• State Resolved
  • Locked Locked
  • Replies 8 replies
  • Answers 2 answers
  • Subscribers 63 subscribers
  • Views 273 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.

TPS23758: IC Fails, short to PGND

eljeffo
Intellectual 560 points
Part Number: TPS23758

My design is based on the TPS23758-EVM

So I'm working on some PCB layout optimization, and am currently in the process of hand soldering everything together.

For about 50% of my boards (6 of 12) things work just great.

The other 6 seem to fail with the TPS23758 shorting VPD (POE Positive) directly to PGND.  Of the 6 failed chips I've replaced several, only to have the following IC die similarly (usually overheat).

With VPD->PGND being a dead short, it's no wonder the chip gets hot enough to make my finger stick to it.  My finger is now branded with the TI logo. :)

I'm curious, what are perhaps 2-3 common mistakes I could have made that make some of my boards super IC killers?  

I can see my POE switch trying to power on the chip 3-4 times, and at that point cuts off power for longer and longer (I assume due to residual heat)

On one chip, sparks flew out and took out a corner of the chip.

I'm not perfect but my soldering isn't thaaaaat bad :)

Prior to power on, there is no dead short.  Power on once with POE, and dead short appears within seconds.

I pulled the diodes out and checked operation.  Seems ok.  

I verified resistors to be the correct value.

I changed the Mosfet on the output just for good luck.

I pulled the transformer (LDT1018) and tested all leads, seems good (connected where it should, not connected where it shouldn't)

I even ordered a new batch of TPS23758 chips to replace the dead ones... the new ones die as well (probably something on the PCB!)

I moved all parts to a new PCB (keeping all old parts) except putting in a new TPS23758, but the TPS23758 dies! 

No luck.  Once a PCB becomes an IC killer board, it's always an IC killer board.

Any ideas? Maybe there's something conceptual that would reduce my search space down to a number of components and I could swap those out even if they "seem good".

Thanks again everyone!

Jeff

  • 0
    TI__Mastermind 21320 points

    Hello Jeff,

    It is hard to not point to the soldering. If something was wrong with the design, I would expect it to break all of the boards. 

    Once a board becomes bad, I would try removing the IC and measuring the resistance between VDD_RTN. Then place an IC and remeasure. 

    Additionally, if there is a short, you would see a spike in current even at low voltages (ex. 2V input would draw something like 1A+). So you could try testing it without applying the full 48V. 

    Additionally, could ESD be an issue here?

    Lastly, just so you know, we use either a professional assembly company to build boards, or technicians to replace IC's. I have tried some IC's, and my luck is 50% as well. If you are verifying a design, I would suggest looking into a professional company.

    If this post answers your question, please indicate so by marking this thread as resolved. Thank you.

     

    Regards, 

     

    Michael P.

    Applications Engineer

    Texas Instruments 

  • 0 in reply to Michael Pahl
    Intellectual 560 points

    Hi Michael,

    Your guidance is well taken.  I was hoping you would say something like, did you know a bad D## or Bad Q## can cause your chip to go nuts?

    I'll try your VDD/RTN measurement advice.  What's the meaning of a change in resistance?

    Are there certain pins on the IC where shorting is most likely to cause complete failure?  Like RSNS/CP? RTN/VSS? FB/Comp? VDD/VPD?

    In most case in digital circuits, shorts just cause logic malfunction.  But I appreciate with all kinds of voltages flying around there might be some really important combos to check.  Let me know if you have any favorites to check.

    Good advice on the low voltage short check... I'll try it next time!

    ESD - Maybe? In the last 3 chips I grounded myself pretty reasonably just to take out that as a factor, but EMI and ESD seem to be magical cousins.

    But if that's not the case then.. I guess it is my soldering.  I just wish I could find where I went wrong.  Usually for other chips that don't destruct at some point I find it, but this circuit has been tough! :)

    And yes, to prove it's just me, I have 10 boards coming to me professionally made.  Just to show me how bad I am. :)

    Thanks as always!

    Jeff

  • 0 in reply to eljeffo
    Intellectual 560 points

    So I can share that when I measure the VDD_RTN resistance it's 7M ohm when meter negative is on RTN.  But swapped around it's 2M Ohm.  This is regardless of a working/non-working (short) board.

    On the board where the IC blew off the drain corner haha, the resistance starts about 7M then heads towards infinity. :)

  • 0 in reply to eljeffo
    TI__Mastermind 21320 points

    Hey Jeff,

    If the board is thermally overheating, I would think this would either be a short between VC_RTN or VDD_RTN. Or potentially VDD_VSS. 

    That would be the main areas that would even have enough power to heat the circuit enough to blow off a corner. 

    The only other thing that comes to mind is if there is a diode placed backwards or if the transformer was placed backwards. 

    Let me know how the 10 boards turn out! 

    If this post answers your question, please indicate so by marking this thread as resolved. Thank you.

     

    Regards, 

     

    Michael P.

    Applications Engineer

    Texas Instruments 

  • 0 in reply to Michael Pahl
    Intellectual 560 points

    Michael,

    Does it ring any bells with you if I get a secondary output of about 2.5V?  It happens here and there.  PD detection at the switch is fine.  But just output isn't 5V.

    Does it seem more like my feedback (however that magic works with this chip is) is bad?  Or does it seem more like a fried IC?

    I'll let you know how the other 10 boards turn out.  I'm 0 for 3 today so it's... a sad day :( hahaha

    Jeff

  • 0 in reply to eljeffo
    Intellectual 560 points

    Resoldered the TPS23758 for the 5th time.  Works.  Okay I'm just going to go with what you said... my soldering skills need work.

    What is it again that it takes to be one of those cool technicians you speak of?  I want to be awesome like them!

  • 0 in reply to eljeffo
    TI__Mastermind 21320 points

    Hello Jeff,

    If the output voltage is incorrect but the IC is soldered good, it could be something with the feedback circuit or the timing of the secondary side. 

    I know in Dallas they offer 2 year degrees to become a soldering technician, so don't beat yourself up on it if you haven't had technical training!

    If this post answers your question, please indicate so by marking this thread as resolved. Thank you.

     

    Regards, 

     

    Michael P.

    Applications Engineer

    Texas Instruments

  • 0 in reply to Michael Pahl
    Intellectual 560 points

    Michael, 

    When I get a gob of money and time, I'm going to sit down and learn how to solder for 2 years.

    Until then I'll just ... straighten up my IC placement :) hahaha

    I have some older boards where I looked and the problem doesn't look the same, but... we'll come back to that another day.  Main thing is the newest design / layout works ok.

    I fixed the remaining boards with a similar problem.

    Thanks again!

    Jeff