Connect with us

Flyback duty cycle locked

Discussion in 'Electronic Design' started by Hammy, Mar 15, 2010.

Scroll to continue with content
  1. Hammy

    Hammy Guest

    I built this 10W flyback. See schematic.

    http://i39.tinypic.com/ops5jd.png

    It was working perfectly for several days of testing then I powered it
    up and it stopped regulating the duty cycle remains locked at a little
    over 4% the outputs are at about ten percent of there regulated
    values.

    I noticed my once nice and clean DCM waveform over my current sense
    resistor (R4_D) is now oscillating at about 1.5MHz.

    Scope shot current sense resistor (R4_D) 10mV/div.

    http://i42.tinypic.com/jzd1n5.jpg

    I suspected the overcurrent protection may be kicking in but there is
    no indication of this. If the OCP is active the SS pin should be
    ramping up and down trying a restart but this isn't happening.

    The controller is the TPS40210.

    http://focus.ti.com/lit/ds/slus772c/slus772c.pdf

    I'm down to switching out the controller, unless someone can offer an
    explanation.

    I'd prefer not to pull the controller out because it would be a bitch
    without damaging the traces it's a fine pitch 10 pin MSOP-PowerPAD
     
  2. Hammy

    Hammy Guest

    correction its oscillating at about 33MHz.
     
  3. D9/Q4 could be causing the oscillation. Short D9 and see if the
    symptoms become clearer.
     
  4. Hammy

    Hammy Guest

    I tried as you suggested. No luck now the controller gate drive is
    erratic almost like burst mode. I know some controllers have this
    feature but to my knowledge the TPS40210 doesn't.

    I was going to try as MooseFet suggest but I think at this point its
    safe to say the controller somehow got damaged.

    The gate waveform is almost triangular versus the nice sharp clean
    transtion when I first put this together. I must have inadvertently
    shorted the gate drive or something.

    I'll redo the board . I'll place a shottky clamp on the gate drive
    just in case. I did check all the pins for negative transients and
    they were all fine. I always put a clamp on the CS pin regardless; D20
    in the schematic.
     
  5. Joerg

    Joerg Guest


    The initial ramp-up is suspiciously steep. Sure nothing at or behind the
    transformer has shorted out? Shorted winding? Crack in the toroid core?
    D4/D4 shorted out?


    Probably doesn't matter if you didn't get an unconnected pin warning
    upon netlist generation: The emitter of Q4 is sort of naked.
     
  6. Hammy

    Hammy Guest

    I'm pretty sure the initial spike on the current sense isn't from
    drain source conduction but from the gate driver. The current sense
    signal is RC filtered then fed to the controller which also has
    leading edge blanking.

    I pulled the transformer out so I could test for shorts and there is
    none. I can't see any cracks in the core but I measure the correct
    inductance with my LC meter. If there was a hairline crack in the core
    would you still get a correct inductance measurement?

    The behaviour of the controller is just screwy.

    Its just the way the model is drawn,its connected on the board. I
    didnt even noitice that good eye. ;-)
     
  7. Joerg

    Joerg Guest


    Then your Isense current would shoot straight up like a rocket.

    Yeah, looks like the chip is kaputt :-(

    Only because I wear 1.5x glasses when at the computer :)
     
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

-