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.

TPS61260: Possible Ground Connection / Re-work issue ?

Part Number: TPS61260

We have a problem with some 10% of boards that have a circuit featuring the TPS61260.

I think my question may boil down to possible ground connection issues.

I would also like to know what is the Max IFB for this part (not sure that this is actually specified on the data sheet) ?

I would welcome feedback from anyone who has had experience of using this part in a production environment, and in particular has undertaken re-work of if it in-situ ?

Although the particular board is not our design / layout, and they have been assembled by a contractor, having had a look at these there is nothing obviously wrong with the circuit / layout, or indeed the assembly of same.

Out of an initial (pre-production) batch of c 50 Pcs, we have 4 Pcs that have failed.

The input is 3.0V nom, with a 170K / 1M divider to give 3.5V nom output, and the max average output current set at 75mA nom (although typically only drawing 10mA max) with RI at 27K, (+ 4.7uH inductor, plenty of capacitance both input and output).

The good boards all put out 3.5V - as designed.

But for the bad boards the output is around only 1.6V.  We tried changing the TPS61260 for a fresh one, and the output voltage then came up to 2.2V, (also tried isolating the output first - but didn't make much difference).  I actually got one of our experienced production people to change the device, as it looked really difficult to re-work with the ground pad underneath the package !  For the 1.6V output I measured only around 0.2V on the feedback pin (wrt to the PCB ground plane).  Does this suggest the problem we are seeing is with the ground connection of the device ?

Any thoughts or comments from anyone who has experienced similar issues ?

Thanks in advance.

  • Hi Alan:

    Please share the schematic with me, it will be more clear about the surrounding components. 

    May you share the waveform of the good and bad boards? Including the Vin, Vout, L.

    Here are some points that I suspect:

    1. Improper RI: Ri should be 2k-20K

    2. Weak power supply. What's the power supply? Does it have enough current capability?

    3. Layout and soldering related issue. 

  • Hi Xie

    In answer to your questions:

    1. Improper RI: Ri should be 2k-20K

    Oops this was a typo on my part, I meant to write 2.7K - maybe I should have written it as 2K7 - sorry.

    2. Weak power supply. What's the power supply? Does it have enough current capability?

    We're using a (TTi) bench power supply for test purposes - current limit set to 200mA - way higher than what it needs - typical current draw is actually in the 10-15 mA range.

    3. Layout and soldering related issue. 

    I would be surprised if the layout is an issue.  But I do suspect the soldering (of the ground connection) on the failed boards could be the problem.  For those four boards that have failed test, I have doubts as to our ability (i.e. kit) to successfully & reliably re-work these boards by replacing the TPS61260 with a new device.  But even though our tools may be imperfect (for this re-work), the fact that when we tried this (replacing the TPS61260 with a brand new device), we observed an output of 2.2V after replacement (as opposed to 1.6V before), should tell us something.

    Also - as mentioned previously on the units giving a 1.6V output - I measured (in round figs) 0.2...V on the feedback pin (which is consistent with our FB divider ratio) - its just not the 0.5V that it is supposed to be. 

     

    I was hoping for some comments from other TI customers who may have experienced issues in a production environment, and also what they recommend as a means to manage any issues.

    Regards

    Alan.

    P.S.  I will try attaching a photo and a schematic - but not sure how readable this will be.

     1054_001.pdfP.S.  I will try attaching a photo and a schematic - but not sure how readable this will be.

  • Hi Alan:

    May you share the waveform? Both good and bad boards, including the Vin, Vout, L.    It would be helpful to deeper analysis. 

    And please allow me reply after the weekend ...Thank you~ 

  • Hi Minqiu

    Unfortunately I do not have anything I can send you.  I'm afraid our production facility is a bit light on engineering kit.  That said I did try probing with an old analogue scope but could see nothing interesting, even on highest res (10M probe).  I suspect I would have to pull off some of the input and output caps to see anything.  The one place that I think would have been really interesting would be to probe between the ground plane and the ground connection actually on the TPS61260 - but of course there is absolutely no way to get to that when it is in-situ.

    [Of course - if the device was good, and soldered in place properly, it should be at the same potential as the ground plane. ...]

    So far, we have just played around with changing components on one of the faulty boards, even though we suspected the issue was the TPS IC, we went for the easy stuff first, such as output cap, inductor, current limit resistor etc, replacing with new unused parts.

    My only other thought (bit of a sink or swim one) - is perhaps rather than removing the TPS61260 from the other failed boards, maybe we try heating it with a Metcal hot air solder blower to see if we get lucky and get the ground connection to reflow, but somewhat risky - too little and it won't work, too much and we may cook the device.  I am also assuming here that the device is good and the problem is with the ground pad solder connection - but it may not be.

    Kind Regards

    Alan.

  • Hi Alan:

    Sorry for keeping you waiting. Full time in lab these days.

    I still prefer to check the waveform even not interesting. It's really a key to analysis.  Here is a blog, which may help you to catch the waveform.   https://e2e.ti.com/blogs_/b/powerhouse/archive/2016/11/29/tips-in-debugging-a-boost-converter

    According to your photo, it see there is some soldering flux. I heard some cases caused by such things. Some soldering flux is not insulate actually.

    And for such package, the hot air is recommended for manual soldering.

  • Hi Alan:

    As long time no hear from you, I assume you have solved the issue. If not, just reply below. Thank you!