Connect with us

LM317 exploding electro

Discussion in 'Electronic Design' started by John Morris, Dec 7, 2006.

  1. John Morris

    John Morris Guest

    I know someone is going to make me feel stupid for asking this, but
    ....

    I have built the standard LM317 PS circuit with variable pot, etc. on
    a solderless proto board. Input is 37VDC from a plugpack.

    It works fine for about 20 seconds, then the 10uf 63V electro in
    parallel with the pot heats up and explodes.

    I have checked the circuit a dozen times and cannot find a fault.

    Can anyone offer a clue? I have heard some brands of LM317 have
    differing pinouts.

    Thank you,

    John Morris (the moron)
     
  2. Guest

    "John Morris дµÀ£º
    "
     
  3. Phil Allison

    Phil Allison Guest

    "John Morris"

    ** That 10uF electro MUST have been installed with reverse polarity.

    Maybe it was wrongly labelled.




    ....... Phil
     
  4. maxfoo

    maxfoo Guest

    This is a schematic of what your circuit should look like.
    http://www.national.com/images/pf/LM317/00906301.pdf

    max input voltage is 37Volts on a LM317.
    Try using a resistive voltage divider to cut the input
    voltage a bit.

    Also the max current output is 1.5 amps, put a series current limiting resistor
    with the POT, most likely the min resistive value on the pot is your problem.
     
  5. Phil Allison

    Phil Allison Guest

    "Phil Allison"
    Or it is really a 6.3 volt cap.



    ...... Phil
     
  6. maxfoo

    maxfoo Guest

    hehehe... I once had a board that had a tantalum cap in ass-backwards, the thing
    blew up making a loud BANG! and then caught on FIRE!, scared the SHIITE out of
    me. The flame was an inch high and the whole lab stunk for hours, put a nice
    black spot on the pcb. Damn a-hole assemblers got me again.
     
  7. Eeyore

    Eeyore Guest

    I rather doubt that.

    What's this *standard circuit* anyway ?

    Have you made sure you have local decoupling for the regulator ?

    Graham
     
  8. There's no way that should happen. Even if you get the electrolytic
    backwards, the adjut pin on a good LM317 only supplies 1.37 milliamps,
    which is unlikely to blow a capacitor in 20 seconds.


    Use your voltmeter and check that the LM317 is getting a PLUS voltage
    to its pin 1 (the leftmost pin). Not Minus. Not AC.

    My bet is it's AC.
     
  9. John Morris wrote:
    (snip)
    (snip)

    Either the capacitor was in backwards (the - terminal goes
    to ground) or the regulator is unstable and is oscillating
    at high frequency, heating the capacitor. You should look
    at the data sheet for the recommended (minimum) ceramic or
    film capacitors that should be connected as close as
    possible to the pins of the LM317, to stabilize it.

    See page 3 of:
    http://www.fairchildsemi.com/ds/LM/LM317.pdf
     
  10. Perhaps he's looking at the pin view (i.e., backwards).
     
  11. ehsjr

    ehsjr Guest

    There should not be a cap in parallel with the
    pot. Post your schematic so we can see what you
    are talking about. No "standard" LM317 circuit
    blows capacitors.

    Ed
     
  12. Phil Allison

    Phil Allison Guest

    "ehsjr"

    ** You had better look at the original Natsemi app notes for the LM317.

    The schem with the additional protection diodes and 10 uF cap across the
    pot.

    Improves ripple rejection by around 15 dB, according to the published data.




    ...... Phil
     
  13. ehsjr

    ehsjr Guest

    Thanks! You're absolutely right. I did not consider the
    circuit with the protection diodes (or any LM317 vr circuit
    with other than the input and output caps and the voltage
    divider) the "standard" LM317 circuit the OP asked about,
    but perhaps I should.

    Ed
     
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

-