Singing Tesla Coil

Discussion in 'The Projects Forum' started by Phil_Jahelka, Jul 12, 2010.

  1. Phil_Jahelka

    Thread Starter New Member

    Jul 12, 2010
    11
    0
    Hey everybody!

    I'm working on a Singing Tesla Coil for a 4-H electricity project and I just finished my first draft of the circuitry. I was wondering if anyone would be kind enough to give me some tips or point out any glaring mistakes I made. Thanks for your help.

    Relevant info: the voltage regulator is a LM317, the Op-Amp is a LM741, the PWM is a UC28025, and the MOSFETs are FDA20N50F's, the 18.5V power source is an old laptop power supply, and the audio is coming from an old Walkman CD player. In addition, I would really like to avoid using GDTs. They're prohibitively expensive and difficult to make. And if I need them, it means I fried the MOSFETs which are the majority of the expense of this project, so I don't really care about cost after those fry.
     
  2. bertus

    Administrator

    Apr 5, 2008
    15,646
    2,345
    Hello,

    Is the BIG CAP connected to the mains?

    We do NOT allow circuits connected to the mains.
    They can harm poeple.


    The H-bridge configuration is not right.

    Bertus
     
    Last edited: Jul 12, 2010
  3. Phil_Jahelka

    Thread Starter New Member

    Jul 12, 2010
    11
    0
    Two Questions: First, how is the H-Bridge messed up? Second, how am I supposed to power a Tesla Coil Primary without using the mains? Also, the cap is used to smooth the DC output of the Bridge Rectifier.
     
  4. bertus

    Administrator

    Apr 5, 2008
    15,646
    2,345
    Hello,

    You should have at least an isolation transformer between the mains and input of the circuit.
    For information on H-bridges take a look at this PDF:
    H-bridges_at_purdue_university

    Bertus
     
  5. Phil_Jahelka

    Thread Starter New Member

    Jul 12, 2010
    11
    0
    I'm sorry, I still don't see the problem. When A goes high, M1 and M3 turn on. When B goes high, M2 and M4 turn on; with dead time between the two, naturally. Also, the only thing I saw in that Purdue thing that I don't have are Schottky's, are those what I need for the H-Bridge to work? Honestly, I didn't think I needed them because the coil runs at ~100kHz, was I wrong?
     
  6. bertus

    Administrator

    Apr 5, 2008
    15,646
    2,345
    Hello,

    Please, change the ciruit showing the transformer.
    Otherwise I have to close the thread.

    The main problem lies in the gate to drain voltage.
    In the following thread there is a discussion between to members about that:
    http://forum.allaboutcircuits.com/showthread.php?t=33615

    You can also have a look at the attached PDF on high side drivers.

    Bertus
     
  7. windoze killa

    AAC Fanatic!

    Feb 23, 2006
    605
    24
    Bertus, glad you got in regarding mains. This one is positively deadly. The whole cct is virtually live.
     
  8. beenthere

    Retired Moderator

    Apr 20, 2004
    15,815
    282
    That is a very true statement. Any circuit running off the mains is a lethal accident waiting to happen. The generated arc will find any conductor (such as people) an attractive path to ground. Any failure in the H bridge will tend to connect the drive circuit to the mains.
     
  9. Phil_Jahelka

    Thread Starter New Member

    Jul 12, 2010
    11
    0
    Ok, here's a new schematic with an isolation transformer. Also, I'm finally understanding why I need drivers. Thanks for the help. Once I find drivers, I'll post a new schematic. Also, how would the audio section do? Did I mess it up as bad as the drive section?
     
    Last edited: Jul 12, 2010
  10. Phil_Jahelka

    Thread Starter New Member

    Jul 12, 2010
    11
    0
    OK, here's the next draft. I included the drivers, but I'm not sure what value of resistors to use on the MOSFET connections. Also, I'm not sure if I need the boot-strap circuitry, but I included it anyways. Sorry for not showing all the connections on the far right. Resolution is starting to become an issue and I promise I wired it right. Again, thanks for your continued patience and help. Also, I finally understand how the drivers work after reading (http://forum.allaboutcircuits.com/showthread.php?t=33615&page=2). That's pretty damn clever to change how the cap's voltage is referenced!
     
    Last edited: Jul 13, 2010
  11. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    Why don't you attach your .ASC file (it's in the SwitcherCad directory) so that those with LTSpice can give you a hand cleaning it up?
     
  12. Phil_Jahelka

    Thread Starter New Member

    Jul 12, 2010
    11
    0
    Also, I posted the problem on another forum and a guy suggested I just use the PWM's Op-Amp for the audio instead of my crazy bias... setup... thing... Anywho, here's that version of the schematic.
     
  13. Phil_Jahelka

    Thread Starter New Member

    Jul 12, 2010
    11
    0
    Wow, thanks for the offer. Here are my two current versions.
     
  14. Phil_Jahelka

    Thread Starter New Member

    Jul 12, 2010
    11
    0
    Also, disregard that part I said about not needing the bootstrap circuitry. I'm finally starting to understand how those drivers work, it just took me a while. The datasheet doesn't do a very good job of explaining how they work.
     
  15. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    OK, I simply did most of the cosmetic clean-up. I have a bit too much on my plate at the moment to actually perform an analysis on it.

    Note that the H-bridge is symmetrical now. They are much more easily recognized and understood if drawn in such a manner. It's somewhat contrary to the general guidelines for schematics, but makes perfect sense to do it that way - and you will most frequently see such symmetrical circuits drawn that way.
     
    Last edited: Jul 13, 2010
  16. Phil_Jahelka

    Thread Starter New Member

    Jul 12, 2010
    11
    0
    Sir, that is absolutely beautiful. Thank you!
     
  17. Gated

    New Member

    Jul 7, 2010
    15
    0
    Hello,

    I have a large tesla coil that I am constructing, and for it is also audio modulated,

    The driver is a bit different, since it has gate drive transformers and a non-inverting and inverting gate drive feed for the gates,
    also, it uses primary current feedback to get as close as possible to zero current switching, however, my system pulses alot of amps, which is why primary feedback works, however, for your system, you can do a far simpler one using secondary (or antenna, Id just preffer secondary CT feedback), a hex inverting smitt trigger, and a pair of UCCs (TI's gate drivers, inverting and non-inverting)

    note that this drivers pin "3" is ENABLE, this means that you can interrupt the driver from continous operation from the feedback of the secondary (it is always in tune with the secondary's resonant frequency)
    for that pin, I provide a signal, which is usually a square wave with variable on/off time, so I can change arc apprearance, thickness, etc.
    however, if you use an audio modulator (I included a schematic) you can feed that to the 3rd pin, and achive a singing arc output.

    in the schematic, SEC1 and SEC 2 is 1 turn of the secondary coil ground, going through a current transformer (ferrite) the turn ratio is 1:150
    -INTERUPT is the input signal to the 3rd pin that ive mentioned, which is where the modulators signal connects to.
    -the output of the 2 drivers (pin 6+7 shorted) are inverting and non-inverting, the inverting output is to have the 3uF capacitor, the inductor shown is the gate drive transformers primary (depens whether you are running a halfbridge-or fullbridge)
    turn ratio is 1:1:1:1:1 for fullbridge, for half bridge it will be 1:1:1
    I use about 15 turns for this.
    phasing is also a MUST. since you dont want shoot through with your switching devices. so the driver I find is a bit simpler, but you must correctly phase your gates with the transformer

    another neat thing, if you leave the INTERUPT input open (no signal to it)
    the driver runs uniterupted from the feedback of the secondary, meaning that it produces arcs, but are not audio modulated, (still neat arcs)

    slight edit about GDT, this can be directly driven, but GDTs dont have to be a trivial task to make, and if you have an oscillioscope to see your drive waveforms, it helps quite a bit, I just like to use GDT due to the ease of phasing.
    also, where the 10K resistor between the gate and source, I would rather have zeners (back to back, anode to anode if that makes sence) instead of the resistor, it will prevent overvoltage to the gate. (TVS of the same ratings would also work, but zeners are more common and should be enough for this application.

    I hope this helps you out.

    1.jpg

    mod.jpg
     
    Last edited: Jul 13, 2010
  18. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    Just some general "stuff"...
    You had tried to fill in some values for resistances, etc. - but in some cases you changed the reference designator instead of the value field.
    "R3", "C5", "U2", etc. are reference designators. Each must be unique in order for a simulation to run. You can create a schematic that has duplicate reference designators, but the idea of the reference designators is that referring to "R3" means just one part in the schematic, and it's a resistor.

    You had a lot of "white space" in the original schematic. It's not bad to do that in the beginning, but as you've discovered, it's difficult to convert such a large schematic to an image for uploading - you can't see the entire thing without panning all over the place.

    It's important to not overlap reference designator and value fields (like you did with the diode bridge), as that makes it difficult/impossible to determine what the values are without right-clicking on the component.

    There's enough white space in the revised schematic to make it easy to read, and I've used labels to eliminate some of the wiring that was running all over the place.

    You're using 18.5v for Vdd. That's not a typical value; usually Vdd is +3.3v to +16v; lately +5v is much more popular. I didn't look up the datasheets for your H-bridge drivers, but they'll probably need somewhere around 12v to 15v - that will need to be separate from the logic level controls.

    I haven't looked at the datasheet for U1, but it seems to me that R1/R3 are mighty low values. If pins 1,2,3 are an opamp (like I think they are) then the best you'll get out of it is a clipped sinewave, because you have the NI input grounded. It probably needs to be biased at Vdd/2.
     
  19. Phil_Jahelka

    Thread Starter New Member

    Jul 12, 2010
    11
    0
    OK, I think I finally figured out the audio section. The PWM biases the RAMP Valley to 1.25 V with a Valley to Peak Voltage of ~1.8. By adding Bias with the voltage regulator, I get the audio to sit right in the middle of RAMP. With that, I should be able to use the volume control on the Walkman to make the audio voltage swing right. If it's still too small, then I'll use the PWM's Op-Amp. As for my 18.5V messing with logic voltage levels, I looked up their data sheets and I'm pretty sure they'll still work, but I'm not positive.
     
  20. Phil_Jahelka

    Thread Starter New Member

    Jul 12, 2010
    11
    0
    New design! Uses the internal 5.1V Reference for the audio bias and adjustable gain using the PWM's op-amp.
     
Loading...