H-bridge keeps shorting problem

Discussion in 'The Projects Forum' started by ajrenzetti, Mar 28, 2011.

  1. ajrenzetti

    Thread Starter Member

    Dec 20, 2010
    40
    0
    Hi, im currently making an induction heater based on the following design: http://danyk.wz.cz/induk3_en.html

    The only differences are:
    -Im using 10ohm resistors instead of 22ohm
    -instead of using 4 igbt's, i'm using an H-bridge igbt module (MG100Q2YS1) which has a max power dissapation of 800W.
    -Im using 120vac 60hz instead of 220vac
    -I have a 1kw heating element in series with the variac and bridge rectifier; it measures 20ohms.
    -I have a 3200uf 250v dc cap instead of a 2200uf 500v cap.


    Anyway, for some reason, when i turn it on the waveform (on my oscilloscope) never seams to change much, even when its tuned. I tried heating something up but nothing happened. A while back i had some nice mosfet modules instead of an igbt module, and when i removed the heating element and tried it they both blew up, including the ir2153 driver. I recently hooked up the variac output to a 500w isolation transformer and tried that. when i tried that it worked for about 2 seconds and then the variac started humming loudly. after turning it off i soon realized the resistance of my igbt module when in the off state was almost 0 ohms. Im guessing something burnt it out, but its rated for 800w and the isolation transformer was rated for 500w.

    anyway, my questions are:
    Why does my half bridge keep shorting when its practicly the same design as others that work. also, is my igbt burnt out? and if it is, how could it have happened when it was connected to a 500w isolation transformer.

    What should i do to get this thing to work?
     
  2. kubeek

    AAC Fanatic!

    Sep 20, 2005
    4,669
    804
    What is the inductance of your coil? What is your drive frequency? Are you sure you have the two fast diodes in dashed box in place? Are you sure the tank circuit on the output is in resonance?
     
  3. ajrenzetti

    Thread Starter Member

    Dec 20, 2010
    40
    0
    ill have to check what the inductance of the work coil is, but i know it was being driven to resonance because i tested the tank circuit out with a square wave signal generator and and recorded the frequency that my oscilloscope displayed. the frequency of the main circuit was the same as the test, which was 236khz.

    for the diodes, it states on the page that instead of using that kind of diode, an igbt or mosfet with a built in diodes would work as well.

    I forgot to mention, i bought 10 ir2153's and im now down to 4... it seems every once in a while they stop generating the signal. whats wrong? is there any way to avoid this?
     
  4. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    I can't get to the page you linked to; a router must be down someplace.

    Is your heat sinking adequate? If parts are getting hot, they can fail suddenly.
     
  5. ajrenzetti

    Thread Starter Member

    Dec 20, 2010
    40
    0
    the igbt module was hooked up to a small aluminum heatsink. It was barely even warm when it stopped working. its a very large module meant for high power applications. im still not sure how it could have failed.
     
  6. shortbus

    AAC Fanatic!

    Sep 30, 2009
    4,000
    1,512
    If you were using both sides of the modules on at the same time, you could of had something called shoot through, basically a short which will kill the modules. Post your schematic of how Your circuit is, not telling what you changed from the original, but how its wired now. Easier for people to help you.
     
  7. ajrenzetti

    Thread Starter Member

    Dec 20, 2010
    40
    0
    ok heres the schematic. hope it helps
     
  8. CDRIVE

    Senior Member

    Jul 1, 2008
    2,223
    99
    I can't find a common in your circuit. The neg end of the bridge and C10 are floating.
     
  9. kubeek

    AAC Fanatic!

    Sep 20, 2005
    4,669
    804
    You need identical resistors prallel with c6 and c15.
    You are missing the diodes across the main transistors. Those ARE important.

    I speak that language so if you have any questions to the text feel free to ask.
     
    Last edited: Mar 29, 2011
  10. ajrenzetti

    Thread Starter Member

    Dec 20, 2010
    40
    0
    woops i forgot to add common to part of the circuit on the schematic... thanks for noticing that, ill update the correct schematic when i get home.

    oh and the diodes arent missing from the circuit for the main transistors, as they are built in to them.
     
  11. kubeek

    AAC Fanatic!

    Sep 20, 2005
    4,669
    804
    They may be in them, but still you should have a veryfast diode in parallel beacuse the one in the transitor is not fast enough.
    Does the circuit work without the load?
     
  12. ajrenzetti

    Thread Starter Member

    Dec 20, 2010
    40
    0
    The waveform looks ok at the inverter output, but only if i have my heating element in series with the mains power supply.

    What exactly do you mean by "without the load". i know what the load is but im not sure how i should exclude it. Are you saying if i take the tank circuit out, or if i simply dont put a metal rod within the work coil?
     
  13. kubeek

    AAC Fanatic!

    Sep 20, 2005
    4,669
    804
    I mean excluding the tank circuit, maybe slightly loading the driver with some resistance, then meybe adding some inductance in series. You need to check what is going on with the circuit, for example if there is sufficient dead time, what voltage gets induced on the transistors etc.
     
Loading...