Stacking Supply Voltages

[shortbus=]

A while back I made a supply much like this.
I rectified and filtered the power line to get 140VDC. The built a small isolated 0 to 30V DC supply that sat on top of the 140V. In this case we never needed to go below 140 and never above 170. So the switching supply was very simple and small. (low voltage parts)

ronsimpson , Would you show how this was done, please? Didn't want to hijack the thread but do want to know how this is done, for an application of mine. Everyone says it won't work, that the highest voltage source will cancel the lowest. I need to have a continuous 100VDC with a periodic 170VDC spike. Any help in doing this will be much appreciated.

 

[ronsimpson]

Here is an example of making a 30 V signal on top of a 150V DC.
The LT1001 is just just a picture. We need a audio power amp here.

Shortbus=,

tell me about this spike.
100VDC I understand.
How often does the spike come? 1/sec?
How long is the spike? 10mS?
Sharp edges like a square wave?
How flat is the top? Can it slope a little?
How much current?

 

[shortbus=]

Thank you very much for this Ron. And to MysticalDork, I apologize for hijacking your thread, I'll ask the mods to spin this off to a separate one of its own. But didn't know how else to get Ron's inputs on this.

This circuit I'm trying to build is for a DIY die sinker/ram EDM machine. Most of the plans out there are for the earlier type R-C oscillator type. But I'm trying to build the next generation of them, which was a pulsed type.

The extra 70VDC is to get the pulse to break down the spark gap between the electrode and work piece more reliably. The one set of plans for this type machine, the builder admits that it 'skips' starting the spark pulse fairly often. The electrode and the work are in a bath of oil and the break down voltage of the oil is ~170VDC, but after the break down of the oil the actual working voltage of the spark is in the 25 to 35 VDC range.

So I want the extra voltage for around 2 to 5 microseconds. The actual "on" time of the sparking will vary (depending on electrode size) from 25 to 100 micro seconds. with an "off" or recovery time of 25 to 45 microseconds depending on the "on" time.

The wave shape is a square or rectangle shape the positive side of the wave should be as square as possible but the wave also has a normal, for the process 'droop or drop' to it. I'll try and post a picture of the wave shape I'm looking for.

Here is a link to explain what I'm trying to do - http://www.engineersedge.com/edm.shtml
And another that kind of shows the wave form needed - http://www.edm-products.com/Dielectrics/ifase/ifase_1.htm

 

[ronsimpson]

shortbus=,
While you are cutting it looks like the current is in the 1 to 500A range and the voltage might drop down to 40 volts. (time from 1/1000 to seconds)
To get the spark started it looks like 200 to 400 volts is needed. I don't know how much current.

If I were to do this I would build two different supplies. (diode ORed together)
1) heavy, current limited or power limited. adjustable.
100 to 5,000 watt adjustable. Max voltage, maybe 100V, but will run at 35V. (fast turn off)
2) Max voltage, maybe 400V or what ever the spark start voltage is. (low power) It will current limit and collapse when the spark starts. (then turns off)

All supplies off.
Turn on supply1. Ramp up to 100V. (below spark voltage) I think even 50V is OK.
Turn on supply2. The voltage heads up until the spark starts.
The current increases and supply2 gives up and turns off, leaving supply1 to hold the high current.
Supply1 will be in current or power limit mode.
Turn off supply1 until the spark clears. Then turn back on.
Wait for supply2 to start the next spark.

 

[shortbus=]

Thanks again Ron!! And thank you to the mod that moved this from the other tread!! (It was JimB who did the deed!).

The voltages and currents you listed look like the ones used in the links. They are for a commercial machine that works with very large electrodes and work pieces. What I'm attempting is on a smaller scale so amps will be scaled down. Do I built the mechanical parts of the machine about 15-20 years ago, then started to try to learn enough electronics to build a power supply for it.

There are quite a few plans out there for the R-C style. All using ~100VDC and ~2A to 5A. I lucked into a couple of servo motor power supplies with 100VDC, 20A, and going to use the transformer - rectifier - capacitor from one, for the main voltage. So your saying to make a smaller 100VDC supply to add on? Do you think that around 0.5A would be enough for the spark start part of the circuit? That's what I was aiming for any how.

Your the first person on any of the forums that has understood what I'm trying to accomplish. They all say it won't work, I show that it's done in industry all the time but they still say electronics doesn't support making sparks like this. My first thought was to use an inductor to get the higher add on voltage, kind of like a one shot boost convertor, with no cap, just a discharge into the spark gap.

About the diode "or". The guy's at AAC said that won't work either. That the higher voltage would reverse bias the diode of the lower voltage and shut it down. But when I saw that you were doing something similar, I had to ask you. Thanks again for your help!

 

[Tony Stewart]

When putting multiple sources in series, it is imperative they both be much lower impedance than the load. Impulse sources tend to be inductive flyback with a zener. Pulse duration depends on L & Rload

 

[shortbus=]

Thank you JimB for the thread move!!

Tony I'm not understanding what you said. I'm a newbie to designing circuits. In this particular case the load will be a spark gap of around 0.001 inch. From all the literature and from running these machines for ~13 years at work the spark gap is somewhere around 1 to 1.5 ohms resistance. The details of the circuits are very close guarded secrets that companies don't share. I understand most of the parameters of the work but the circuits are unknown.

 

[Tony Stewart]

Source impedance for impulses will be the ESR of the Cap.

Load impedance of the arc will be inversely proportional to the energy (J) and density (/sq.mm.) in the arc. ( according to Faraday and others after him) The voltage to arc 0.001" only has to be > ~1kv/mm or where 1mm ~ 0.040" or >25V/0.001 This is for wire to wire in 50% humidity. These are just Rules of thumb from my experience and it increases to 3kV/mm for flat surfaces and increases with vacuum or high pressure and lowers with dust to 0.5kV/mm for wires or point sources.

What industrial process is this used for?

 

[shortbus=]

Thanks for the answer, this is for an EDM, electrical discharge machine or machining. The links in post #3 will explain it better than I can. Using an electrode made of graphite or copper, that is in reverse of the form you want, it "burns" into the work piece that form or shape. Anything made from plastic or rubber is made in a mold that was made with an EDM machine. It doesn't matter how hard the metal is, as long as it is conductive it will EDM. Look at it as welding in reverse.

Every one not familiar with EDM quotes the 1KV/mm for a spark gap, but this is not taking place in the open air. It is done submerged in dielectric oil. And as the oil is used and filtered there is very small particles of metal and the electrode in it. This also helps to make the spark gap more conductive, like your dust in the air quote.

Some of the DIY EDM machines use as low as 50VDC as the open circuit voltage. But they also have problems with 'missing' sparks, the spark gap not ionizing. The industrial machines now use a 'spike' voltage to assure that the spark always happens, this is what I'm trying to do with my circuit.

So as to the ESR of the capacitor, your saying to look for a higher ESR value? Instead of the more often used low ESR?

 

[ronsimpson]

50 volt supply sends voltage to the spark gap through D1.
L1, MOSFET, D2,D3 and IC (not shown) makes a boost up supply that can make hundreds of volts of boost. Current goes through D2,D3 to the spark gap. C2 should be small like 10uF or 1uF. When the spark starts the IC will shut off the MOSFET. This is very simple. The "50V" supply needs to be special but you can start with what you have. ron@gradllc.com

 

[shortbus=]

Thank you again Ron. What would the IC that's not shown be? I found a trans former that has a 150V secondary on it that I will rectify and use for the spark starter/spike, like the first schematic you showed. That should give me around a 210VDC starting spike.

The more I read on inductors the less I understood. I really think the voltage spike needs to be more controlled, and from what I gathered an inductor can't be made to give a certain voltage when turned off. Is this correct?

The more I look at this last schematic, the more I understand what you did. Wish I had the skills to think like this. Thank you so much.

 

[Reloadron]

Maybe I can help a little with the Inductor thing. Let's keep it pretty simple. We know a capacitor opposes voltage change right? Well an inductor opposes current change. That is what they do and they do it quite well. If we hang a big capacitor across the output of a power supply and the load is changing constantly that capacitor will help to oppose voltage changes caused by the changing load. Well an inductor will do the same but for current. When a current passes through an inductor magnetic lines of flux emit from the core. If the current remains a constant those flux lines do the same. However, if the current rapidly drops the lines begin to collapse but if the current jumps back up the lines again emit. This action opposes the sudden changes in current. An Inductor does not regulate voltage or current, it merely opposes a change in current.

As to the EDM thing. While I can understand your objective here there are a few things I don't quite understand. You want to induce a submerged arc by using a voltage. I understand that much. EDM uses a pulsed arc. Before the arc is initiated there is an open circuit voltage between the electrode and workpiece. The instant the arc occurs that open circuit voltage drops rapidly and the current which was zero increases rapidly. The resulting voltage and current will be a function of the arc gap. Would that be correct? The actual arc is like a resistor between the electrode and workpiece. The actual voltage drop between the arc points becomes the arc voltage and is generally a low voltage with a high current. I have seen EDM arc voltages around 9 or 10 volts with currents between 20 and 200 Amps. Some start with an open circuit voltage of 90 volts and use an RF pulse to initiate the arc. Some simply just pulse a high enough voltage to arc. The actual arc voltage is always fed back to a servo motor to control the arc gap as material is removed from the workpiece. What are you using or plan to use for current and how do you plan to move or dive the electrode as material is removed and the arc gap changes?

Ron (The Other Ron)

 

[ronsimpson]

Reload(good name),
If some people are using RF to start the arc then remove C2 or make it small and let there be ripple. I was thinking of letting the boost up supply have 100khz spikes on it. (needs protection, but that is simple)

I was thinking the base power supply probably needs to voltage limit at 50V or something. And power limit at 1000 watts. (this is a small system)
50V @ 0 to 20A = 0 to 1000W,
25V @ 40A
12V@ 80A
I don't know if we need to power limit or current limit. Either way it needs to survive continuous shorts. I like the idea of power limiting. I see the idea of watching the arc voltage to know when to advance the tooling.

The boost up supply could easy be 100 watt. So 500V @ 2A or 250V@4A or 120V @ 2A. That should start the arc. I think 400 to 500 voltages limit and 100W power limit. As soon as the arc starts it will current limit the boost up supply and cause it to shut down. Then the arc will cause the main supply to current or power limit.

The other Ron

 

[shortbus=]

Thank you Ron the other. I understand the basics and the differences between the cap and inductor. But where my knowledge stops is about the actual working of an inductor when it is turned off after it is on. The way I understand it, the voltage can go to a very high unpredictable voltage to keep the same current. But like many things electronic, this is just my understanding, from self learning. Now in a DC-DC boost convertor this voltage is controlled with the control chip, but I''m not using a chip. It was going to be a one shot convertor with no regulation.

An EDM power supply is kind of like a constant curren welding supply used for stick or TIG welding. It counts on the voltage to start the arc, then voltage drops for current transfer. The add on/spike voltage is just too make sure the arc starts with every pulse, it will only last a fraction of the total on time of the arc. Lost arcs or pulses are a big problem with all of the DIY designs and the earlier industrial machines. Never knew they were using RF as an arc start.

You have the EDM circuit down pretty good. My cutting pulses will be around 100VDC open circuit. The pulses will come from charged caps of various values, that are switched to meet the requirements of the electrode size being used. To get a good finish in EDM all of the pulses should be close to the same value, that's why the use of caps. The ram/electrode will be moved by a stepper motor and ball screw, that I built years ago.

The stepper/servo control is regulated by a window comparator wired to the electrode feed. A high voltage will advance the electrode into the cut, a low one will retract it. In the window will mean it is cutting correctly, sitting still. The newer machines may use the 9 - 10V gap voltage and can since most of them are CNC. But the old ones like I'm trying to duplicate, used 25 - 35V gap voltage, as do most of the DIY ones out there. This is to prevent gap shorting due to less precise control than CNC.

You guy's that have done electronics for your whole life don't realize how hard and complicated it is to understand this stuff. When your trying to do it at 67 years old after two strokes. But I thank you all for helping an old guy to complete his project.

 

[Reloadron]

Now I understand what you are getting at with inductors. This is a good read on the subject and it is kept simple with nice little animations.

My friend the EDM machine, such fond memories. I was not a weld engineer, never even played one on TV. However, there was a time during my career when I reluctantly became involved with nuclear welding systems and EDM machines. Been retired two years and it was around 15 years before I retired so it was some time back. We had a large row of 5 or 6 CNC machines called Okuma Street (named after the machines). Every now and then a tool would break off in a part being machined. This is a ***** when the part is at the $25K plus point in the machining process. So how to remove a broken tap or drill? Place the part on an EDM machine and let the machine borough down into the broken tool. Slow but steady and bubble by bubble the snapped tool material was washed away until the broken tool would collapse on itself. I did redesign an old 60s vintage EDM machine which used Thyratron tubes to run on an updated solid state power supply. That was long ago but I was forever sucked into weld department issues after that. EDM and TIG Weldment are just a few things that fascinated me over the years.

Youngstown area huh? My future daughter-in-law is a Youngstown girl.

Ron

 

[Reloadron]

Reload(good name),
If some people are using RF to start the arc then remove C2 or make it small and let there be ripple. I was thinking of letting the boost up supply have 100khz spikes on it. (needs protection, but that is simple)

I was thinking the base power supply probably needs to voltage limit at 50V or something. And power limit at 1000 watts. (this is a small system)
50V @ 0 to 20A = 0 to 1000W,
25V @ 40A
12V@ 80A
I don't know if we need to power limit or current limit. Either way it needs to survive continuous shorts. I like the idea of power limiting. I see the idea of watching the arc voltage to know when to advance the tooling.

The boost up supply could easy be 100 watt. So 500V @ 2A or 250V@4A or 120V @ 2A. That should start the arc. I think 400 to 500 voltages limit and 100W power limit. As soon as the arc starts it will current limit the boost up supply and cause it to shut down. Then the arc will cause the main supply to current or power limit.

The other Ron

All of the stuff like this I worked with was normally current limited. Arc voltage were set and maintained by AVC (Arc Voltage Control) which was just a matter of maintaining a gap as material was removed but to my memory it was all about current control. Like I said, I could have done without much of that stuff during my career.

Ron

 

[shortbus=]

Reload(good name),
If some people are using RF to start the arc then remove C2 or make it small and let there be ripple. I was thinking of letting the boost up supply have 100khz spikes on it. (needs protection, but that is simple)

I was thinking the base power supply probably needs to voltage limit at 50V or something. And power limit at 1000 watts. (this is a small system)
50V @ 0 to 20A = 0 to 1000W,
25V @ 40A
12V@ 80A
I don't know if we need to power limit or current limit. Either way it needs to survive continuous shorts. I like the idea of power limiting. I see the idea of watching the arc voltage to know when to advance the tooling.

The boost up supply could easy be 100 watt. So 500V @ 2A or 250V@4A or 120V @ 2A. That should start the arc. I think 400 to 500 voltages limit and 100W power limit. As soon as the arc starts it will current limit the boost up supply and cause it to shut down. Then the arc will cause the main supply to current or power limit.

The other Ron

Now we have two 'other Ron's'

The problem with using too high of a voltage for the spike voltage is that there is a chance that the main voltage wouldn't keep conducting after the spike quits. The breakdown voltage of the oil is ~170VDC at 0.001 inch. This give a predictable undersize to make the electrode. If the spike voltage is higher the gap would be bigger and then the problem of missed sparks would crop up again.

 

[shortbus=]

Now I understand what you are getting at with inductors. This is a good read on the subject and it is kept simple with nice little animations.

My friend the EDM machine, such fond memories. I was not a weld engineer, never even played one on TV. However, there was a time during my career when I reluctantly became involved with nuclear welding systems and EDM machines. Been retired two years and it was around 15 years before I retired so it was some time back. We had a large row of 5 or 6 CNC machines called Okuma Street (named after the machines). Every now and then a tool would break off in a part being machined. This is a ***** when the part is at the $25K plus point in the machining process. So how to remove a broken tap or drill? Place the part on an EDM machine and let the machine borough down into the broken tool. Slow but steady and bubble by bubble the snapped tool material was washed away until the broken tool would collapse on itself. I did redesign an old 60s vintage EDM machine which used Thyratron tubes to run on an updated solid state power supply. That was long ago but I was forever sucked into weld department issues after that. EDM and TIG Weldment are just a few things that fascinated me over the years.

Youngstown area huh? My future daughter-in-law is a Youngstown girl.

Ron

I ran EDM for ~13 years and became fascinated with the process. Most of the time was making plastic molding molds, but saved many other parts for guy's like you were talking about. Taps and such never seem to break at a good time.

The nuclear welding, did you work for B&W? Did a lot of work for their R&D lab in my apprenticeship.

Some of the older machines when I first worked with EDM still had power supplies with tubes and big selenium rectifiers in them.

 

[shortbus=]

Ok, would you guy's look at this? The schematic leaves out some of the stuff like voltage dividers, gate drivers, and etc. But it's the basics, and want to know if I'm on the right track. Thank you again for the help!



Sorry for the thumbnail, but when I tried to do it fullsize only part of it showed on screen.

 

[Tony Stewart]

I have done spark plug testing in clean air as well as clean and lightly contaminated transformer oil.
I haven't done EDM but this may help. Are you designing from scratch , fixing a unit or trying to improve it?

EDM's can discharge from 10A to 250A depending on application, I think. YOu don't need much power to initiate the arc just enough to charge the gap capacitance load and reach breakdown threshold, which will reduce as the oil gets contaminated. If 210V is all you want then this will limit your gap size in future. But I would go for 1kV with a small coil around probe and size the ferrite and coil inductance to match impedance of electrode resistance and gap capacitance in oil for effective use and lower resonance effects during the arc. There will always be resonance after the arc extinguishes below the holding current. But I dont know how much that drops your 100V supply between arcs.

I used a typical 1.5mm gap titanium fine tipped auto spark plug and which arc'd at 3kV +/-0.1 in air and 21kV in clean mineral oil. (for huge power transformers)
Thus 2kV/mm in air and 42kV/mm in clean oil. 25kV/mm is the min. standard for clean oil. T
his can easily drop to 50% with the slightest amount of contamination and 10% of this level with heavy contamination.

So if you want to guarantee the arc for first use in clean use 42kV/mm to start in clean oil
or with 0.001" gap 42kV/mm*25.4mm/" = 1kV or maybe 500V.

What matters is the extinction voltage or Holding current (just like SCR's) can vary with contamination, meaning the level of discharge between sparks from your 100V main supply.
This I do not know.

So conceptually you only need a simple inductor diode or spark to initiate ionization and the follow-on current is determined by the sum of the ESR of your 100V caps, the carbon electrode resistance and the gap ionization resistance.

With sufficiently large Q=CV , the ionization resistance should be low as the caps and thus the carbon electrode dissipates most of the power for consistent welds but is cooled by the oil.

Depending on the energy levels E= 1/2CV² the oil will breakdown and produce ;
Hydrogen, Methane, Ethane, Ethylene, Acetylene in this order according to heat and energy levels. Any oxygen dissolved in the oil will of course make it explosive which is normally < 5000 ppm but depends on exposure to air and cellulose material. 1~10W discharges will probably just generate Hydrogen that gets dissolved in the oil. Maybe 500W discharges energy bigger energy larger combustible gas molecules. The oil detonates with the metal, then extinguishes quickly so it may be silent or quite loud.

DGA or dissolved gas analysis costs around $75 to get results from companies like Weidmann. Oil circulation is important for cooling and dispersing gas that dissolves in the oil as material gets eroded and vaporizes and solidifies in suspension.

If you can find out the extinction voltage which is determined by the Holding Current and the gap length.
The energy transfer between discharges and thus change your Cap size with switches and gap size ( series capacitance) , by the change in E= 1/2C(Vi²-Vf²)

Then you can determine how much current is needed for the 100V supply.
The 1kV spike have enough current to charge up the dielectric oil between electrode and plate. You can measure this or guess it is around 1000 pF
The flyback inductor current must be greater than the dielectric current to reach 1kV

here is a design to consider http://bee.mif.pg.gda.pl/ciasteczkowypotwor/Racal/9470-9479.pdf then reconsider your requirements.