• State Resolved
  • Locked Locked
  • Replies 7 replies
  • Answers 1 answer
  • Subscribers 63 subscribers
  • Views 1640 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.

Original question:

BQ25895: BQ25895 as UPS for Raspberry PI 3B+

BQ25895: PMID problems

Paul Beam1
Prodigy 200 points
Part Number: BQ25895
Other Parts Discussed in Thread: BQ25890

I have 2 problems with the PMID output on the bq25895:

1.  The output voltage occasionally drops for 5V to 3.8V (battery voltage is ~4V) which resets my system.  The voltage stays low for ~38mS, but the load processor resets, so I don't know if the drop in load causes the recovery or whether the 25895 recovers itself.  This is with no USB input, and running purely on battery.  The average current draw that I measures is ~380mA, but I am uncertain if there are spikes.  I don't see in the registers any record of  error.  How do I troubleshoot this?  The EVM does not have this behavior.  The 25895 does not appear to get hot.  

2.  When using a high voltage adapter for charging, PMID voltage tracks VBUS voltage!  So, if I apply 9V on VBUS, 9V appears on PMID, which is really bad.  This does occur on the EVM.  Is this intended behavior?  I did the measurement with no load, and I would be afraid to attach a load if this is the way things work.

Regards,

Paul

  • 0
    TI__Guru**** 166136 points

    Hi Paul,

    Regarding 2, yes.  When VBUS has a power supply applied, the IC is in charge mode.  There is only a reverse blocking, current sensing FET between VBUS and PMID so PMID will follow VBUS.  When there is no power supply at VBUS, the IC changes to OTG boost mode, where the buck converter operates in reverse as a boost converter and provides a 5V regulated voltage at PMID.  

    Regarding 1, the changeover from buck charge to OTG boost mode, after VBUS is removed, takes about 30ms.  So PMID will droop to a diode drop below the battery voltage on the EVM or your board.

    Regards,

    Jeff

  • 0 in reply to Jeff F
    Prodigy 200 points

    Jeff,

    #2 is a real bummer.  Based upon a reply to another question regarding a Raspberry Pi, it seems like a better solution for me to add a boost regulator that takes battery voltage input and produces 5V output.  I'm aiming for the behavior like a USB battery that always outputs 5V for attached devices.

    If I do the boost regulator, it will ultimately make #1 moot, however, I am not changing from VBUS=5V to VBUS = 0 -- that is, I am not plugging and unplugging USB.  I am simply on battery only, and periodically, the voltage drops for 40mS, which resets my system, and then the process repeats forever.  Unfortunately, it keeps the system from completely booting since it gets to a point and then the power supply drops, resets, and we start all over again.  It appears like it is trying to protect from overload.  Is there something I can measure or a register value what would indicate the 25895 has reset?  

    Regards,

    Paul

  • 0 in reply to Paul Beam1
    TI__Guru**** 166136 points

    Hi Paul,

    Assuming you have the recommended capacitance at PMID (~60uF) and the diode from SW to PMID, the boost converter should be able to provide up to 3A.  If you system has a transient load that requires more than that, the system will dip.  Also, if your connection from the battery to the BAT pin is highly resistive, a smaller surge might cause the voltage at BAT to dip below the undervoltage threshold for OTG operation.

    REG0C bit 6 is the boost fault bit.  You would have to read that bit during the fault, which might not be possible if the system fails.  If you have an oscilloscope and can capture PMID, BAT and either the load current or the battery current, I could diagnose the problem.

    Regard,

    Jeff

  • 0 in reply to Jeff F
    Prodigy 200 points

    I do not have a diode from SW to PMID.  Looking at the evm schematic, it shows D2 as "DNP."  

    I have 44 uF on PMID ( 2 x 22uF).  The evm schematic shows potential for 6 10uF, but 5 are marked "DNP."

    I can easily add another 22uF, and I'll see if I have an appropriate diode lying around.

     

  • 0 in reply to Paul Beam1
    TI__Guru**** 166136 points

    Paul,

    This schematic above is for the BQ25890 not the BQ25895.  The 895 schematic is further down in the user's guide.

    Jeff

  • 0 in reply to Jeff F
    Prodigy 200 points

    I'm an idiot!  Sorry, I did not see there were multiple schematics in this doc.  

  • 0 in reply to Paul Beam1
    Prodigy 200 points

    I added a diode and another 22uF cap, and so far so good!  I'm not sure what I will do about the high voltage charge issue as my device will surely be destroyed if PMID exceeds 5V.  However, this will get me through the demo!

    Thanks,

    Paul