(Flyback Driver)
Once again the circuit I use is almost a solid state Tesla coil; it is
flyback transformer from an old television
set driven at high frequency. The schematic I currently use is
shown below. The
technical
name is a Hartley Oscillator configuration (sort of)
of a Line
OutPut Transformer (flyback.)
Circuit
Rundown:
The circuit is basically 120 volts AC in a center tapped 25.2 volt step down transformer (which will die if you draw too much current from it), into a full wave bridge rectifier (using the center tap and one hot of the transformer), and then a large filter capacitor, which gives me over 16 volts DC at just over 2 amperes in the primary side of the circuit WITH NO LOAD. The capacitor is necessary as it charges during the peak of the AC cycle, and releases during the trough, smoothing things out nicely. A large filter cap makes a huge difference in performance (see "filter cap notes" below). The resisters are a voltage divider providing a lot of power to the primary winding, and a little base current for the feedback winding and transistor. The transistor switches the primary coil on and off at resonant frequencies to create high frequency input.
Circuit Operation:
The DC voltage charges the primary coil and consequently magnetizes the ferrite core of the transformer, which induces a reverse current in the "feedback" coil, which obviously goes against the current passing through it from the voltage divider canceling it out, which turns off the transistor. The EM field in the ferrite core then collapses with no primary current to support it. After this happens, the transistor begins to conduct again because there is no longer any current in the feedback winding. So our DC input is rapidly switching on and off. This induces current in the secondary, first one direction (primary conducting) and then the other (primary not conducting (basic laws of inductors, as the EM field collapses, it creates a current rush in the other direction)). This all happens at a high frequency like 15-40 kHz, and the output is usually around 10-25kV depending on your flyback and component choices. You can see some oscilloscope graphs below of it's operation. The first picture is during unloaded operation, the second is when a hand is placed on a globe. The yellow is a voltage reading taken from the top of the primary coil, and the blue is taken from the ground. I'm not sure if I hooked up the probes in the correct places for proper voltage measurements, but this experiment did show the frequency change quite clearly, which can be read in the bottom right hand corner of each picture. I believe this is showing that the average voltage is 6.27V so the peak should be around 13, which means 4 volts is being burned someplace (16V supply remember), Part of it is gone because the cap is constantly draining with a load and therefore it is only really maintaining about 15V and the other V is being burned by the transistor. Ignore the 32V Pk-Pk, I don't know what it means, that is what the graph is displaying and may be caused by resonant rise, but I do not know. I also do not know what the hiccup is on the yellow line but it occurs right when the transistor starts decelerating the voltage increase.
Circuit I:
Here is a picture of my first system. It was powered by 9-volt batteries or a variable dc power supply from radio shack. This flyback did not work for a plasma globe because it's output was rectified which means it's output was DC, which won't work for a plasma ball. Remember, the light emitted by the plasma is caused by electrons shifting orbits in the atoms, it is when they move that they emit light. If the voltage flows in the same direction continually, the glass will build up a charge until current is zero (like a capacitor charging,) and no light will be emitted. When I tested this circuit on my professional globe a very interesting effect was observed, it looked like a pink aurora that spread across the glass, and moved away when one blew on it. This was due to particles only being able to flip charges on the edge of the "cloud", blowing on it pushed the charged particles away so they discharged, and new charges could develop. It was very cool, but to dim to see with the lights on, so I dropped the rectified flyback (you can use them if you remove the diodes, but it isn't easy!).
Circuit II:
Here is my new working but quirky system. I finally got around to removing the flyback from the dead console in the garage and after winding my own primaries it works great!! (despite the unsightly rust spots, and great skepticism on my part!) This flyback was meant to be used with a cascade, but since we want AC we cannot do that (see "flyback notes" below). This means we are limited to about 8-10kV output.
The circuit is basically 120 volts AC in a center tapped 25.2 volt step down transformer (which will die if you draw too much current from it), into a full wave bridge rectifier (using the center tap and one hot of the transformer), and then a large filter capacitor, which gives me over 16 volts DC at just over 2 amperes in the primary side of the circuit WITH NO LOAD. The capacitor is necessary as it charges during the peak of the AC cycle, and releases during the trough, smoothing things out nicely. A large filter cap makes a huge difference in performance (see "filter cap notes" below). The resisters are a voltage divider providing a lot of power to the primary winding, and a little base current for the feedback winding and transistor. The transistor switches the primary coil on and off at resonant frequencies to create high frequency input.
Circuit Operation:
The DC voltage charges the primary coil and consequently magnetizes the ferrite core of the transformer, which induces a reverse current in the "feedback" coil, which obviously goes against the current passing through it from the voltage divider canceling it out, which turns off the transistor. The EM field in the ferrite core then collapses with no primary current to support it. After this happens, the transistor begins to conduct again because there is no longer any current in the feedback winding. So our DC input is rapidly switching on and off. This induces current in the secondary, first one direction (primary conducting) and then the other (primary not conducting (basic laws of inductors, as the EM field collapses, it creates a current rush in the other direction)). This all happens at a high frequency like 15-40 kHz, and the output is usually around 10-25kV depending on your flyback and component choices. You can see some oscilloscope graphs below of it's operation. The first picture is during unloaded operation, the second is when a hand is placed on a globe. The yellow is a voltage reading taken from the top of the primary coil, and the blue is taken from the ground. I'm not sure if I hooked up the probes in the correct places for proper voltage measurements, but this experiment did show the frequency change quite clearly, which can be read in the bottom right hand corner of each picture. I believe this is showing that the average voltage is 6.27V so the peak should be around 13, which means 4 volts is being burned someplace (16V supply remember), Part of it is gone because the cap is constantly draining with a load and therefore it is only really maintaining about 15V and the other V is being burned by the transistor. Ignore the 32V Pk-Pk, I don't know what it means, that is what the graph is displaying and may be caused by resonant rise, but I do not know. I also do not know what the hiccup is on the yellow line but it occurs right when the transistor starts decelerating the voltage increase.
Circuit I:
Here is a picture of my first system. It was powered by 9-volt batteries or a variable dc power supply from radio shack. This flyback did not work for a plasma globe because it's output was rectified which means it's output was DC, which won't work for a plasma ball. Remember, the light emitted by the plasma is caused by electrons shifting orbits in the atoms, it is when they move that they emit light. If the voltage flows in the same direction continually, the glass will build up a charge until current is zero (like a capacitor charging,) and no light will be emitted. When I tested this circuit on my professional globe a very interesting effect was observed, it looked like a pink aurora that spread across the glass, and moved away when one blew on it. This was due to particles only being able to flip charges on the edge of the "cloud", blowing on it pushed the charged particles away so they discharged, and new charges could develop. It was very cool, but to dim to see with the lights on, so I dropped the rectified flyback (you can use them if you remove the diodes, but it isn't easy!).
Circuit II:
Here is my new working but quirky system. I finally got around to removing the flyback from the dead console in the garage and after winding my own primaries it works great!! (despite the unsightly rust spots, and great skepticism on my part!) This flyback was meant to be used with a cascade, but since we want AC we cannot do that (see "flyback notes" below). This means we are limited to about 8-10kV output.
Below is
my
bridge rectifier, and
step down transformer. A while back, during a test of my
globe,
the transformer died (for reference when I say transformer I mean the
step down transformer, and when I am talking about the flyback, I will
say flyback), I fried a few actually, they seem to have a
tendency to burn
out, as I had a tendency to draw to much current from them, since this
picture was taken, I have installed a 2 amp fuse, and reduced the
number of transistors to one (see notes below about paralleling
transistors,) and run it off the center tap ground instead of
the other secondary (see the black wire). This gave me 12.6
volts,
which only works if you have a huge filter cap. Not shown is the
large
heat sink I put on the transistor (I tried several, see "Transistor
choice" below), which is necessary to keep them from frying.
Circuit
III:
I will have a picture soon, but not much has changed, so you'll have to wait. I switched to the original primary (actually I switched flybacks, and used it's original primary, and I am in the process of converting the old unit). I added a large filter capacitor (45,000 uf). I switched to the TIP 3055 transistor, and I like it. Operation is way better, even running of the 12V input.
I will have a picture soon, but not much has changed, so you'll have to wait. I switched to the original primary (actually I switched flybacks, and used it's original primary, and I am in the process of converting the old unit). I added a large filter capacitor (45,000 uf). I switched to the TIP 3055 transistor, and I like it. Operation is way better, even running of the 12V input.
Twin
Flyback:
Just for kicks I built a twin flyback supply, which uses two identical flybacks with 1 feedback winding, and the primaries wired in series (opposite polarity), and the secondary returns connected together and grounded. The high voltage "outs" arc together, and can really light up a fluorescent light bulb quite well. Not really any good for a plasma globe, but I thought it was cool nonetheless.
Arcs:
And here is an arc generated by the power supply to a ground wire (I apologize for the poor image quality, small arcs in dark rooms are hard to photograph, but just for reference that arc is over a 1 inch gap!). This picture was taken with an overstressed transformer and a small filter cap, my arcs are now thinner, but I could easily fix that by running 25 volts input, which really is much nicer for a plasma globe (I'm going to have to order some 4 amp transformers.)
Just for kicks I built a twin flyback supply, which uses two identical flybacks with 1 feedback winding, and the primaries wired in series (opposite polarity), and the secondary returns connected together and grounded. The high voltage "outs" arc together, and can really light up a fluorescent light bulb quite well. Not really any good for a plasma globe, but I thought it was cool nonetheless.
Arcs:
And here is an arc generated by the power supply to a ground wire (I apologize for the poor image quality, small arcs in dark rooms are hard to photograph, but just for reference that arc is over a 1 inch gap!). This picture was taken with an overstressed transformer and a small filter cap, my arcs are now thinner, but I could easily fix that by running 25 volts input, which really is much nicer for a plasma globe (I'm going to have to order some 4 amp transformers.)
Flyback
Notes:
Modern flyback are difficult to use, as you need to literally cut them open and remove or bypass the diodes inside of them. If you are simply looking for high voltage, not for a plasma globe, then they are fine as is, with some circuit modifications (something to step up the 12V to about 100 V, before feeding it into the existing primary, or winding your own primary which is much easier though less efficient). Older, not rectified flybacks are better, but we still have two kinds, those used with a cascade and those not. Only the really old ones did not use a cascade (think vacuum tube televisions). The flybacks I have collected over time (5 of them) all were designed to be used with a cascade (or "silicon tripler" as it was called). They work good, but the older ones are better as they put out a higher voltage.
Power Input Notes:
This circuit runs best on 12 volts input. The experts say it is not good to run higher voltages than that, as it can damage the flyback, I say different. While further testing is needed, I believe if you double the number of turns in the primary, you can safely double the input voltage, as the secondary voltage turns ratio will not be messed up. This will allow you to draw more current, and give a more powerful system. Sounds pretty good huh? Maybe not, I need to test it to be sure, but there is a possibility that the increased impedance of the extra turns primary will limit incoming current, and you will be stuck with the same wattage output. Like I said, I have to try it to be sure. Check back in a few weeks and I might have it done. Just know that 12 volts will nicely drive a 12-in plasma globe with a metal electrode, a glass one needs a bit more.
Filter Cap Notes:
The filter cap is a very important part that I overlooked in the past. The circuit will not run without one, at all. I used a small 4,000 uf filter cap initially, and it caused the circuit to run very poorly, and draw a tremendous amount of current (5 amps plus!) I had to parallel five transistors to get the power out of it that I do now that I have a 45,000 uf filter cap, with one transistor. Make sure it is rated for 20+ volts if you are going to use a 12 volt supply (if you can go for 35+ volts, as this allows you to upgrade to 25 volt input power for when you need to, but remember you will draw about 4 amps of current with that voltage, which will kill your transformer if you run it too long, so be careful, or get a better transformer).
Flyback Diode:
One component not in my diagram above is the flyback diode. No I am not talking about the diodes that rectify the output, I am talking about a fast switching diode that is placed across the primary coil. Why would one put a diode there? I'll tell you! When the electromagnetic field set up in the ferrite core collapses, it not only induces a reverse voltage in the Secondary coil, but also in the primary. This voltage does nothing because the transistor will only conduct one way, but that pull is bad for the transistor and will cause it to heat up. A diode allows a return path for this induced primary voltage, taking the load off the transistor. In theory this is great, but I've tried using them and my performance died, I'm not sure why. Some professionals put a capacitor across the primary and this is said to improve circuit operation, I've tried that as well and results were not spectacular. The idea makes sense since the cap will charge when the transistor is on, and when it is off it will discharge through the inductor primary coil, giving the system some push and allowing back EMFs to go someplace. I need tom monitor current consumption and see if there is improvement to be gained there, because as far as output is concerned, I got nothing extra using a cap.
General Construction Notes:
Please note, If you attempt to build on of these circuits, do yourself a huge honkin' favor. Buy a variable, not regulated, power supply from radio shack (1.5-12V, 300mA). Use this as a power supply until you get the circuit working, then put in your homemade jobby (step down transformer, rectifier, and filter cap). Using the radio shack power supply will allow you to get everything set up correctly, as when you turn it on, and something isn't right, it will not run, but nothing will be damaged. I have blown out a lot of parts hooking something up wrong and powering it up, you could have everything wrong and it still won't die if you are using that weak power supply. Also, the "Alternative power supply" in the schematic above is much simpler to use, and it seems to give good results, and it works well with Radio Shack's 25.2V-2Amp step down transformer, I haven't burned one up with this configuration. Just point your diodes stripe side toward circuit positive, blank side towards transformer. Note however, a three prong plug is always recommended, use the green wire for everything grounded for best performance, and to ensure touching your components does not shock you.
PC Board:
I am attempting to make a printed circuit board for my future drivers, but have not started on that yet. Maybe over the summer.
Transistor Choices:
A note on transistors, if you plan on paralleling them you had best put an equalizing resistor on their outputs for best operation, otherwise bad things happen like excessive heat and only nominal performance increase. The main reason to parallel transistors is to share current, not get extra amplification, if that is your goal, your gonna have hot transistors. Think about it, the transistors turn on from base current, if two transistors are in parallel, each one only turns on half way, so they dissipates twice the power and get hot. Anyway I have experimented with a bunch of transistors, and I will list my results below.
1. The MJ15015, Very quiet, but seems a bit delicate, medium output power (note, I did not try it with a proper filter cap, and my transistors were about 30 years old, so newer ones may work better, I'll have to try them again).
2. The 2N3055, Very quiet until you load it down, then some audible noise is made, some are very hearty, some are not (don't buy them from Radio Shack, their quality varies too much) Pretty powerful, gets hot medium speed. A decent choice, not the best.
3. The NPN NTE284, Loud and annoying, very hearty, very powerful (it makes poor plasma though because it switches too slow, but once again I did not try it with my better filter cap,) not my first choice.
4. The TIP 120 Darlington, Also loud and annoying, pretty powerful, pretty hearty, I need to test it with a properly filtered supply to decide if I like it or not. (note it may be possible to construct a much more efficient driver using a Darlington, as they have obseen gain, so the voltage divider could be built to pass only a tiny current, and the feedback winding designed accordingly, I need to try it, slow switching speeds may be a problem however.)
5. The MJE 13007, Very silent, not overly powerful, somewhat hearty. I like the TIP case, but the 2N3055 is a better choice for performance.
6. The TIP3055, Silent, more powerful than the 2N3055 but runs much cooler, very hearty, TIP case. My current favorite choice. pretty sensitive to over voltages.
7. The KSE13009, seems about the same as the TIP3055, awesome transistor, but not available locally (not that I get my parts from radio shack anymore anyway.)
8. The 2N3773, gets Hot quick, very powerful. Awesome transistor with a fanned heat sink.
9. The TIP31, Super transistor, runs fast and hot, DON'T USE MORE THAN 12V, it will die instantly! Lower output voltage (slower switching speed?) but really high current.
10. The BUF420M, Some professionals use this transistor on globes up to 22-inches wide, so it must be good, I wasn't very impressed but my circuit was not tuned well for such low gain, virtually indestructible however.
Tuning and Observations:
This is a section that is left out of virtually every web page I have looked at about these circuits, how to optimize them; they usually say something like "adjust resistor values for your particular transistor and flyback." What they heck does that mean? Adjust how? Well, I don't know (sort of, I have no math to help you, no formulas, just observations and I will list them now.) For a higher output voltage, use fewer primary turns (duh) the problem with this is that with only 2 or so turns your feedback winding has trouble generating enough current under load, and your output severely drops when you touch the glass. More primary turns gives more feedback current at a lower voltage and hence more current in your discharges, but substantially lower output voltage to the point of being useless with a plasma globe. I generally use the flybacks included primary which usually has 2 turns, and my best circuit has a 1 turn feedback winding. One thing you can do to combat that stalling under load associated with fewer turns is to use a smaller resistor in the voltage divider, like 14 ohms instead of 22, this seems to give me better performance no matter what I am doing. The current passed by the transistor is directly related to the transistors gain, and the base current (base current * gain = output current) so to ensure full "on" of the circuit you need a base current that will achieve this, hence low value resistors as low as 14 and up to 500 ohms in the divider) I've tried a stepped up ratio 1k and 100 ohms and it worked but output suffered under load really badly, I will try a Darlington and see if I can get away with it because a lot of power is uselessly burned up in that divider, and less current admitted means a more efficient driver. Because of this whole gain thing, excessive heating occurs when you parallel transistors, a process which I gave up on for the most part, they never turn on all the way because the current is shared and the fastest one wins, when comparing current draw of the circuit overall it just doesn't make that much difference anyway, you need equalizing resistors to make it work, or like thirty transistors and hope a few match. One thing I am in the process of experimenting with is using this circuit to drive a comparator op amp to drive a mosfet, this will always ensure full on, and maximum power transfer, with minimum heat wasted during switching. I'm not sure if it will work initial results are very poor but I am likely wiring something wrong, I need to try again. That would be the ultimate circuit, minimal switching losses, and always at resonance, sweet.
The
best thing I can tell you right now is when designing a
circuit, go for a transistor with the smallest possible saturation
voltage, the saturation voltage is the volts burned in the transistor
when it is full on, so a smaller
saturation voltage means the transistor will run cooler. A higher
gain helps also, but may not be necessary.
Also design your voltage divider to have a low output voltage
and a high current, so use low value resistors. Try to match the
voltage output of the divider with the output of the feedback winding
(so if your voltage divider puts out .5V with a 500 and a 14 ohm
resistors, try to design your windings to put out that voltage also, so
the current can be effectively canceled, and not shunted) this needs
further exploration however, as it is not possible with a two winding
primary; it is possible double the voltage would be best, I'm still not
sure.
Varieties of Plasma Globe Circuits:
There are several preferred methods for constructing plasma ball power supplies, mine is one of the more popular types among hobbyist, but not necessarily the best for raw power.
1. Hartley Oscillator, It is true the Hartley Oscillator runs at resonance regardless of tuning so it is the simplest method, but sometimes you do not want resonant frequency for certain plasma effects. Nevertheless, it is in my humble opinion the best driver for beginners, and I think it makes some nice plasma effects, which is why I use it. It is not the most powerful and a lot of power is wasted in the divider, so I'm going to experiment with Darlington transistors to see if I can fix that. Or perhaps I can isolate the oscillator somehow from the power supply of the transistor, that would be nice.
2. Square Wave Generator, What a lot of people use, is a square wave generator, like a 555 integrated circuit chip, which makes for a more complicated circuit, and they sometimes tend to whine loudly, which is annoying. Also when the load is changed (like when you touch the ball) the circuit is no longer in resonance, which you may have so carefully tuned it to be in, and output significantly drops. Basically it works the same way as system with a feedback winding but the timer is set to oscillate at a fixed frequency, which is what trips the transistor. The only advantages I can think of for this circuit is you can use an automotive ignition coil instead of a flyback (but contrary to popular belief, you can use a Hartley Oscillator for this as well, I'll explain in a moment) and you can (on a well designed system, with variable resistors) you can change the frequency to get different plasma effects. A popular variation is the use of a MOSFET instead of a transistor. A MOSFET is a special transistor, that is designed to handle high current, and is switched on my voltage rather than current (so the gain is volts in, current out, or mhos, yes mhos, which is ohms spelled backwards, don't you love engineering?) I can't say much about it, except the systems I have seen are usually somewhat complicated (I believe these circuits are usually designed to have a variable frequency and duty cycle output, which is very good, because sometimes your transformer resonance output is not ideal for your gas combination, as I just said.) Duty cycle is the amount of time each cycle that the signal generator is generating voltage, the rest of the time the output is 0 volts.
3. ZVS driver, You have probably seen these before if you looked up flyback driver on Google, they operate in a push pull configuration on a tiny tank circuit (sort of.) Here is a schematic.
Other Random Notes, That are Noteworthy:
Now hear this, modern flyback transformers must have the high voltage diodes in them bypassed to work correctly, and the intended voltage input is usually high (around 100-150 volts, which is how they make them so small) so to rig it to work one must (actually should, not really must) place an intermediate transformer between the flybacks original primary winding (the highest resistance one in the coil, excluding the secondary of course), and your hand wound primaries. Simply wind a bunch of turns on your secondary transformer, and several feed back turns (as if it were the flyback in my above schematic, except more turns) then wind many turns (10/1 is fine) from another wire, and connect it directly to the flybacks original primary coil. This will however cause the coil to operate at the intermediate transformers resonance, not the flyback's. You could simply wind the feedback winding on the flyback core, this works, but may not be ideal; more experimentation is necessary. I tried rigging up an intermediate transformer and results were very poor, but it may have been poorly designed. This system can also be used to drive an ignition coil with a Hartley Oscillator, I will try this soon to see what kind of results I get, just for kicks. By the way, if you need a flyback, e-mail me and I'll tell you where you can buy one that will work well for a plasma globe.
Modern flyback are difficult to use, as you need to literally cut them open and remove or bypass the diodes inside of them. If you are simply looking for high voltage, not for a plasma globe, then they are fine as is, with some circuit modifications (something to step up the 12V to about 100 V, before feeding it into the existing primary, or winding your own primary which is much easier though less efficient). Older, not rectified flybacks are better, but we still have two kinds, those used with a cascade and those not. Only the really old ones did not use a cascade (think vacuum tube televisions). The flybacks I have collected over time (5 of them) all were designed to be used with a cascade (or "silicon tripler" as it was called). They work good, but the older ones are better as they put out a higher voltage.
Power Input Notes:
This circuit runs best on 12 volts input. The experts say it is not good to run higher voltages than that, as it can damage the flyback, I say different. While further testing is needed, I believe if you double the number of turns in the primary, you can safely double the input voltage, as the secondary voltage turns ratio will not be messed up. This will allow you to draw more current, and give a more powerful system. Sounds pretty good huh? Maybe not, I need to test it to be sure, but there is a possibility that the increased impedance of the extra turns primary will limit incoming current, and you will be stuck with the same wattage output. Like I said, I have to try it to be sure. Check back in a few weeks and I might have it done. Just know that 12 volts will nicely drive a 12-in plasma globe with a metal electrode, a glass one needs a bit more.
Filter Cap Notes:
The filter cap is a very important part that I overlooked in the past. The circuit will not run without one, at all. I used a small 4,000 uf filter cap initially, and it caused the circuit to run very poorly, and draw a tremendous amount of current (5 amps plus!) I had to parallel five transistors to get the power out of it that I do now that I have a 45,000 uf filter cap, with one transistor. Make sure it is rated for 20+ volts if you are going to use a 12 volt supply (if you can go for 35+ volts, as this allows you to upgrade to 25 volt input power for when you need to, but remember you will draw about 4 amps of current with that voltage, which will kill your transformer if you run it too long, so be careful, or get a better transformer).
Flyback Diode:
One component not in my diagram above is the flyback diode. No I am not talking about the diodes that rectify the output, I am talking about a fast switching diode that is placed across the primary coil. Why would one put a diode there? I'll tell you! When the electromagnetic field set up in the ferrite core collapses, it not only induces a reverse voltage in the Secondary coil, but also in the primary. This voltage does nothing because the transistor will only conduct one way, but that pull is bad for the transistor and will cause it to heat up. A diode allows a return path for this induced primary voltage, taking the load off the transistor. In theory this is great, but I've tried using them and my performance died, I'm not sure why. Some professionals put a capacitor across the primary and this is said to improve circuit operation, I've tried that as well and results were not spectacular. The idea makes sense since the cap will charge when the transistor is on, and when it is off it will discharge through the inductor primary coil, giving the system some push and allowing back EMFs to go someplace. I need tom monitor current consumption and see if there is improvement to be gained there, because as far as output is concerned, I got nothing extra using a cap.
General Construction Notes:
Please note, If you attempt to build on of these circuits, do yourself a huge honkin' favor. Buy a variable, not regulated, power supply from radio shack (1.5-12V, 300mA). Use this as a power supply until you get the circuit working, then put in your homemade jobby (step down transformer, rectifier, and filter cap). Using the radio shack power supply will allow you to get everything set up correctly, as when you turn it on, and something isn't right, it will not run, but nothing will be damaged. I have blown out a lot of parts hooking something up wrong and powering it up, you could have everything wrong and it still won't die if you are using that weak power supply. Also, the "Alternative power supply" in the schematic above is much simpler to use, and it seems to give good results, and it works well with Radio Shack's 25.2V-2Amp step down transformer, I haven't burned one up with this configuration. Just point your diodes stripe side toward circuit positive, blank side towards transformer. Note however, a three prong plug is always recommended, use the green wire for everything grounded for best performance, and to ensure touching your components does not shock you.
PC Board:
I am attempting to make a printed circuit board for my future drivers, but have not started on that yet. Maybe over the summer.
Transistor Choices:
A note on transistors, if you plan on paralleling them you had best put an equalizing resistor on their outputs for best operation, otherwise bad things happen like excessive heat and only nominal performance increase. The main reason to parallel transistors is to share current, not get extra amplification, if that is your goal, your gonna have hot transistors. Think about it, the transistors turn on from base current, if two transistors are in parallel, each one only turns on half way, so they dissipates twice the power and get hot. Anyway I have experimented with a bunch of transistors, and I will list my results below.
1. The MJ15015, Very quiet, but seems a bit delicate, medium output power (note, I did not try it with a proper filter cap, and my transistors were about 30 years old, so newer ones may work better, I'll have to try them again).
2. The 2N3055, Very quiet until you load it down, then some audible noise is made, some are very hearty, some are not (don't buy them from Radio Shack, their quality varies too much) Pretty powerful, gets hot medium speed. A decent choice, not the best.
3. The NPN NTE284, Loud and annoying, very hearty, very powerful (it makes poor plasma though because it switches too slow, but once again I did not try it with my better filter cap,) not my first choice.
4. The TIP 120 Darlington, Also loud and annoying, pretty powerful, pretty hearty, I need to test it with a properly filtered supply to decide if I like it or not. (note it may be possible to construct a much more efficient driver using a Darlington, as they have obseen gain, so the voltage divider could be built to pass only a tiny current, and the feedback winding designed accordingly, I need to try it, slow switching speeds may be a problem however.)
5. The MJE 13007, Very silent, not overly powerful, somewhat hearty. I like the TIP case, but the 2N3055 is a better choice for performance.
6. The TIP3055, Silent, more powerful than the 2N3055 but runs much cooler, very hearty, TIP case. My current favorite choice. pretty sensitive to over voltages.
7. The KSE13009, seems about the same as the TIP3055, awesome transistor, but not available locally (not that I get my parts from radio shack anymore anyway.)
8. The 2N3773, gets Hot quick, very powerful. Awesome transistor with a fanned heat sink.
9. The TIP31, Super transistor, runs fast and hot, DON'T USE MORE THAN 12V, it will die instantly! Lower output voltage (slower switching speed?) but really high current.
10. The BUF420M, Some professionals use this transistor on globes up to 22-inches wide, so it must be good, I wasn't very impressed but my circuit was not tuned well for such low gain, virtually indestructible however.
Tuning and Observations:
This is a section that is left out of virtually every web page I have looked at about these circuits, how to optimize them; they usually say something like "adjust resistor values for your particular transistor and flyback." What they heck does that mean? Adjust how? Well, I don't know (sort of, I have no math to help you, no formulas, just observations and I will list them now.) For a higher output voltage, use fewer primary turns (duh) the problem with this is that with only 2 or so turns your feedback winding has trouble generating enough current under load, and your output severely drops when you touch the glass. More primary turns gives more feedback current at a lower voltage and hence more current in your discharges, but substantially lower output voltage to the point of being useless with a plasma globe. I generally use the flybacks included primary which usually has 2 turns, and my best circuit has a 1 turn feedback winding. One thing you can do to combat that stalling under load associated with fewer turns is to use a smaller resistor in the voltage divider, like 14 ohms instead of 22, this seems to give me better performance no matter what I am doing. The current passed by the transistor is directly related to the transistors gain, and the base current (base current * gain = output current) so to ensure full "on" of the circuit you need a base current that will achieve this, hence low value resistors as low as 14 and up to 500 ohms in the divider) I've tried a stepped up ratio 1k and 100 ohms and it worked but output suffered under load really badly, I will try a Darlington and see if I can get away with it because a lot of power is uselessly burned up in that divider, and less current admitted means a more efficient driver. Because of this whole gain thing, excessive heating occurs when you parallel transistors, a process which I gave up on for the most part, they never turn on all the way because the current is shared and the fastest one wins, when comparing current draw of the circuit overall it just doesn't make that much difference anyway, you need equalizing resistors to make it work, or like thirty transistors and hope a few match. One thing I am in the process of experimenting with is using this circuit to drive a comparator op amp to drive a mosfet, this will always ensure full on, and maximum power transfer, with minimum heat wasted during switching. I'm not sure if it will work initial results are very poor but I am likely wiring something wrong, I need to try again. That would be the ultimate circuit, minimal switching losses, and always at resonance, sweet.
Varieties of Plasma Globe Circuits:
There are several preferred methods for constructing plasma ball power supplies, mine is one of the more popular types among hobbyist, but not necessarily the best for raw power.
1. Hartley Oscillator, It is true the Hartley Oscillator runs at resonance regardless of tuning so it is the simplest method, but sometimes you do not want resonant frequency for certain plasma effects. Nevertheless, it is in my humble opinion the best driver for beginners, and I think it makes some nice plasma effects, which is why I use it. It is not the most powerful and a lot of power is wasted in the divider, so I'm going to experiment with Darlington transistors to see if I can fix that. Or perhaps I can isolate the oscillator somehow from the power supply of the transistor, that would be nice.
2. Square Wave Generator, What a lot of people use, is a square wave generator, like a 555 integrated circuit chip, which makes for a more complicated circuit, and they sometimes tend to whine loudly, which is annoying. Also when the load is changed (like when you touch the ball) the circuit is no longer in resonance, which you may have so carefully tuned it to be in, and output significantly drops. Basically it works the same way as system with a feedback winding but the timer is set to oscillate at a fixed frequency, which is what trips the transistor. The only advantages I can think of for this circuit is you can use an automotive ignition coil instead of a flyback (but contrary to popular belief, you can use a Hartley Oscillator for this as well, I'll explain in a moment) and you can (on a well designed system, with variable resistors) you can change the frequency to get different plasma effects. A popular variation is the use of a MOSFET instead of a transistor. A MOSFET is a special transistor, that is designed to handle high current, and is switched on my voltage rather than current (so the gain is volts in, current out, or mhos, yes mhos, which is ohms spelled backwards, don't you love engineering?) I can't say much about it, except the systems I have seen are usually somewhat complicated (I believe these circuits are usually designed to have a variable frequency and duty cycle output, which is very good, because sometimes your transformer resonance output is not ideal for your gas combination, as I just said.) Duty cycle is the amount of time each cycle that the signal generator is generating voltage, the rest of the time the output is 0 volts.
3. ZVS driver, You have probably seen these before if you looked up flyback driver on Google, they operate in a push pull configuration on a tiny tank circuit (sort of.) Here is a schematic.
As far as power output goes, this
is the one to beat, these puppies routinely put out 100W (whereas my 12
driver does 2A*12V=24W.) And given there is only like 6 parts,
this is
a tough one to beat, I'll be honest with you though, I have no idea how
this works, but here is my best theory. This is obviously a push
pull
driver, the 470 and 10K resistors along with the 12V zenar diodes make
a 12V regulated supply to turn on the MOSFETs. The inductor
stores energy to increase your supply voltage to double the input, and
the cap resonates with the primary coil. So if I use the above
abstractions, I can see how it turns on, but how it turns off is a
little more difficult, my current theory is that once it is on, the
fast switching diodes short out the gate drive and the current drops to
zero, so the MOSFETs shut off. I do not understand how you get a
push pull out of this but that is my current theory, I am obviously
missing some details; it likely has something to do with the fact that
the diodes cross to the opposite sides, but I don't see how this gives
any certain resonate frequency, or how the cycle starts up. If
you know please don't hesitate to e-mail me, so I can pass along good
info to my readers.
Other Random Notes, That are Noteworthy:
Now hear this, modern flyback transformers must have the high voltage diodes in them bypassed to work correctly, and the intended voltage input is usually high (around 100-150 volts, which is how they make them so small) so to rig it to work one must (actually should, not really must) place an intermediate transformer between the flybacks original primary winding (the highest resistance one in the coil, excluding the secondary of course), and your hand wound primaries. Simply wind a bunch of turns on your secondary transformer, and several feed back turns (as if it were the flyback in my above schematic, except more turns) then wind many turns (10/1 is fine) from another wire, and connect it directly to the flybacks original primary coil. This will however cause the coil to operate at the intermediate transformers resonance, not the flyback's. You could simply wind the feedback winding on the flyback core, this works, but may not be ideal; more experimentation is necessary. I tried rigging up an intermediate transformer and results were very poor, but it may have been poorly designed. This system can also be used to drive an ignition coil with a Hartley Oscillator, I will try this soon to see what kind of results I get, just for kicks. By the way, if you need a flyback, e-mail me and I'll tell you where you can buy one that will work well for a plasma globe.