• State
  • Locked Locked
  • Replies 3 replies
  • Subscribers 62 subscribers
  • Views 1026 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.

BQ24170 problems when attaching PSU (2-cell design)

Henrik Leon
Prodigy 40 points
Other Parts Discussed in Thread: BQ24170, CSD17313Q2

We've received quality complaints regarding a design where BQ24170 is used (EVM design is copied in all essential parts - same FETS, same values on other componetns). It has 2 cells and has a +12VDC PSU. Not all PCB's are affected, let's say 90% are, it is of course urgent to understand the problem. We're stuck right now and have spent almost a week not understanding the problem.

In analysis, we see that startup sequence is strange on problematic boards. ACDRV is not driven low as expected during BATFET -> ACFET switching, instead ACDRV stays at SRN level. We see a huge current rushing through ACFET during this 30 ms period - where ACFET should be OFF. ACDRV does not exceed AVCC voltage, which indicates +6V gate boost driver is not working.

Any ideas are appreciated, thanks.

Blue=VREF

YELLOW = AVCC

RED = ACDRV

GREEN = SW

Non-working sample:

Working sample:

  • 0
    TI__Guru 56650 points

    1. In order to tell ACFET is on or off, we needs check ACDRV and CMSRC at same time. I didn't see the CMSRC in the waveform.

    2. Also, what is external Cgs and Cgd value of ACFET (input FET)? If Cgd/Cgs is too big, the ACFET's gate may couple some voltage during the input source hot plug-in.

  • 0 in reply to Wang5577
    Prodigy 40 points
    1. Yes you are correct. thanks.
    2. The ACFET is the same as in EVM: CSD17313Q2, and we also use external Cgs and Cgd; 47n and 4n7.

    Big breakthrough this late afternoon. We finally discovered that the ACFET is blown on broken systems. They measure 2-3 Ohms between D and S, so no wonder we saw current rush during 30 ms - the current from the PSU rushed to the battery until the power switcher disabled the BATFET.

    I now have to figure out what killed the ACFET. Could it be that Cgd is too big (despite being identical to the EVM), which couples a few volts as you helpfully point out, thereby accidently enabling the ACFET on some boards at hot plug-in? Maybe that would create a current rush destroying ACFET in the process?
  • 0 in reply to Henrik Leon
    TI__Guru 56650 points

    Great, Yes. Please reduce Cgd to 1nF that can reduce the ACFET loss during input source hot plug-in.