Half bridge induction heater mosfet problem

Discussion in 'The Projects Forum' started by ajrenzetti, Nov 5, 2011.

  1. ajrenzetti

    Thread Starter Member

    Dec 20, 2010
    40
    0
    Ok, well so far ive been able to get my induction heater to work somewhat. theres just just one problem... the induction heater only works for a short amount of time, only heating up the workpiece a bit, but then all of a sudden it stops. the light on my 48v 950w power supply goes on indicating that the terminals are being shorted. When i test the mosfet leads with an ohm meter it reads around 5 ohms or less when it shouldnt be indicating anything. ive posted the schematic of the circuit(without the tank circuit, which has a tank capacitor of 4uf and a work coil with 5 turns). what exactly is goin on with my inverter? if you need more specs on it ill do my best to include them.
     
    • ih4.png
      ih4.png
      File size:
      26.2 KB
      Views:
      186
  2. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    R1 and R2 are WAY too large. They should be around 10-30 Ohms.

    You should also have 10v-12v Zeners from the gates to the source terminals, cathodes to the gates. This will prevent the MOSFETs from getting zapped due to excessive Vgs.

    You probably need much larger heat sinks.

    You don't mention anything about the characteristics of your heater coil. It probably has much too low of a resistance/impedance for the frequency that you are trying to operate it at, but we have no idea what that might be as you have not provided the information.

    C2 through C5 will likely be getting VERY warm.
     
  3. ajrenzetti

    Thread Starter Member

    Dec 20, 2010
    40
    0
    woops i forgot to change that 10k to 10 ohms on the schematic. i actually have 10ohm resistors there.

    im going to try adding those zeners you were talking about and see if they help.
     
  4. ajrenzetti

    Thread Starter Member

    Dec 20, 2010
    40
    0
    ok well i tried it but when i turned the inverter on the gates didnt turn on, and when i took the diodes off the gate did turn on, but shortly after shorted just like the rest. besides, i heard that mosfets wont turn on unless the gate charge is 15v more than the drain charge.

    is there something else that may be affecting the mosfets? i was wondering if the source was recieving too much inductive feedback and broke. could that be the problem?
     
  5. ajrenzetti

    Thread Starter Member

    Dec 20, 2010
    40
    0
    i was thinking of using 75v zeners instead of the 12v ones so that way the gate voltage can exceed the the source voltage but not go overboard. Would this work?

    if not, then why arent the regular zeners allowing the gates to turn on when i apply the signal?
     
  6. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    That depends on the MOSFETs in question. Some are logic level; Rds(on) specified at Vgs=4.5v or Vgs=5v. Many are standard; where Rds(on) is only specified at Vgs=10v.
    I can't tell what mosfets you're using. It looks like they start with IRFP, but after that is anyone's guess.

    You have a 1nF cap in series with the xformer primary. I don't know what your xformer ratio or inductance is, and I have no idea what your total gate charge is because I don't know what your mosfets are.

    I have a feeling that your Vgs is marginal. That would toast the MOSFETs very quickly.
     
  7. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    As I said above, I think that your Vgs as it is without the Zeners is marginal. Adding the Zeners added a small amount of capacitance, which caused the Vgs to be below the threshold value.

    I think you have insufficient drive to the gates. The 1nF cap may be the problem, but you should also address the other items I mentioned.
     
  8. ajrenzetti

    Thread Starter Member

    Dec 20, 2010
    40
    0
    the mosfets i am using are irfp064n. ive also added the link to the transformer im using to drive the gates. http://www.murata-ps.com/data/magnetics/kmp_786.pdf

    also, what exactly do you mean when you say "marginal". the word has me a little confused.
     
  9. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    Sorry I confused you. By "marginal", I meant that I think that your Vgs is not high enough when the MOSFETs are ON.
     
  10. praondevou

    AAC Fanatic!

    Jul 9, 2011
    2,936
    488
    Can you post a bigger picture so more details get visible, most importantly the connections between crossing lines. For now it's only guessing.

    The turn ON threshold for the n-channel MOSFET you are using is specified at 2 to 4V VGS (voltage from gate to source) @ 250uA drain current.


    putting a 75V zener from gate to source or nothing is the same thing. have a look in the datasheet. It says +-20V Vgs absolute maximum ratings.


    You mentioned a 48V power supply. Does that mean the voltage at "MAIN" is 48V?
    The rated voltage for the IRFP064 is 55V so I wonder if this could be a problem.

    Are the MOSFETs getting hot?

    If, as Sgt said, Vgs is too small power dissipation of the MOSFET will also increase (depends on your load current). The specified current Id for a given gate voltage you find in figure 3 of the datasheet.

    Do you have an oscilloscope to measure your gate signal?
     
  11. ajrenzetti

    Thread Starter Member

    Dec 20, 2010
    40
    0
    This has always confused me, but the vgs is the voltage needed to turn the mosfet on, right? what voltage should i be applying to the gate to turn them fully on? the main is 48v and the mosfet drivers are being powered by 12v, so with the transformer being a 1:1:1 ratio, the mosfets should be recieving 12v, right? This voltage seemed to work fine when i made a zvs induction heater, so maybe the current is too low? i was thinking maybe it was because of the size of the transformer; that it was simply too small to carry the required current.

    yes my mosfets are getting really hot. i also noticed that they dont break until the frequency is at a certain level, which im pretty sure is just below the resonant frequency. anything below that they just get very very hot.

    also, what exactly is the difference between a logic level mosfet and a standard mosfet?
     
  12. praondevou

    AAC Fanatic!

    Jul 9, 2011
    2,936
    488
    Are the MOSFETs mounted on a heatsink? How big is it?

    Logic level MOSFETs turn on with a lower Vgs. If possible you should be using a normal MOSFET (like you are doing).

    For the IRFP064 I would use at least 12V, better 15V. Gate threshold voltage is quite low, but it is not fully turned on at 2 to 4V and can therefore not provide enough drain current without dissipating to much.

    Do you have an oscilloscope? You should measure the slopes and voltage levels on the gates (Vgs). It would also be interesting to measure Vds directly at the MOSFET pins to see if there are any voltage peaks.

    I wouldn't use a 55V rated MOSFET in a switching circuit supplied with 48V. There is no room for voltage transients.

    If they are overheating then:

    - they are either not mounted on a heatsink
    - switching slope (gate charge time) are too slow
    - drain current is too high

    So what current are you drawing and what is the max switching frequency?

    EDIT: you posted the transformer series specs, but which model are u exactly using ?
     
  13. ajrenzetti

    Thread Starter Member

    Dec 20, 2010
    40
    0
    the transformer part number is 78602/2C.

    are you sure the vgs should be around 12-15v? because i tried driving the mosfets directly from the mosfet drivers instead of using a transformer but they never turned on. someone here told me that was because the gate charge had to be around 15v greater than the drain voltage.

    The power supply im using is rated at 900w, with an output of 48v exactly. i tried using a voltage doubler to double my 115v outlet power to 230v, but the moment i turned it on the mosfets broke. ive also tried using mosfets with a higher voltage rating but that didnt do any good, they still broke.

    i do have an oscilloscope, but it is rather old. it does display the waveform correctly but i cant get a frequency measurement out of it. by the looks of it, however, id say my vco frequency is at around 30-60khz, most likely 50khz.
     
  14. praondevou

    AAC Fanatic!

    Jul 9, 2011
    2,936
    488
    In order to turn the MOSFET on you need to charge the gate to source capacitance. So you apply a voltage from gate to source. The drain voltage could be anywhere as long as it is inside the specs.
    If you have the time have a look here: http://www.electronics-tutorials.ws/transistor/tran_7.html

    Please post the waveform of the gate signal indicating adjusted voltage/div.
     
Loading...