Need used sinker power supply for custom machine - or do I?

I need to cut complex cavities and holes in a conductive metal.

Sort of hard to explain, but what I want to build is a sinker EDM that works on material as it's moving. Further, it won't work in the simple vertical format that most sinkers are set up in, so I need to be able to move my plunge axis wherever I please.

Problem is, I don't really know much about EDM. I've never worked in a shop that had one.

I understand the rudimentary fundamentals of how the machines work. My initial thought it just to build one, but I don't want any more projects. Realistically though, just the simple RC circuit would probably suffice, but I'm worried about tool geometry holding up. Basically, I want whatever shape is on my graphite to not wear "too much"....and it's my understanding, which may be completely wrong, that the "look ma, I made an EDM" circuit with a transformer, a few capacitors and a resistor will erode my graphite. I will have excellent dielectric fluid flow at the spark gap.

My guess, any technology that has integrated circuits driving it instead of vacuum tubes would probably work fine. Honestly, vacuum tubes would probably work for what I want to do. Precision is NOT necessary, it's all cosmetic work. What is necessary is that it put out the spark whenever it's close enough to spark, and I guess that it also drive the servo (I could honestly probably get by listening to the bacon fry, but again, and please correct me if I'm wrong, I think that when the spark gap varies too much, you end up with erosion of your graphite). The less bells and whistles the better.

So do I buy the whole big giant machine so I can cut a few holes to learn? Or do I just look for a used power supply? How much hair am I likely to pull out trying to get a vintage sparker to spark assuming I hook one lead to the part and the other to the electrode? I'd really prefer single phase if that's reasonable. Material removal rate will not be very large...any machine that can say run a 1/4" diameter graphite through 1/2" in under 20 minutes would probably be more than enough. But again, I'm ignorant to the world of EDM, so I don't even know how reasonable that is.

Hi Snowman:
I'm trying to wrap my head around what you're trying to accomplish and I'm having some difficulty.
You wrote:
"what I want to build is a sinker EDM that works on material as it's moving"
So which bit is moving...the electrode or the workpiece?
How is the one different from the other in your mind?
Is the part rotating perhaps?

I ask because all sinkers move the electrode relative to the workpiece...the convenient way of course is to be able to do both, and modern CNC sinkers operate just that way.

So something is always moving...in a plunge burn it's the electrode, in a vector burn in any other orientation other than Z it's the workpiece, if it's a compound vector with a Z component it's both.

Any CNC sinker can do this, so there is no need to re invent this particular wheel unless there is something truly unique about what you need to accomplish.

Moving on to how it works at a really basic level:
The voltage across the spark gap essentially controls what motion happens.
There is a sensing circuit to monitor that voltage and keep the spark gap within the proper range; when the voltage is too high the electrode is made to advance, when it's just right discharge occurs, when it's too low the electrode slows its advance, stops or backs up to restore the gap voltage to its desired range.
In addition, the electrode has dumb motion superimposed on it...it is asked to advance to a destination and it may or may not have jumps imposed to temporarily retract the trode to allow cut debris to be flushed.

So in effect the control tells the electrode: "Go to these coordinates, while you're going there pulse a bit every X seconds, but stay within this voltage range across the gap while you're going there whenever you advance.

This requires the control to sense the gap voltage, to drive servos or steppers or a hydraulic servo valve and ram, and to create discharge currents at the rate and intensity you specify.

You must have all three capabilities in the control, and you must have the right motion system to respond to the signals the control sends, so one without the other is useless to your needs.
So if you buy an old power supply from a scrapped machine you're only part way there.

Regarding the practicalities of cannibalizing an old beater and hoping to run it single phase, well I forsee some problems.
Old machines are overwhelmingly hydraulic servo valve (Moog Valve) machines so there is a hydraulic pump, a dielectric pump and odd other bits all of which are 3 phase in the vast majority of machines, so single phase is definitely going to be a homebrew design, not an adapted commercial design.

Hydraulic servo valves were state of the art in the 1980's but the guts are not replaceable anymore and are also not fixable if they fail.
The heart of these systems is the Moog valve and once they start to leak or gum up from neglect they are toast, and so far as I know, cannot be rebuilt.
The Moog valve is what drives the hydraulic ram of the Z axis up or down in response to the gap sensing circuitry...obviously it shunts hydraulic fluid from the pump and accumulator into the proper end of the hydraulic cylinder depending on what signal it receives from the control.
So if you buy a beater you have to get lucky.

More modern machines use a servomotor and ballscrew to do exactly the same thing.

So when you contemplate this project, think of all these things and see if you can meet your needs with an off-the-shelf solution instead.

Cheers

Marcus
Implant Mechanix • Design & Innovation > HOME
Vancouver Wire EDM -- Wire EDM Machining

Thank you Marcus.

With your explanation, my clarity is enhanced.

I am having difficulty with the concept of feed rate as it compares to voltage (at least I think it’s voltage?). In a simple plunge of consistent surface area, voltage is maintained as you say by feed rate. But feed rate is dependent upon material removal rate. In a Cnc system I don’t imagine you program a feed rate, instead, a voltage?

If I’m plunging say a hemispheric shape, are you still setting the voltage and the current just increases as surface area increases?

And you are very correct...I have zero interest in troubleshooting a vintage hydraulic servo valve.

I really don’t have an issue with three phase on anything that can have a vfd or rpc, but I am not crazy about trying to run 3-phase control electronics off a manufactured 3rd leg on a machine that creates emi...just sounds like more frustration.


Sent from my iPhone using Tapatalk

Hi again Snowman:
You wrote:
"voltage is maintained as you say by feed rate".
Actually sort of the opposite occurs; the voltage across the gap is the driver and the electrode moves in response to that.
There really is no "Feed rate" ; what actually occurs is that the electrode travels toward the workpiece at high speed until the gap voltage drops enough to trigger a spark which blasts a bit of the workpiece away, thus making the gap bigger so the voltage goes up and the electrode advances until the gap voltage drops again and the next spark is discharged, opening the gap again, increasing the voltage again, advancing the trode again and on it goes.

The "feedrate" depends on how quickly chunks are chewed off the face of the workpiece: spark intensity, spark frequency, surface area of the burn face, adequacy of the flushing; all these factors have a pronounced effect on the size and number of chunks chewed away per unit of time.

No matter how quickly the trode is driven toward the workpiece, the rate of advance of the burn is completely governed by the state of the gap and how quickly it can be made to dissolve away in front of the electrode, and not at all by the speed with which the electrode is driven forward by the motion control system...fast electrode motion toward the gap is super useful in pumping dielectric fluid into the gap to clean it and enable proper gap voltage and timely discharge, but it makes the process faster only by that indirect mechanism.

So it's fundamentally unlike chipmaking where you set a feedrate and that feedrate governs the advance of the tool through the material.
Instead, you set the spark duration, intensity, and frequency (which will control surface roughness, electrode wear, and discharge voltage among other things).
You then point the trode at the workpiece and it will go as fast or slow as it goes...you cannot force it with a "faster" setting.
You can only manipulate those settings to knock off bigger chunks at a faster rate but you do so at the expense of other things.

Cheers

Marcus
Implant Mechanix • Design & Innovation > HOME
Vancouver Wire EDM -- Wire EDM Machining

implmex said:

So it's fundamentally unlike chipmaking where you set a feedrate and that feedrate governs the advance of the tool through the material.
Instead, you set the spark duration, intensity, and frequency (which will control surface roughness, electrode wear, and discharge voltage among other things).

Click to expand...

And this is what I was having difficulty processing, is that it’s fundamentally different. In chip making, as long as you can maintain a set of conditions you can remove material as fast as you can feed.

So spark duration is most linked to tool wear. Increasing duration also increases material removal?

Intensity controlling surface roughness? (And what is intensity actually controlling? Current?)

And frequency “just” controls how often the other two happens.


Sent from my iPhone using Tapatalk

Hi again Snowman:
So electrode wear like everything else in EDM is multifactorial.
A basic rule of thumb is that EDM is backward compared to chipmaking in that fine burning consumes more of the electrode than roughing does.
This is highly counter intuitive and the explanation for precisely why this is so begins to stretch what I know about the process.

On another note; when you talk about "spark duration" there are actually two components to it.
The first is the pulse frequency...the more sparks per minute the shorter each pulse has to be in order to crowd them all into that minute.
But this is NOT really what's referred to by EDM drivers when they talk about spark duration.

Here's my understanding of it:
Suppose you have 60 zaps per minute so each zap takes one second from start to finish before the next zap occurs.
Within that one second you can have a 3/4 second "power time" and a 1/4 second "recovery time" or you can have a 1/4 second "power time" and a 3/4 second "recovery time" or any variant up to about a 95% power time and down to fractions of a percent power time.
So you can go BLAST wait BLAST wait, or you can go blast WAIT blast WAIT.
The recovery time is when the debris is flushed from the gap and the dielectric oil recovers its dielectric properties.
Obviously, a big BLAST knocks off a big chunk, but too many big blasts too quickly in a row destroy the properties of the dielectric in the gap so it becomes erratic with respect to its ability to conduct a powerful spark that will actually erode anything from the workpiece.

The wait time within the spark cycle is a necessary evil, so the control (or you if you have an older machine) try to minimize it while still maintaining the dielectric properties within the gap.
That frying bacon sound so beloved of old EDMers is the point where you've turned down the spark duration within each pulse to the point where the dielectric can recover enough to support the next spark and it varies with a whole bunch of other things that are changing all the time, so a modern control will sense what's going on and adjust spark duration on the fly.

Linking all this to electrode wear is complex when you consider all the variables (never mind some basics we haven't even touched on yet, like whether the electrode is positive or negative polarity and what it's made of, and what you're burning, and what kind of dielectric and how trashed it is and on and on).

However the basic premise holds that roughing uses up less electrode than finishing does, and this is a concept that you need to wrap your head around.

On a final note; most of what I know comes from Bud Guitrau's excellent book on EDM and how it works.
Bud has a clear concise and highly readable way of describing the process and how to think about it, so I highly recommend you go out and get yourself a copy.

The book is "The EDM Handbook" and it's available on Amazon among other places for about sixty bucks (and is worth every penny).

On a very last note, Bud, if I've understood you wrong and explained any of this wrong please feel free to give me a nice dope slap and correct my babble.

Cheers

Marcus
Implant Mechanix • Design & Innovation > HOME
Vancouver Wire EDM -- Wire EDM Machining

Marcus,

Thanks for plugging my book, but it needs a serious updating. Instead, Snowman can click this link for a fresher version of the EDM Fundamentals. More than just on- and off-time, there is additional information regarding the influences of voltage and servo-voltage which should help Snowman understand these better.

"On a very last note, Bud, if I've understood you wrong and explained any of this wrong please feel free to give me a nice dope slap and correct my babble."

Click to expand...

Not necessary, my friend. Your explanations are clear, accurate and described on a shop-floor level which everyone understands and appreciates. Please continue with your valuable help.

Bud

implmex said:

On a very last note, Bud, if I've understood you wrong and explained any of this wrong please feel free to give me a nice dope slap and correct my babble.

Click to expand...

No, you've done an absolutely wonderful job interpretting MY babble and helping me understand what I'm up against.

Thank you!