# need a circuit to generate a pulse for plasma erosion of electrodes

#### sirdavid

Joined Nov 9, 2009
20
I am wishing to erode electrodes into nano sized particles in water, using a pulsed electric arc between the electrodes to accomplish the erosion.

What I need is a circuit, or a recommendation for a piece of equipment, that could generate the arc.

Do you know of a do it yourself circuit that can be used to approximate the following pulses:

Would it be easier to buy a pre-made piece of equipment, such as a pulse generator? There are many sold cheaply on Ebay, but they do not seem to produce anything other than square pulses.

The study that I'm looking at is: http://www.ipme.ru/e-journals/RAMS/no_81808/tien.pdf

The study explains the process in detail:

The power supply system provides a stable pulse voltage for etching the Ag electrodes in pure water. In order to ionize the aqua medium between the electrodes, the DC arc-discharge system provides a pulse voltage of 70-100 V for 2-3 ms and then maintains a pulse of 20-40 V for around 10 µs. At that moment the etching current can reach 4 A. The well-controlled on and off timing is shown in Fig. 2. The servo control system based on a feed-back loop controls the gap between the electrodes which is equal to a few microns. The upper Ag electrode (usually the cathode) is held by the servocontrol system and the bottom one (usually the anode) is fixed by the electrode holder. The container with deionized water is maintained and stirred by using the magnetic stirrer at room temperature. Silver wires are used as both the positive and negative electrodes. The pure silver wires areetched by the DC pulse arc-discharge in pure water. The parameters of the control system were chosen for optimal conditions of Ag nanoparticle production.
The governing parameters of this system such as the working voltage, selected current, pulse duration (on/off time), electrode gap, and temperature of the deionized water are crucial factors for nanoparticle production. During the arc-discharge, the surface layer of the Ag wires is evaporated and condensed in the water. The transparent solution converts to a characteristic pale yellow color (Fig. 3), and then a silver suspension is created.

I would also like to be able to vary the parameters some: varying the voltage, amperage, timing, etc, if that is not too complicated.

If you are able to make any recommendations, I would greatly appreciate it.

David.

#### KL7AJ

Joined Nov 4, 2008
2,229
I am wishing to erode electrodes into nano sized particles in water, using a pulsed electric arc between the electrodes to accomplish the erosion.

What I need is a circuit, or a recommendation for a piece of equipment, that could generate the arc.

Do you know of a do it yourself circuit that can be used to approximate the following pulses:

Would it be easier to buy a pre-made piece of equipment, such as a pulse generator? There are many sold cheaply on Ebay, but they do not seem to produce anything other than square pulses.

The study that I'm looking at is: http://www.ipme.ru/e-journals/RAMS/no_81808/tien.pdf

The study explains the process in detail:

The power supply system provides a stable pulse voltage for etching the Ag electrodes in pure water. In order to ionize the aqua medium between the electrodes, the DC arc-discharge system provides a pulse voltage of 70-100 V for 2-3 ms and then maintains a pulse of 20-40 V for around 10 µs. At that moment the etching current can reach 4 A. The well-controlled on and off timing is shown in Fig. 2. The servo control system based on a feed-back loop controls the gap between the electrodes which is equal to a few microns. The upper Ag electrode (usually the cathode) is held by the servocontrol system and the bottom one (usually the anode) is fixed by the electrode holder. The container with deionized water is maintained and stirred by using the magnetic stirrer at room temperature. Silver wires are used as both the positive and negative electrodes. The pure silver wires areetched by the DC pulse arc-discharge in pure water. The parameters of the control system were chosen for optimal conditions of Ag nanoparticle production.
The governing parameters of this system such as the working voltage, selected current, pulse duration (on/off time), electrode gap, and temperature of the deionized water are crucial factors for nanoparticle production. During the arc-discharge, the surface layer of the Ag wires is evaporated and condensed in the water. The transparent solution converts to a characteristic pale yellow color (Fig. 3), and then a silver suspension is created.

I would also like to be able to vary the parameters some: varying the voltage, amperage, timing, etc, if that is not too complicated.

If you are able to make any recommendations, I would greatly appreciate it.

David.
I'd look for one of the older tube type HP pulse generators. Even on E-Bay, they aren't cheap, but they do a great job for plasma pulsing. We used them extensively in the old UCLA plasma lab. Pretty indestructible "little" fellas.

eric

#### sirdavid

Joined Nov 9, 2009
20
I'd look for one of the older tube type HP pulse generators. Even on E-Bay, they aren't cheap, but they do a great job for plasma pulsing. We used them extensively in the old UCLA plasma lab. Pretty indestructible "little" fellas.

eric
Can you suggest some specific models? I'm quite unfamiliar with pulse generators.

Joined Jul 7, 2009
1,583
A few years ago a friend sent me an HP 214A and it would be the first thing I would try (i.e., I agree with Eric). If you're in the northwest US, PM me and maybe we can arrange something. Be forewarned, however, that it's a boat anchor (i.e., heavy and consumes a lot of power).

Otherwise, it shouldn't be terribly difficult to construct some circuitry that would make the requisite pulses. The basic design would be a charged capacitor and a fast switch (i.e., a transistor) to apply the voltage. You could drive it with a cheap pulse generator or just make your own. However, this would involve hazardous DC voltages on the order of 100-150 V, so it's not something you should attempt unless you have some circuit building experience. The folks on this forum could help you with its design.

#### sirdavid

Joined Nov 9, 2009
20
http://www.sciquip.com/browses/detailed_item_view.asp?productID=24699&Mfg=HP&Mdl=214A

Something like this should do the trick.

By the way, you do NOT need to try to reproduce that waveform....all that ringing is the NATURAL result of hammering a plasma with a square wave!

Eric
OK.

Wow. Now THAT's the type of information I need: that I don't have to exactly replicate all the minutia of the waves.
Can the device you suggested "provide a pulse voltage of 70-100 V for 2-3 ms and then maintain a pulse of 20-40 V for around 10 µs. At that moment the etching current can reach 4 A"

Is there some way I can use a computer connected to a device to generate pulses?

#### KL7AJ

Joined Nov 4, 2008
2,229
OK.

Wow. Now THAT's the type of information I need: that I don't have to exactly replicate all the minutia of the waves.
Can the device you suggested "provide a pulse voltage of 70-100 V for 2-3 ms and then maintain a pulse of 20-40 V for around 10 µs. At that moment the etching current can reach 4 A"

Is there some way I can use a computer connected to a device to generate pulses?

You can certainly use a computer to generate stepped waveforms and pulses. Depending on your budget, you can get high-falutin' and use Labview to generate an arbitrary wave through a DAQ card. Or you can use low-cost DAQ serial cards from B&B Electronics and write your own software to do it.

If you do get an old school pulse generator, it might be a bit tricky to obtain a two-level pulse. However, you may be able to use the ringing at the trailing edge of the pulse to do that for you, with a little external passive waveshaping.

Eric

#### sirdavid

Joined Nov 9, 2009
20
A few years ago a friend sent me an HP 214A and it would be the first thing I would try (i.e., I agree with Eric). If you're in the northwest US, PM me and maybe we can arrange something. Be forewarned, however, that it's a boat anchor (i.e., heavy and consumes a lot of power).

Otherwise, it shouldn't be terribly difficult to construct some circuitry that would make the requisite pulses. The basic design would be a charged capacitor and a fast switch (i.e., a transistor) to apply the voltage. You could drive it with a cheap pulse generator or just make your own. However, this would involve hazardous DC voltages on the order of 100-150 V, so it's not something you should attempt unless you have some circuit building experience. The folks on this forum could help you with its design.
At this point, I'm looking at all the options, so I don't want to disregard any suggestion.
I live in the southwest, Fredericksburg, Texas, to be exact, so I suspect I could not just swing by and pick it up. But, how much would you charge for shipping and all?

Please note on the volt specifications: the DC arc-discharge system provides a pulse voltage of 70-100 V for 2-3 ms and then maintains a pulse of 20-40 V for around 10 µs. At that moment the etching current can reach 4 A

To be honest, the only circuit I ever built was was an LED driver circuit. It did work nicely, but I can't say it had high voltage.
My tools include a soldering gun, some solder and flux, some copper etchant, I think I still have some copper plated PCB, and a Fluke 189 multimeter.

I certainly would want to consider your warnings about the dangerous voltage, but if a reasonably prudent and careful and person can deal with it, I may be OK in that department. (But if you think not, I would consider what you have to say.)

I've been looking at Ebay under the search terms "pulse generator" and about 355 hits come up. Most of the actual pulse generators that come up seem to be in the low voltage range, about 0-10 volts. That's not enough for what I'm trying to do (vaporize electrodes into nano sized particles into water using a pulsed arc).

Beyond that, I guess I'll toss the ball into your court and see what you have to say.

Thanks again.

David.

#### shortbus

Joined Sep 30, 2009
8,311
Hi David, what you are building is a form of a sinker EDM "Electrical Discharge Machine". I've been building one also, but mine is a 15Amp. version.

EDM discharge really isn't a "two step" as far as the pulse goes. The discharge Starts out at the higher voltage and "ionizes" the "dielectric fluid"(water) then as the voltage drops and the current transfers and erodes the metal and a certain amount of the electrode. This all happens during the "on period" of the pulse. During the "off" period of the pulse the eroded material is washed away, the gap cools, and then it's ready for the next "on" pulse.

Here is a link to EDM plasma; http://cadm.zut.edu.pl/pub/prawie wszystko o edm (ang).pdf

Another one; http://erosion.de/Wissenswertes/mwkcirp97.pdf

I ran a EDM for 13 years while I was working. If you want I can post what I'm doing on my design.

Cary

#### sirdavid

Joined Nov 9, 2009
20
Hi David, what you are building is a form of a sinker EDM "Electrical Discharge Machine". I've been building one also, but mine is a 15Amp. version.Cary
That's cool to know. I don't know if more amperage for what I'm trying to do is necessarily better, although it would be nice to have the capacity if I should need it. Then I could experiment with wider parameters.

EDM discharge really isn't a "two step" as far as the pulse goes. The discharge Starts out at the higher voltage and "ionizes" the "dielectric fluid"(water) then as the voltage drops and the current transfers and erodes the metal and a certain amount of the electrode. This all happens during the "on period" of the pulse. During the "off" period of the pulse the eroded material is washed away, the gap cools, and then it's ready for the next "on" pulse.Cary

Another one; http://erosion.de/Wissenswertes/mwkcirp97.pdf

I ran a EDM for 13 years while I was working. If you want I can post what I'm doing on my design.

Cary
Actually, I worked in several machine shops for 12 years, and in one of them, I ran 2 wire EDMs for probably 2 or 3 years.

Yes, I would definitely like to see your work.
What is your goal? to cut metal, or to produce particles?

#### sirdavid

Joined Nov 9, 2009
20
You can certainly use a computer to generate stepped waveforms and pulses. Depending on your budget, you can get high-falutin' and use Labview to generate an arbitrary wave through a DAQ card. Or you can use low-cost DAQ serial cards from B&B Electronics and write your own software to do it.

If you do get an old school pulse generator, it might be a bit tricky to obtain a two-level pulse. However, you may be able to use the ringing at the trailing edge of the pulse to do that for you, with a little external passive waveshaping.

Eric
I am definitely low budget. I found B&B Electronics serial cards, which one would work best?

"Passive wave-shaping?" That sounds exotic. Let's say I obtained an HP 214A (it's the only one I've found so far that is close to the range I need in volts), how would I do "wave shaping?"

Joined Jul 7, 2009
1,583
David:

As I mentioned in my PM, I'm not interested in selling my HP 214A, but I did get it out and turn it on (it's been in the corner for a couple of years). Into a 50 ohm load, I set it up to get a 10 μs wide pulse with a period of 60 μs like you show in your picture.

Your picture shows a pulse roughly 20 V at 4 A. This means a 5 Ω load. I hooked a 5 Ω load to the HP 214A and got the attached waveform. This is the voltage across the load measured through a 1/10 attenuation differential amplifier. Since the pulse amplitude is about half a volt, that means the HP 214A is putting about an amp into the load. The output impedance of the pulse generator is 1200 Ω on the 100 V range, so there's a pretty bad impedance mismatch which is probably causing the ringing.

Conclusion: assuming my HP 214A is representative of the HP 214A's designed capabilities, it isn't capable of providing the current levels you want.

Don

#### Attachments

• 21.9 KB Views: 87

#### sirdavid

Joined Nov 9, 2009
20
David:

As I mentioned in my PM, I'm not interested in selling my HP 214A, but I did get it out and turn it on (it's been in the corner for a couple of years). Into a 50 ohm load, I set it up to get a 10 μs wide pulse with a period of 60 μs like you show in your picture.

Your picture shows a pulse roughly 20 V at 4 A. This means a 5 Ω load. I hooked a 5 Ω load to the HP 214A and got the attached waveform. This is the voltage across the load measured through a 1/10 attenuation differential amplifier. Since the pulse amplitude is about half a volt, that means the HP 214A is putting about an amp into the load. The output impedance of the pulse generator is 1200 Ω on the 100 V range, so there's a pretty bad impedance mismatch which is probably causing the ringing.

Conclusion: assuming my HP 214A is representative of the HP 214A's designed capabilities, it isn't capable of providing the current levels you want.

Don
First, thanks for taking the time to check out your machine.
Tt's a bummer to know that my current line of inquiry has turned up a dead end. On the other hand, it's great to know, because I won't spend any more time on the HP.
Now I'll have to look for something else.

Actually, a fellow named Ken Ashcroft /Prolex Design told me much the same thing earlier today, he said the max on the 214a was 2 amps, and he said:

All I have is an HP 214A pulse generator that will supply 100v peak into 50 ohms. That's only 2A max. and with no multiple output vs time programming - sorry. Good luck on your search.

So kudos to him and you both.
Thanks, guys.

Next question, what will work?

David.

#### beenthere

Joined Apr 20, 2004
15,819
Given that is is difficult to output a 100 volt pulse, why not place the voltage (with the proper limiting resistor and possibly a reservoir capacitor on the anode side) on the anode and use a FET to let the cathode conduct to ground?

#### sirdavid

Joined Nov 9, 2009
20
Given that is is difficult to output a 100 volt pulse, why not place the voltage (with the proper limiting resistor and possibly a reservoir capacitor on the anode side) on the anode and use a FET to let the cathode conduct to ground?
Thanks for taking a look.
Would this be on the 214A?
Would it allow for stepped waves and meet my other requirements?

David.

Joined Jul 7, 2009
1,583
One option would be to look into commercial power amplifiers. In the 1980's, I needed to do something similar to what you want to do (although the power levels weren't quite so high). I bought an amplifier for it. IIRC, it was around $3k, so would probably be around$5-10k now. You may not have that kind of coin in your budget.

Another thing to check out is the EDM industry. Someone may produce a supply that could get near to what you want to do. Talk to some equipment dealers and you might find some junked equipment you could fix. However, it's possible this type of equipment might not be sophisticated enough for what you want to do.

The most promising option, it seems to me, is to build your own. First, you can buy inexpensive pulse generators (see B&K, Global Specialties for new ones). I bought a Global Specialties 4001 a decade or two ago (when they were Continental Specialties) and the thing continues to work and is very useful. It's based on simple TTL circuitry. If I was on a budget, I'd buy a used one on ebay (in particular, I'd look for an HP, Tek, or Philips that had a right of return if I didn't like it). There are many sites on the web that will instruct you on how to build your own pulsing circuits, but this is an area where I think it's best to buy a product and spend your time on what's less easy to buy.

If you do have access to some coin, you could consider the B&K 4076, a 50 MHz arbitrary function generator. I have one on my bench right now on loan and I'm pretty impressed with it -- there's a very nice pulse generator buried in it. While it costs a bit over $2k, it's something I would consider as a nearly ideal general purpose tool. That leaves the power switching. Since AAC is primarily concerned with safety, this isn't something you should attempt if you're a beginner -- this would best be left to an experienced experimenter, as you're fooling with voltages that can kill you (and you don't get a second chance). If I was trying this, I'd rectify the 120 VAC line voltage and store the energy in a big capacitor (rated around 200 V or more). Then I'd use the pulse generator to drive a MOSFET (rated appropriately) to switch this stored energy to the load. Now, if in your original post you meant that there was an initial 100 V pulse of about 2 μs (not 2 ms -- I suspect a misprint in the paper) followed by a 20-40 V pulse of 10 μs, then you're going to need more sophisticated pulsing and switching circuitry to accomplish this. You can still use the pulse generator to control the pulse width and repetition, but another circuit will be needed to shape the pulse that drives the gate of the MOSFET. When you get to that point, the EEs on here can help you with the design. (If in fact a 2 ms 100 V pulse is needed to initiate a plasma followed by the 20-40 V pulses, this is also doable.) If you have the coin alluded to above, the B&K 4076 can very likely produce the exact pulse shape you'd want by programming the arbitrary function generator portion. If you decide you want to purchase one of these, let me know and I'll write a quicky program to see if the desired waveshape can be generated (it very likely can, although you might be limited a bit as to the detail allowed in the shape by virtue of the sampling rate). I'd imagine you need fairly fine control over this shaped pulse because you want to control the interesting physics that's going on in this process. You also need to address the servo system design, which might be more complicated than the pulse stuff. You don't say why you want to do this, but I'll assume it's because you want to fool around with some silver nanoparticles. I took a quick look at the paper you referenced; they gloss over the physics (probably because they don't understand it) and don't give any data on the kinetics. Nevertheless, it's interesting work and, if I were you, I'd certainly contact L. Stobinkski for more details (he/she is probably a physical chemist and can give you lots more data). I wouldn't want to attempt this experiment until I had contacted one or more of the authors and extracted lots more practical details -- they've left out some crucial information. But it sounds like something that would be fun to work on (it sounds like the nano equivalent of a shot tower). Thread Starter #### sirdavid Joined Nov 9, 2009 20 One option would be to look into commercial power amplifiers. In the 1980's, I needed to do something similar to what you want to do (although the power levels weren't quite so high). I bought an amplifier for it. IIRC, it was around$3k, so would probably be around $5-10k now. You may not have that kind of coin in your budget. Another thing to check out is the EDM industry. Someone may produce a supply that could get near to what you want to do. Talk to some equipment dealers and you might find some junked equipment you could fix. However, it's possible this type of equipment might not be sophisticated enough for what you want to do. The most promising option, it seems to me, is to build your own. First, you can buy inexpensive pulse generators (see B&K, Global Specialties for new ones). I bought a Global Specialties 4001 a decade or two ago (when they were Continental Specialties) and the thing continues to work and is very useful. It's based on simple TTL circuitry. If I was on a budget, I'd buy a used one on ebay (in particular, I'd look for an HP, Tek, or Philips that had a right of return if I didn't like it). There are many sites on the web that will instruct you on how to build your own pulsing circuits, but this is an area where I think it's best to buy a product and spend your time on what's less easy to buy. If you do have access to some coin, you could consider the B&K 4076, a 50 MHz arbitrary function generator. I have one on my bench right now on loan and I'm pretty impressed with it -- there's a very nice pulse generator buried in it. While it costs a bit over$2k, it's something I would consider as a nearly ideal general purpose tool.

That leaves the power switching. Since AAC is primarily concerned with safety, this isn't something you should attempt if you're a beginner -- this would best be left to an experienced experimenter, as you're fooling with voltages that can kill you (and you don't get a second chance). If I was trying this, I'd rectify the 120 VAC line voltage and store the energy in a big capacitor (rated around 200 V or more). Then I'd use the pulse generator to drive a MOSFET (rated appropriately) to switch this stored energy to the load.

Now, if in your original post you meant that there was an initial 100 V pulse of about 2 μs (not 2 ms -- I suspect a misprint in the paper) followed by a 20-40 V pulse of 10 μs, then you're going to need more sophisticated pulsing and switching circuitry to accomplish this. You can still use the pulse generator to control the pulse width and repetition, but another circuit will be needed to shape the pulse that drives the gate of the MOSFET. When you get to that point, the EEs on here can help you with the design. (If in fact a 2 ms 100 V pulse is needed to initiate a plasma followed by the 20-40 V pulses, this is also doable.) If you have the coin alluded to above, the B&K 4076 can very likely produce the exact pulse shape you'd want by programming the arbitrary function generator portion. If you decide you want to purchase one of these, let me know and I'll write a quicky program to see if the desired waveshape can be generated (it very likely can, although you might be limited a bit as to the detail allowed in the shape by virtue of the sampling rate).

I'd imagine you need fairly fine control over this shaped pulse because you want to control the interesting physics that's going on in this process. You also need to address the servo system design, which might be more complicated than the pulse stuff.

You don't say why you want to do this, but I'll assume it's because you want to fool around with some silver nanoparticles. I took a quick look at the paper you referenced; they gloss over the physics (probably because they don't understand it) and don't give any data on the kinetics. Nevertheless, it's interesting work and, if I were you, I'd certainly contact L. Stobinkski for more details (he/she is probably a physical chemist and can give you lots more data). I wouldn't want to attempt this experiment until I had contacted one or more of the authors and extracted lots more practical details -- they've left out some crucial information. But it sounds like something that would be fun to work on (it sounds like the nano equivalent of a shot tower).
It seems I need to contact the authors of the study before I do anything else and clarify which unit is correct.

Can you think of any other questions I need to ask them?

#### rjenkins

Joined Nov 6, 2005
1,013
What you describe sounds very much like a 'Spark erosion' system.

Some of these are extremely simple electromechanical devices, basically a spring loaded, movable electrode holder surrounded by a chunky solenoid coil, with the coil winding in series with the electrode circuit.

The electrode is initially in contact with the workpiece so a short circuit, giving a large current in the solenoid which lifts the electrode.

This open-circuits the coil, causing a high-voltage 'flyback' pulse and an arc across the electrode gap.

Once this dies out and is extinguished by the surrounding liquid, the electrode is pushed back into contact by the spring and the cycle repeats.

#### shortbus

Joined Sep 30, 2009
8,311

This is what I'm basing my EDM on. It will do anything that you need for what your trying to do.

Here are some photos of the mechanical parts of my machine. And a photo of a pulse form from one of the books I have on edm.

You do know that you don't have to use a pulse to erode the metal don't you? When the process was first discovered in the 1940's in Russia and up until the late 1960's they used a RC relaxation type circuit. It was only for more speed and to get around patents that they started using pulse.

tells about the history and different types of machines.

#### SgtWookie

Joined Jul 17, 2007
22,221
Well, since different ideas are being kicked around, might as well toss an idea in... bear with me a minute or so.

Marlin P. Jones & Associates is currently having a sale on 24v 4.2A switching supplies, for under \$10 each:
http://www.mpja.com/email/11-10-09a.asp?r=%%ref%%&s=4

Datasheet: http://www.metapo.com/products/New ips/IM101C12-1M.pdf

So let's say you get four of these 4.2A 24v switching supplies, wire them in series, and use a 555 timer to drive a power MOSFET's gate for the pulsing.

Opinions?