EDM process technical questions (not a production set up!)

superUnknown · Jan 14, 2014

I built a little 30kHz 60VDC 10amp power source and am having success making millions of sparks.

I've studied the Machinery's Handbook and poked around the internets. I haven't had any success with metalic electrodes, but graphite works, barely. Maintaining spark gap is next on the list, then de-ionized water as dielectric.

The frequency and current are adjustable, and the spark gap will soon be. On my first attempt cutting a razor blade, it took about 45 minutes using air as the dielectric. Obviously having re-cast trouble, so I'm hopping switching to water will be a huge leap. I'm going to the drug store to get de-ionized water and have a submersible pump to do the flushing. Nothing for filtering just yet.

Since I've never seen an EDM in the flesh, I'm thinking there are lots and lots of little traps that I'm missing. I'd appreciate you pointing out any you see, bearing in mind this is just a little hobby toy.

Thank you!

Red James · Jan 14, 2014

Frequency is NOT the primary factor in your current. EDM's are more concerned with ON-TIME and OFF-TIME of a square wave. It sounds as though you have a rectified 30 KHz power supply. If your signal is a square wave, even at a constant frequency you should get decent but slow results. Try reversing the polarity from what works with your pencil (low grade graphite), use the di water (distilled may be easier to find and should work until its too contaminated). Hope this helps a little!

andrewmawson · Jan 14, 2014

The dielectric used in EDM has a huge influence on what happens. My sinker EDM uses a refined non volatile fluid that holds of the discharge until the voltage at last breaks down the fluid and there is a discharge that explosively removes material. You certainly want to avoid arcing at all costs

Read this link and it will tell you a lot about the process:

http://www.oelheld.com/fileadmin/content/pages/Prospekte_pdf/oh_Senkerosion_dt.pdf

Sorry that isn't in English if you go to their site you can read it in English - it is a MUST in your case

Andrew

superUnknown · Jan 14, 2014

Cool, the on/off time is sort of adjustable, so I'll try it with 40/60, 50/50, 60/40?

superUnknown · Jan 14, 2014

That's a great website. Thank you!

Keyepitts · Jan 14, 2014

You can also use kerosene as a dielectric, make sure the part is submerged at least by at least a half-inch though. Flushing is pretty much mandatory to remove the debris from the burn area and keep the electrode from arcing; you will not have much success until you have a good flush set up.

Zonko · Jan 14, 2014

This got me wondering.....the *old school* quick and dirty tap burner circuit uses a capacitor. Its charged via an impedance large enough to prevent a sustained arc. That simple.

Are you charging a capacitor, using some other form of matching network ? Or feeding directly into the load. How is the current controlled ? Shunt resistor, current sense xfmr ? Or current cotrol by passive components.

Gets me really curious because the way i understood it was that a pulse of extremely short duration but high power is optimal. Basically the total opposite of a simple transistor circuit. Thinking along the lines of using some sort of L/C network and a few IGBTs. Those buggers can handle several 100 A pulse. (And cost only a few bucks these days)

I won't have the time to actually build one right now, but should we ever really screw up a tap in a very valuable piece...i will have to build one...probably.

superUnknown · Jan 14, 2014

Zonko said:
This got me wondering.....the *old school* quick and dirty tap burner circuit uses a capacitor. Its charged via an impedance large enough to prevent a sustained arc. That simple.

This is all square wave control out of an Arduino. I tried to use IGBTs, but the gate trigger needed waaaaay to much oompf. IGBT is overkill for this low voltage. I'm using a 20amp Darlington direct from the 20mA output on the Arduino. The gate impedance controls the ON/OFF percentage. Current limiting resistors control the spark current. The spark gap voltage is Armstrong control, soon to be Arduino controlled as well. DC source is 60V 10amp (max). Nothin' fancy, just having fun learning!

Zonko · Jan 14, 2014

Thanks for the response. But what in gods name is "Armstrong control". Google spits out no circuit topology under that name, i have to assume it is a commercial product ? There is an Armstrong circuit, but thats an analog oscillator.

Btw:
A totem pole voltage follower should drive your IGBTs. Pretty much any half decent pair of NPN/PNP transistors larger than To-92 package should do.

In case you have not seen it yet:
C. Scott Ananian : MAS 863 : final project : power electronic design

Some good stuff there. Is a bit complicated and "messy" , but it seems to have all sorts of bells n whistles.
The one thing which really bugs me is the complicated drive circuitry. 2 "Bootstrapping" IR21 series and a charge pump.....wtf.

But i bet there is a reason for it. I will study the circuit a bit more.

andrewmawson · Jan 15, 2014

Key to success is your servo circuit. It needs to approach the work under control with the voltage across the gap being monitored as part of the feedback loop, and as the discharge happens withdraw the electrode to quench any arcing, then start approaching again. On my diesinker there is a Moog proportional hydraulic valve controlling the ram hydraulic cylinder.

A very crude but effective tap burner can be made by a solenoid, mounted vertically, with a soft iron armature in it with the electrode at the lower end. The electrical circuit is made when the electrode contacts the work, the armature is pulled up breaking the circuit, then descends to restart the process. You need a weak spring above the armature to prevent it flying up too far

Using copper tube and a constant water drip this will remove taps quite nicely and would go through your razor blade in seconds.

Andrew

superUnknown · Jan 15, 2014

Zonko said:
Thanks for the response. But what in gods name is "Armstrong control"
Btw:
A totem pole voltage follower should drive your IGBTs. Pretty much any half decent pair of NPN/PNP transistors larger than To-92 package should do.

In case you have not seen it yet:
C. Scott Ananian : MAS 863 : final project : power electronic design

hahahahaha Armstrong is tradesman slang for manual operation: e.g. old cars that had no power steering are "Armstrong" steering.
I wrestled with the IGBTs for a day and gave up, but I will try your suggestion for a TIG welder inverter circuit that I am dreaming up. Thank you!
I have seen that MAS project, I agree too complicated. "Bootstraping" turned me off too.

superUnknown · Jan 15, 2014

andrewmawson said:
Key to success is your servo circuit. It needs to approach the work under control with the voltage across the gap being monitored as part of the feedback loop, and as the discharge happens withdraw the electrode to quench any arcing, then start approaching again. On my diesinker there is a Moog proportional hydraulic valve controlling the ram hydraulic cylinder.

Andrew

Ja, I figured spark gap was critical. I'm building a simple voltage controlled servo. Spark gap over 25Volts moves the electrode towards the work, under 20V moves the electrode away. Simple voltage divider on the 0-60V discharge circuit to get a 0-5V signal to the arduino measuring the spark gap and pulsing to a DC gearmotor.
What remains to be seen is if the feedback is fast enough to keep the discharge working and if the arduino can handle all the electrical noise...

andrewmawson · Jan 15, 2014

The discharge needs to stop, not be maintained. Then the electrode creeps down to the work to restart it. Meanwhile the electrolyte that was vapourised by the first discharge collapses / condenses and it is that vapourisation and collapse is what is removing much of the material like the cavitation that happens when a boat propeller runs too fast.

Andrew

Scottl · Jan 15, 2014

andrewmawson said:
Key to success is your servo circuit. It needs to approach the work under control with the voltage across the gap being monitored as part of the feedback loop, and as the discharge happens withdraw the electrode to quench any arcing, then start approaching again. On my diesinker there is a Moog proportional hydraulic valve controlling the ram hydraulic cylinder.

A very crude but effective tap burner can be made by a solenoid, mounted vertically, with a soft iron armature in it with the electrode at the lower end. The electrical circuit is made when the electrode contacts the work, the armature is pulled up breaking the circuit, then descends to restart the process. You need a weak spring above the armature to prevent it flying up too far Using copper tube and a constant water drip this will remove taps quite nicely and would go through your razor blade in seconds.

Andrew

Andrew,

That is very interesting! Would you use a simple R-C power supply with that? Years ago I tried building one of the simple EDM circuits that use light bulbs, capacitors, and an SCR. Trying to maintain electrode current using manual means was a real problem. The backlash in the old milling adapter I used caused the circuit to break every time I tried to back off the electrode.

andrewmawson · Jan 15, 2014

Several years ago I had a commercial 'tap disintegrator' that was simply the solenoid arrangement I mentioned powered by a 24 volt ac output from a transformer. It had a stand with a small tank of water, a throttling tap, and a bit of rubber hose. It worked. Maybe not the most sophisticated but hey - it worked

Andrew

Zonko · Jan 16, 2014

superUnknown said:
hahahahaha Armstrong is tradesman slang for manual operation: e.g. old cars that had no power steering are "Armstrong" steering.
I wrestled with the IGBTs for a day and gave up, but I will try your suggestion for a TIG welder inverter circuit that I am dreaming up. Thank you!
I have seen that MAS project, I agree too complicated. "Bootstraping" turned me off too.

*Big grin here* Now that was funny.

For an inverter you will most likely need to drive a bridge configuration of IGBTs.
The IR 21XY series is good stuff. This bootstap capacitor based principle works. I have worked with several circuits using it over the past year and i like it very much. Simple and easy low and highside drive in a $1.10 IC. Splendid.

The heavy duty approach is to use a switchmode supply to produce several isolated DC rails. One for the low side and one for each high side.
Use Opto-couplers and totem Pole stages to drive the IGBTs. That is how big VFDs etc. do it.
But things can go wrong there. The transformer which creates the isolated voltages needs to have low winting to winding capacitance or you can burn out drivers. Some folks tried using cheap old mains frequency transformers , that failed badly, since they were not designed to isolate high frequency VFD voltages they have too much capacitance.

Actually i am currently working on a Stick welding machine. The circuit is partly proven, it works, but I am stuck because i never find time to get it all done. Too much other nasty problems and intresting challanges in my life right now. I hope to have the first really usable test setup in summer.

I will send you some specs and pics then, in case you are intrested.

superUnknown · Jan 16, 2014

Zonko said:
Actually i am currently working on a Stick welding machine. The circuit is partly proven, it works, but I am stuck because i never find time to get it all done. Too much other nasty problems and intresting challanges in my life right now. I hope to have the first really usable test setup in summer.

I will send you some specs and pics then, in case you are intrested.

Hahahaha. We are on the same page my friend! Usually, by the time the parts arrive for one project I have started to lose interest and am dreaming the next. I'll get back to the stalled projects in a decade or three. As you write; too many interesting challenges!
Please do send me specs and pics. I am very interested. You see, I have a little DC TIG welder that I'd like to build a module for outputing square wave AC TIG (aluminum). At first glance, nothing too difficult something like a motor driver H bridge that can handle 150 amps?

Nerdly · Jun 7, 2018

this was a super cool project...did you end up going any further?

EDM process technical questions (not a production set up!)

Aluminum

Stainless

Hot Rolled

Aluminum

Aluminum

Aluminum

Stainless

Aluminum

Stainless

Hot Rolled

Aluminum

Aluminum

Hot Rolled

Diamond

Hot Rolled

Stainless

Aluminum

Plastic

Similar threads