Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Wiring a gate driver correctly (keeps burning out)

  1. Apr 26, 2013 #1
    I bought the TC4421 to drive a IRFP250N power MOSFET. I'm using a 12v battery as the power source, and an Arduino to make a square wave input at ~1kHz. I have things wired this way:

    TC4421:
    GND to ground
    VDD to +12v line on battery
    Input to PWM output from Arduino
    Output to gate of IRFP250N

    The FET is driving a flyback transformer to make plasma. I put about 10 turns it. Everything works well at first for the first few times: I have it on for about 1-4 seconds at the time. Neither component gets hot, I have them both heatsinked on copper and the FET BARELY gets warm, the TC4421 is at room temp.

    All of the sudden, everything stops working, the TC4421 gets VERY hot and it just dumps current to ground regardless of what's going on at the input. Am I wiring something wrong, and/or can I fix it somehow? I blew through 3 of them already and am running out, plus they're about $3 a piece.

    Thanks!
     
  2. jcsd
  3. Apr 28, 2013 #2
    We will need some schematic to even have a clue.
     
  4. Apr 28, 2013 #3
    Last edited by a moderator: May 6, 2017
  5. Apr 29, 2013 #4
    I think your FET is already damaged due to inductive voltage spike of the coil.
    Try using free wheeling diode across the coil but that will also prevent quick dying of current, so probably won't produce plasma.
     
  6. Apr 29, 2013 #5
    A-a-L -- I agree, the energy in the coil needs to go somewhere, a FWD is needed there - you may want to have this physically from Ground to Source on the MOSFET ( as opposed to just acreoss the terminals of the transformer) - if you have long leads to the transformer, the connection will have high parasitic inductance, which should also be accounted for ( although 200V MOSFET with 12 supply is probably OK with parasitics like this) .

    Also - you do not show a gate resistor, note the MOSFET DS shows 3.9Ohms - The Gate of a MOSFET is almost purely capacitive ( and non-linear with voltage) - the Gate resistor helps dampen oscillations in the gate signal. With that Driver - You may also want to have a R(GS) - to discharge the Gate. I have a feeling at some point you are not turning the MOSFET off - due to Gate Capacitance ( charge) and /or oscillation.

    Also - a pull down resistor in the control signal between the Arduino and the Driver.

    Lastly - having a scope trace will help tremendously in trouble shooting this type of circuit - just keep in mind you are dealing with pulses that may be much greater then 12V.
     
  7. Apr 29, 2013 #6

    jim hardy

    User Avatar
    Science Advisor
    Gold Member
    2016 Award

    i'd think you might relocate the mosfet to low side of transformer.

    Observe mosfet has source to gate maximum voltage of 20 volts. It doesn't take much of a kick to exceed that.

    Perhaps you are failing MOSFETS too ? You didnt say....

    here's a handy appnote from IR about using mosfets, see particularly paragraphs 3 and 4.
    http://www.irf.com/technical-info/appnotes/an-936.pdf

    old jim
     
  8. Apr 29, 2013 #7
    HEXFET devices are particularly sensitive to gate-to-source overvoltage, and as Jim wrote, you'll have plenty of that when your FET switches off. The flyback primary can easily pull the source terminal of your FET several hundred volts below your circuit common (especially when you use the secondary for arcing) and create a low impedance gate-to-source conduction path in the process. Your FET driver will then look into a short for DC current to flow through, made worse due to the lack of a gate resistor, and the FET itself will probably start switching erratically, if at all, and likely end with all its terminals shorted.

    What kind of flyback transformer are you using?
    Where did you get it from?
    Do you know how many windings of how many turns are on it?

    A word of caution, some flyback transformers can store a good chunk of energy, which makes their high voltage very dangerous.

    A couple of years ago, a colleague of mine was sent to the hospital for observation after experiencing a rather severe electrical shock. He was very hesitant to go since "he felt fine" and it took some effort to convince him. This proved to be the right call, though, as he had a heart attack during the night. You often won't be aware of the effects of an elecrical shock until much later.

    Edit:
    Allowing the current in the primary to freewheel will defeat the purpose of using a flyback transformer. You want to force the amp-turns into the secondary during the discharge phase.
     
  9. Apr 30, 2013 #8
    I don't know if I blew the mosfet, is there any way to check? I know nothing about circuits, but would some back-to-back zener diodes across gate/ground help save it?

    I got the transformer from ebay, it's a replacement part for a CRT TV. I wanted to make a plasma speaker but it didn't work so then I want to just make plasma (it's for a demonstration to kids, don't worry I won't let them touch it).

    And how would I store the amp-turns into a secondary? Sorry I don't know how these transformers work as you can probably tell.

    Thanks a lot for the help everyone.
     
  10. Apr 30, 2013 #9

    mfb

    User Avatar
    2016 Award

    Staff: Mentor

    GS and DS zener diodes can help to save MOSFETs.

    To check the transistor, you can measure the gate-drain current versus voltage. It should be nearly zero for low voltages, and increase quickly after some threshold.
    At some GS voltage where the transistor conducts, check the DS current. It should be nearly independent of the DS voltage (if the voltage is not too small), and some multiple of the GS current (given by the data sheet?).
    At zero GS voltage, DS current should be basically zero independent of the DS voltage (within the transistor specifications, of course).

    If that works, the transistor is probably fine.
     
  11. May 1, 2013 #10
    You can use a multimeter to test the basic functionality of your MOSFET:
    YouTube: How to test an RF MOSFET with a digital multimeter

    You should use your 12 V supply (or a 9 V battery etc.) instead of the low-voltage multimeter to charge the gate capacitance of the IRFP250N.

    If the MOSFET checks out, it might still be damaged and have degraded performance, but at least you can rule out shorts between its terminals.

    If current can no longer flow in the primary, its amp-turns must necessarily shift to whatever windings will allow current to flow. The flux in the core must be a continuous function of time.

    If you know nothing about circuits and have no experience with flyback transformers, maybe you should try out a proven design first. And be careful about it, they really can be quite dangerous.

    Edit:
    My previous post was meant to encourage you to move your FET to the low side of the transformer, as Jim advised. I don't know if that was clear.
     
    Last edited: May 1, 2013
  12. May 2, 2013 #11
    Do you have any links to some proven designs I could build? Thanks.
     
  13. May 3, 2013 #12

    jim hardy

    User Avatar
    Science Advisor
    Gold Member
    2016 Award


    None that I've built. My search turned up these, see if any of them look to be based on real world tests:

    http://www.angelfire.com/80s/sixmhz/flyback.html

    http://amasci.com/tesla/plasplan.html (this fellow is pretty good amateur site)

    http://www.teslaboys.com/Plasma/PlasmaPower/ I personally like the looks of his fifth one
    http://www.teslaboys.com/Plasma/PlasmaPower/PlasmaGlobeDriver2.gif

    http://www.personal.psu.edu/sdb229/plasma ball power supply.html
     
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook




Similar Discussions: Wiring a gate driver correctly (keeps burning out)
  1. Wiring boolean gates (Replies: 12)

  2. OpAmp burn out (Replies: 5)

  3. Burned out BJT? (Replies: 8)

Loading...