1. This forum section is a read-only archive which contains old newsgroup posts. If you wish to post a query, please do so in one of our main forum sections (here). This way you will get a faster, better response from the members on Motherboard Point.

PIC16F716 physically exploding

Discussion in 'Embedded' started by Fred, May 4, 2007.

  1. Fred

    Fred Guest

    Hi there,

    Does anyone have experiencing seeing a PIC16F716 physically blow its
    lid (in other words, explode)? I have a design that uses three SO pkg
    of the F716 and one of them is intermittently exploding. It's always
    the same one, and never either of the other two.

    I have never seen this before and I don't know what would be causing
    it.

    Any help appreciated. Thanks.

    Fred
     
    Fred, May 4, 2007
    #1
    1. Advertisements

  2. How does something "intermittently explode" ?!?

    What is connected to it ?

    -jg
     
    Jim Granville, May 4, 2007
    #2
    1. Advertisements

  3. Fred

    jcomeau_ictx Guest

    A schematic might give us a clue? Or is it too confidential?
     
    jcomeau_ictx, May 4, 2007
    #3
  4. Fred

    Roman Guest

    All PICs have schottky diodes on all I/O pins. Wire -12 volts strong
    enough to drive several amps onto an I/O pin and it may blow like you
    are describing.

    I/O drivers are current limited. You should not destroy PIC by driving
    an output against output or voltage source between 0-Vdd.

    Your description is not enough, this is just a hint.
     
    Roman, May 4, 2007
    #4
  5. If you have a power supply capable of very high current, it could
    easily be latchup, but you'd have to be severely abusing the part to
    get it to latch up.



    Best regards,
    Spehro Pefhany
     
    Spehro Pefhany, May 4, 2007
    #5
  6. Fred

    Fred Guest

    My power supply is 4xAAs. No pin is seeing more than 6v. There is a
    diode in series with the batts, but other than that, no current
    limiting.

    What kind of thing causes latchup?

    Have any of you personally seen one of these chips blow? The tops
    literally come off, leaving a 1/4" diameter pock-mark.

    Fred
     
    Fred, May 4, 2007
    #6
  7. In the bad old days it was by turning on a parasitic SCR when
    you apply a voltage above Vdd to an I/O pin. Most parts have
    protection diodes to prevent that from happening.
    I've not seen that particular one blow up, but I've seen others
    do it. I remember one 14-pin DIP op-amp that had supply and
    ground pins centered on each side. If you put the chip in
    upside down it would quite reliably blow a chunk out of the
    package leaving a nice crater. If you took a drinking straw
    and held one end against the part when you flipped the power
    switch you could shoot a chunk of epoxy a fair distance with
    some accuracy.
     
    Grant Edwards, May 4, 2007
    #7
  8. Add some current limiting, or better, a LDO regulator.
    Current injection into the parasitic SCR all CMOS devices have.
    can be -ve or +ve spikes, of short duration - just need to
    be long enough, and of enough magnitude, to fire the SCR.
    Once fired, it attempts to crowbar the Supply
    - in this case, loosing :)

    What is connected to this chip ?

    -jg
     
    Jim Granville, May 4, 2007
    #8
  9. Fred

    Walter Banks Guest

    The only time I have seen top off and crater is when the part
    joined the 110 club. 220 or 240 in other countries

    I have seen some very hot parts from latch up but never one to
    explode.

    w..
     
    Walter Banks, May 4, 2007
    #9
  10. Fred

    Mark Walsh Guest

    Where is this device used? We have controllers in remote areas that are
    subject to lightning strikes. The spark gaps and input protection usually
    handle it, but sometimes we get controllers back with missing chips and lots
    of epoxy powder.

    Mark Walsh
    Rogue Engineering Inc.
     
    Mark Walsh, May 4, 2007
    #10
  11. All PICs have schottky diodes on all I/O pins. Wire -12 volts strong
    enough to drive several amps onto an I/O pin and it may blow like you
    are describing.[/QUOTE]

    In the bad old days of Unibus and wirewrap boards, a wireman had a pin
    on a Biploar PROM that was not connected, so he 'helpfully' connected
    it to a spare pin on the edge connector. Shame it was -15V.

    When the system was turned on and the board was tested in an open rig the
    complete top of the PROM was blown across the lab.

    Mind you that engineer always had the misfortune of having new boards
    catching fire, blowing up, smoking, sparking and many other disasters.
     
    Paul Carpenter, May 4, 2007
    #11
  12. They don't prevent it, only increase the current required to make it
    happen. A few hundred mA for a fraction of a microsecond ought to do
    it at room temperature, less if the part is hot.

    Best regards,
    Spehro Pefhany
     
    Spehro Pefhany, May 4, 2007
    #12
  13. NiCd cells by any chance? They can supply extremely high fault
    currents.

    Best regards,
    Spehro Pefhany
     
    Spehro Pefhany, May 4, 2007
    #13
  14. Not Schottky.
    Or a brief pulse (eg. from static discharge or lightning).
    Not instantly anyhow, if the ambient is reasonable.

    Best regards,
    Spehro Pefhany
     
    Spehro Pefhany, May 4, 2007
    #14
  15. Fred

    Fred Guest

    Mind you that engineer always had the misfortune of having new boards
    Yes, it's amusing when chips blow up ... except when they're your
    own :-( ...

    Here's a schematic of the Exploding PIC (part U10):

    http://handyboard.com/tmp/exploding-pic.png

    Some notes:
    * there are 3 PIC16F716's on the board (it's improperly labeled as a
    715). This is the only one that's blowing.
    * Vdd is 4xAA alkaline run through a Schotty diode, so it could draw
    ample amounts of current.
    * the chip is wont to blow up even when nothing's plugged into the
    motor outputs or sensor inputs.
    * 6 of about 50 units have had the chip blow. in an earlier production
    run (last year), none exhibited any problem.

    The only thing I can imagine that I'm doing differently with this PIC
    (versus the other 2 on this board, and all of my previous designs) is
    that I'm driving its oscillator-in (pin 16) with the oscillator-out of
    one of the other PICs.

    Can this possibly explain anything?

    Thanks all for your suggestions!

    Fred
     
    Fred, May 5, 2007
    #15
  16. Fred

    Fred Guest

    Hi Spehro, can you elaborate on this? Thank you. Also, it has blown
    with nothing plugged in.

    Fred
     
    Fred, May 5, 2007
    #16
  17. What clamps the motor inductve energy ?
    With no regulator, and battery/diode, there is nothing to stop
    high short-term voltages, and that WILL fry chips.
    They may be stressed, and fail some time later.

    There are no series ESD resistors in the sensor lines - also bad
    practice.

    -jg
     
    Jim Granville, May 5, 2007
    #17
  18. Probably a layout issue (causing latchup)-- with those motors it just
    requires a lack of consideration of the flow of ground return
    currents.


    Best regards,
    Spehro Pefhany
     
    Spehro Pefhany, May 5, 2007
    #18
  19. If current (particularly current that is high and/or changes quickly)
    goes through a path with enough inductance, a voltage will appear when
    the current changes quickly. If that voltage is higher than a volt or
    so, depending on the layout, and flows through a "ground" trace, it
    can cause current to flow into or out of an I/O pin because two
    "grounds" are not at the same potential. If there's enough current
    available, latchup can occur. That's not necessarily fatal for the
    chip, but if the power supply (Vdd for the chip) can supply enough
    current, it delivers the coup de grĂ¢ce.

    Oh, and be sure you actually have the L293D and not the L293.


    Best regards,
    Spehro Pefhany
     
    Spehro Pefhany, May 5, 2007
    #19
  20. Fred

    Rocky Guest

    Fred,

    Do you ever use /MCLR ?

    Grounding the pin using even a short track can cause latchup due to
    undershoot transients. You need to resistor in series with the /MCLR
    pin so the cct should be VDD - 1K - 100R - /MCLR and the reset switch
    is connected at the junction of the 1K and the 100R, NOT directly to
    the PIC. We normally use VDD-10K-1K0-PIC with a 1nF to ground at the
    PIC to ensure there are no ugly transients.

    Robert
     
    Rocky, May 5, 2007
    #20
    1. Advertisements

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.