generator (1 kHz to 150 kHz) which is sold online. Then simply just input the square wave pulse to an audio based amp design which uses either a transistors or mosfet type output which is connected to the electrode via a high power resistor to control the output amps. If you plan to use a transistor based design (I do at the moment being old school but a mosfet design would be more efficient) then set your final output power transistor (npn) as an emitter follower (mosfet (n channel) as source follower and use mosfet driver chip) as you only want to switch the 100v D.C. power and not amplify it and don’t forget that it’s a square wave and not a sine wave that’s being sent to the power generator so remove any input capacitors to make it work. The 100v output should hopefully end up to be an exact copy of the input square wave from the square wave pulse frequency generator with no ringing or transients present which is important for low electrode wear.

Thank you for your feedback. I’ve had a look at your videos and they look good but I would recommend the following: Use positive polarity at the electrode for lower electrode wear also 100 kHz pulse frequency is a slow, high wear rate surface finishing pulse so at your machines rating of 5A gives a surface finish of 9 microns, max stock removal 12 mm3 per min, electrode wear 16%, spark gap 0.03 mm per side. Best settings for speed and low electrode wear are at approx 10 to 12 kHz giving a surface finish of 20 microns, max stock removal 25 mm3 per min, electrode wear 1%, spark gap 0.06 mm per side. It’s also better to use a dielectric fluid or kerosene rather than distilled water, unless you plan to use the machine indoors, as water only really works best when fully chemically deionised using a resin which is expensive to use. If you haven’t already seen I would recommend that you view my first video “Easy Peasy EDM” which uses a dc motor speed controller to control the edm z axis via a voltage divider taken from the output of the power generator, it’s a quick and cheap way to build a commercial grade output edm machine with no software required, so you would only have to add your mosfet power generator and away you go!

Thank you for your feedback. The square wave is set by a square wave pulse frequency generator (1 kHz to 150 kHz) which is sold online. Then simply just input the square wave pulse to an audio based amp design which uses either a transistors or mosfet type output which is connected to the electrode via a high power resistor to control the output amps. If you plan to use a transistor based design (I do being old school but a mosfet design would be more efficient) then set your final output power transistor (npn) as an emitter follower as you only want to switch the 100v D.C. power and not amplify it and don’t forget that it’s a square wave and not a sine wave that’s being sent to the power generator so remove any input capacitors to make it work. The 100v output should hopefully end up to be an exact copy of the input square wave from the square wave pulse frequency generator with no ringing or transients present which is important for low electrode wear.

@@edmguru9056 Very helpful thank you. I wouldn't call myself oldschool, but I like your NPN more than a mosfet. My biggest concern with mosfet's has to do with another project I am doing where variable impedance coils are giving me pains (temperature ranges are very wide and inconsistent). Wouldn't I need a snubber zenner diode or a freewheeling diode to keep the mosfet from being fried? Because the edm electrode it self does have some inductance that will cause a voltage peak once the falling edge jitters through? That's where I get uneasy because that could provide either insufficient protection, or an extra quirk to dial in... With this being so tricky as it is... I just think the NPN is a much better solution. I suppose a more modern mosfet with built in avalanche protection would work... But. NPN just works. It can certainly do plenty of frequency, and I could care less about EMI. Would love to hear your thoughts. And I would love your advice to build something like this. Would you be open to mentoring me? I would certainly try to compensate you best as possible.

It would be a good idea to fit a fly back diode, just in case, across any output power resistors to protect the n channel mosfet although I would have thought that any back emf generated would be too low to be a problem if the output power resistors have a low coil count (less than 30) and as the output end is not connected fully to ground, but a short circuit might change this. As for mentoring, I’m sorry but I regard my electrical knowledge as only basic so I wouldn’t be of any help to you but good luck with your project.

@@edmguru9056 Thanks, I just was looking for help understanding your approach to building the machine on a more top down level. I have to be able to understand the circuits involved on my own to get anywhere. And tips on machining with EDM since you have so much experience. Do you post on any forums anywhere?

@@edmguru9056 The avalanche issue from back emf seems to occur mostly at the mosfet, inductance on the pcb itself, and that's why it's so tricky. Apparently just a little extra inductance on the traces can have quite a large back emf at the mosfet. And then there's single pulse vs multi pulse emf... The extra inductance of the circuit (in our example the electrode) probably just amplifies whatever "imperfections" exist in the circuit layout. As you said though, just put a diode in and see if it works is just the most practical solution. Or pick a mosfet with very good avalanche breakdown characteristics.

Hi sorry for the delay but I’ve only just seen your message for some reason? As I don’t know your machine design it would be difficult to help but I’ll try. First off I assume that you take a high power D.C. output (say 100 volts) from an isolated mains transformer (important for safety) and use either a transistor or mosfet type power generator with its output going to your electrode via a high power resistor to control the output amps and the power generator is pulsed by a square wave frequency generator set at the required pulse frequency and duty cycle. Now if this is correct then to control the z axis stepper motor you need to send a control signal taken from your electrode via a voltage divider to a comparator or the like which will control or set, your stepper motor controllers (say a TB6600) direction. Adjusting the voltage dividers control voltage (add a potentiometer) compared to the comparators fixed reference voltage will control your output spark gap volts (aim for just under half open circuit voltage (at not working state) compared to working (when sparking) voltage. A 100v voltmeter will be required connected to the negative (work tank) and electrode to monitor this. From your video it looks like you need to increase the speed (reaction) of the stepper motor by simply increasing the stepper motor controllers pulse frequency, aim for approx 10mm per second to start with and adjust either way from there and you’ll also need a lift timer for flushing fitting to the z axis and all should be good. I hope this all helps but it’s difficult without knowing your actual machine design.

@@edmguru9056 Thanks for your reply i have created windows comparator circuit to get the spark voltage using register divider. but the comparator output is not stable. so i can't able to control the stepper motor. i will upload my comparator video on my channel. can you suggest me the stable circuit so i get the correct spark indication. below is the spark gap control circuit video link.if you require any more detail i will provide you. ruclips.net/video/IgEZDBEBbx4/видео.html is there any way i can connect with you using mail. lots of thanks for reply.

I can’t suggest a stable circuit I’m afraid as I don’t understand the comparator circuit from your video, a schematic of the circuit would be better. Also I don’t know what the capacitors in your design are doing as they’re normally not used in an E.D.M. comparator circuit design using a control voltage via a voltage divider from the output high voltage D.C. supply. There’s no other way at present to contact me online as I don’t do social media etc, sorry.

i have followed the same circuit which you have described in comment. but i am not using the isolation transformer. and i am using the mosfet for high voltage pulse.

@@edmguru9056 I have tried to feed the voltage divider voltage directly feed to ADC of Microcontroller. but still not able to maintain the correct spark gap. what circuit you have followed to maintain the spark gap.

Thank you for your feedback. The control of the z axis is by the voltage feedback from the voltage divider to a “cascade voltage” control or “voltage ladder” which is variable and at any predetermined working voltage level sends a voltage signal via a transistor to the stepper motor controllers (TB6600) direction control, if that makes sense. I know that similar can be done by using a comparator but they have stability issues due to the R.F. produced with E.D.M. and also with the control voltage from the voltage divider having the same pulse and duty cycle frequency as the output at the electrode, which again destabilises a comparator. With this “new” method of E.D.M. control it’s possible to remove most of the pulse and duty cycle frequency from the voltage divider which in turn adds a kind of basic fuzzy logic which helps to blur the transition points to the stepper motor controller so basically if the voltage “burn” at the electrode becomes unstable then the transition points should automatically tighten improving the z axis response or that’s the theory anyway, but it does seem to work and be more stable in practise.

@@edmguru9056 Hey thanks for such a swift reply. I understand the cascade voltage control however I am still unsure of the function of the EDM system. I get the impression that as the voltage of the voltage divider gets lower, say a little lower than half the 100V of the rail the stepper drives the tool lower and if the voltage divider gets a little above half the tool is slowly raised - this leads the tool to constantly hunt for that sweet spot up and down leading to a wonderful controlled "burn". As you say the stepper speed is slower those commercial EDM's but still hunts for the correct height... Using a voltage divider rather than a comparitor means much less RF interference. I suppose a faster stepper can keep the tool in the correct zone faster.. Not sure if thats it but I am new to this and quite fascinated..

Yes that's basically how it works, it's all about maintaining a stable "burn" or working voltage.

@@edmguru9056 Thanks for confirming that.. 😊

Thank you for your feedback. I understand what you are trying to say but this video along with all my videos are genuine and if you view part two you will see the proof that this machine works. If it is a fake how am I doing it when you can see the machine working? E.D.M. machines are not difficult to build it’s just that most if not all home build machines use a flawed control design.

@@edmguru9056 I believe you’re correct. The entire premise of edm is very simple. However people think that means there isn’t nuance to everything… meanwhile schools won’t even teach basic electromagnetic information correctly. Please continue to post! I am reading through Ben’s book… do you have any recommendations for resources?

@@th600mike3 Thank you for your feedback. If you’re planning to build a pulse E.D.M. I would recommend that you take a look at the other builds and their problems on here first as most if not all builds that I’ve seen have a flawed control design. I’ve commented on this before but basically they all use a fast comparator which is easily destabilised by the control voltage having the same pulse frequency and duty cycle as the output at the electrode and also by a total lack of R.F shielding as well as other power generator related design issues.

Thank you for your feedback. It’s difficult to explain how it works in brief but basically it’s Ohm’s Law. You take the voltage feedback from the high voltage output via a voltage divider to a controller whose output dictates in which direction the stepper motors controller should send the stepper motor. In this case the controller is a custom design based on a repurposed battery charge cascade voltage monitor. I hope this helps but if you’re interested in spark erosion it might be an idea to research online how E.D.M. machines work and can be controlled.

@@edmguru9056 appreciate❤😊

Thank you for your feedback. This design so far is still being developed and is fast becoming a complicated wiring build. The plan is to get all the features working properly, especially the anti arc, and then try to slim down the wiring somehow to make it an easier build for home use. The controller and the anti arc circuit are integrated for now but will be developed as separate units on two different circuit boards which eventually will solve a lot of problems.

@@edmguru9056 Cool thanks. I guess that is something you decide. Maybe let me know if you have something for a EDM hobby/diy guy to make and build like this. If you could provide a licensed build plan, you could get a couple of cents to put in your dev project. Myself would be interested if available...All the best...

@EDMGuru - me too! I'm interested in build one too - best wishes from Brazil

Thank you for your feedback. The Z axis uses a 200 steps per revolution Nema 17 stepper motor and a 1204 ball screw drive. The stepper motor pulse frequency to the motor controller (TB6600) from memory is variable between 5Khz and 20Khz and the controller is set on 32 steps micro stepping so 6400 steps per motor revolution and 4mm travel per revolution works out at between 3.125mm and 12.5mm travel per second so it's not that fast compared to commercial E.D.M. speeds.

@@edmguru9056 Very impressive build. If you keep track of the step position you can switch microsteps for retraction moves to get more speed for up movements. At least with TMC drivers that don't use micro switches for microstep settings. BTW: Your pulse generator seems way better than my one. I had stable frequencies up to 40khz. Above 40khz it started to fade but still produces fine sparks. Did some major changes and now I have weird troubles getting it back to 40khz. Maybe ground loops or noise stuff. Still debugging. PS: I switched to the fastaccelstepper library. It can create up to 200khz step pulses without blocking the code. Maybe the best stepper library I used so far.

Thank you for your feedback. I agree that by using code to determine the z axis position is ultimately the way to go and is something that I will eventually have to learn how to do. I use 75 Ohm miniature coax cable on any external wiring that’s carries a low voltage pulse and I found that this make a big difference with RF suppression (make sure the shielding has a good connection to ground) and also remember that it is a square wave that you are amplifying/ switching (I don’t know your power amp design but I assume that you’re using an n channel mosfet as a common drain or source follower so are using high side switching which is the best way for an EDM) so any resistor to the gate/ emitter of any amp stage, or on your power mosfet gate will be effected by induction and therefore you will loose the square wave form at the higher frequencies even if using a mosfet driver IC or charge pump circuit so this will need to be checked and balanced using your oscilloscope. I’ve just designed a new mosfet power generator which should in theory give even better performance and a lower electrode wear rate than my previous transistor based one. It uses three IRFP250N high side switched (using a charge pump) n channel mosfets in parallel, each one as a separate output channel to determine output Amps via three separate output power resistors (200W each min) which is a similar design to before and hopefully I’ll get round to building and testing it sometime soon.

Thank you for your feedback. Yes this design does work but it could be better and I’ll improve it further. You’re right about the popular one it has control problems. The machine I’ve built is a true pulse machine, some designs claim to be a pulse E.D.M. but are really a pulsed R.C. design which have higher electrode wear rates as the actual pulse emitted is comprised of multiple frequencies in the Mhz range. I haven’t seen and don’t know the exact design of it but it probably uses a comparator and therefore will suffer from at least two major flaws: Firstly the control voltage from the E.D.M. via a voltage divider although it’s at a lower voltage still carries the same pulse frequency and duty cycle as the E.D.M. and this destabilises the comparator. Secondly the comparator and its wiring need to be shielded from the R.F. produced by the E.D.M. process as this also destabilises the comparator. These machines transmit a lot of R.F. from their high voltage output and if you stand near to one you can actually, using an F.M. radio, tune into the pulse frequency being used especially at the higher frequencies.

@@edmguru9056 I learn that second problem hard way too. I used a twist pair to a travel switch near the slideway. But once in a while, my machine had weird behavior. I tried to solve that problem by putting RC snubber at some input, increase size of C etc. The problem persisted for more than a year but finally I use a shield wire(a salvage wire from an earplug). The problem has been solved since.

Thank you for your feedback. Just like with the “Easy-Peasy” D.C. motor controller, I simply took an existing D.C. motor control circuit and reapplied it to the E.D.M. world, and it worked! This new controller doesn’t use a comparator as such and is partly based on an existing battery voltage monitoring circuit which uses a “cascade” or “voltage ladder” to monitor voltage and again I’ve simply repurposed it to work with E.D.M. and the results speak for themselves. The anti arc circuit switches off the output pulse from the frequency generator on all the low, possibly arc (localised overheating) forming pulses. If the frequency gen pulse output is off then so is the output pulse from the power generator. About the next video, it’s already done and will be uploaded in a few days. I’ve worked with E.D.M. all my life and all I can say for now is WOW!