Connect with us

irf9640 p-chan mosfet switching buck converter blows without load, whats wrong

Discussion in 'Power Electronics' started by magtec, Jul 29, 2020 at 3:08 PM.

Scroll to continue with content
  1. magtec

    magtec

    6
    0
    Wednesday
    no load 60vdc in adjusted for 30vdc out only load is led with 11k ohm, dc signal very clean only
    transistor fails (shorts) for no apparent reason, using 10k ohm source to gate resistance. gate
    pulled down with irfp250n n-chan using 2.6k ohm to ground. irfp250n biased with 5v regulated source
    through 100k ohm on gate pulled down with d526-y npn through 1k ohm. what am I missing?
     
  2. Harald Kapp

    Harald Kapp Moderator Moderator

    10,057
    2,144
    Nov 17, 2011
    Circuit diagram? A picture is worth a 1000 words.
     
  3. magtec

    magtec

    6
    0
    Wednesday
    schematic attached
     

    Attached Files:

  4. Harald Kapp

    Harald Kapp Moderator Moderator

    10,057
    2,144
    Nov 17, 2011
    This doesn't look like a buck converter at all. There is no switching involved - at least none I can see.
    The circuit looks more like a rather strange form of linear regulator. The IRF9640 is a switching transistor. This type of transistor can be operated in a linear fashion only within a limited range, the so called SOA (Safe Operating Area). See figure 8 in the datasheet. This application note discusses linear operation of MOSFETs.

    Where is this design from? Do you really want to use a linear regulator? Why not use a proven design?
     
  5. magtec

    magtec

    6
    0
    Wednesday
     
  6. magtec

    magtec

    6
    0
    Wednesday
    I was using the output voltage divider (130k, 5k pot, 100) to turn on Q3 at the desired output voltage. Q3 on = Q1 off, this would act to
    automatically set the switching frequency according to the load requirements via the voltage drop. This actually works as far as being able to set an output voltage (30-60vdc IN to 10-60vdc OUT) except for some reason source drain will short without a load. I was wondering if the gate control circuitry had something to do with it. My purpose of this circuit was to simplify function and parts replacement as I am working on experimental systems
     
  7. Harald Kapp

    Harald Kapp Moderator Moderator

    10,057
    2,144
    Nov 17, 2011
    I wonder where switching should occur at all in this circuit.
     
  8. magtec

    magtec

    6
    0
    Wednesday
    As voltage builds at C1, Q3 is forward biased for buck desired output from voltage divider 130k, 5kpot, 100(back-up protection for pot)
    pulling gate voltage of Q2 to ground turning Q1 off. Q1 starts on with bias by the 5 volt regulator divider 1k, 10k pot,100
    applied at Q2 gate pulling Q1 p-channel gate lower than source (on), again I verify this actually works as I am able to adjust the
    output voltage with the 5k pot and since the only factor for switching off Q1 is the output side divider voltage to Q3 which is
    totally dependent, not relying on a set frequency or pulse width, this makes the frequency and pulse width set according to the
    charging of C1. Note: I have an led w/10k across output and have successfully temporarily connected a 3ohm load @ 20v out so I am
    not sure why the irf9640 melts down. I am wondering if the gate circuit is amiss somehow or maybe a current spike?, also notable is
    the 15v zener which was an add on in hopes of a fix, this was not in original design which worked the same without it
     
  9. Alec_t

    Alec_t

    2,843
    760
    Jul 7, 2015
    Q2 won't switch on reliably. The IRFP250N has a Vgs(thr) which can be as high as 4V.
    As noted above, this is a linear regulator design, not a switch-mode one. LTspice simulation confirms no oscillation occurs.
    There is no accurate reference voltage source to establish the output voltage, so the output will vary considerably with changes in MOSFET properties and load.
    A 3Ω Load @ 20V implies a ~ 7A current. If the input is 60V then the P-MOSFET will be dissipating (60V-20V)*7A = 280W (which is way above its 190W rating)! Even half that would require a pretty big heatsink to prevent the MOSFET self-destructing.
    The purpose of C1 should be to suppress pulses. The output is presumably supposed to be DC (not pulsed).
     
    Last edited: Aug 2, 2020 at 4:31 PM
    Harald Kapp likes this.
  10. bertus

    bertus Moderator

    745
    264
    Nov 8, 2019
    Hello,

    As @Alec_t calculated, the power for the mosfet is way to high.
    The datasheet shows a maximum dissipation of 125 W.

    Bertus
     

    Attached Files:

  11. magtec

    magtec

    6
    0
    Wednesday
    Thanks alot to everyone involved, looks like I will stick to frequency switching. I appreciate all the help
     
Ask a Question
Want to reply to this thread or ask your own question?
You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.
Electronics Point Logo
Continue to site
Quote of the day

-