• State Resolved
  • Locked Locked
  • Replies 5 replies
  • Subscribers 63 subscribers
  • Views 718 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

Failure analysis of LM5067

Thomas Ham14
Prodigy 110 points
Other Parts Discussed in Thread: LM5067, STRIKE, LM5064

Alex,

During IEC 61000-4-5 testing, an LM5067 failed. There is a small crater just behind 8 in the top of the part. Can you tell me what functions are in that area so I can try to figure out what damaged the part?

Thanks!!

  • 0
    TI__Genius 12621 points
    Hi Thomas,

    I saw you posted another thread which is a follow on update. Will post your additional info here and close that second thread:

    "It looks like the crater is centered behind pins 8 and 9 on the top of the part."

    I will discuss with a team member who has a bit more experience regarding IEC 61000-4-5 surge testing for -48V hot swap and will get back to you. If you can share a partial schematic (what MOV and/or TVS you are using for example), that would be very helpful!

    FYI - to attach an image, simply click "Use rich formatting" on the bottom right-hand corner of the reply box. Then insert image icon.

    This thread is viewable by the public, so please don't share any confidential information.

    Although we can dig through the silicon design to see what structures there might be (and we certainly will if necessary), I don't believe that would immediately point towards a root cause. Overall if the surge is not sufficiently clamped by external devices, then our IC will experience voltages and currents well beyond the ABS_MAX table and lead to damage.

    Thanks!
    Alex
  • 0 in reply to Alex Triano
    Prodigy 110 points

    Alex,

    There are no options to insert rich text formatting. The only buttons are along the top edge of the reply box. I have used "insert file". 

    Tom

    4150.SCHEMATIC1 _ PAGE1.pdf

  • 0 in reply to Alex Triano
    Prodigy 110 points

    Hi Alex,

    Any news yet on what your teammates think?

    Tom

  • 0 in reply to Thomas Ham14
    TI__Genius 12621 points
    Hi Thomas,

    What level of IEC 61000-4-5 testing are you looking to complete (as well as what is the surge tester source impedance)? My guess is you're testing for Class 4 around 4kV - 5kV.

    At such high levels, at least a couple MOVs are needed (metal oxide varistor). The 2x SMCJ TVS diodes would be insufficient to clamp a lightning strike. If you do have some MOVs on board, what voltage to they clamp at? If they clamp above the TVS voltage, then all the energy will be directed into the TVS diodes first (and they may break).

    Can you verify if the LM5067 was the only device damaged, or did the TVS diodes break as well?
    I would guess the TVS diodes could not clamp the lightning surge, thus letting a high voltage through to the LM5067 and damaging the part by exceeding ABS_MAX.

    My guess is that stronger MOVs or different MOV clamping voltage may be needed (or if there are none, they would be needed).

    After this, a few diodes and resistors can be added to make the design more robust. Check out the LM5064 datasheet (which is a similar device as the LM5067). Check out pages 27-28.
    www.ti.com/.../lm5064.pdf

    My guess is you'll need a few schottky diodes on the SENSE and OUT pins to help protect transients as well as a high power schottky diode on the output (currently there's a TVS, but a schottky would be needed to clamp the negative voltage transient when the device shuts off. A TVS may not be needed there at all, and if so, does not need to be bi-directional - though I understand it's easier to keep a single part number).

    Thanks,
    Alex
  • 0 in reply to Alex Triano
    Prodigy 110 points
    Hi Alex,
    The surge is -500V, 1.2uS rise time and 50uS duration. The only component damaged is the LM5067. There are no MOV's on the boards so I will look into that. I will also look at the LM5067 sheet you mention.
    Thanks,
    Tom