• State
  • Locked Locked
  • Replies 10 replies
  • Subscribers 62 subscribers
  • Views 168 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.

BQ24650: Two-layer PCB unstable

Chris Tanner
Prodigy 10 points
Part Number: BQ24650


Tool/software:

Hi, I have had this two-layer PCB layout running fine for a year or so, but the most recent batch of PCBs has a problem where it goes into an unstable mode:

- the STAT1 LED flashes briefly once every two seconds
- BTST net is as shown on the trace during the charging phase (which cycles on and off)

Charging li-ion at a target of 16.8v, using a 21V 10W solar panel, at 1A max current. While the battery is low voltage and the charge current is high, everything seems OK, until some point (which varies) during charge when it becomes unstable.

I am OK to make modifications to the PCB layout if needed, but would need to know the source of the problem so I can change the appropriate section. Or even another component I can retrofit to existing PCBs to ensure stability.

Gerber attached.

Unstable on BTST, approx 16v at peak.

GerberWM-V12.zip

  • 0
    TI__Mastermind 35396 points

    Hello Chris,

    Can you provide your schematic?

    Best Regards,

    Christian

  • 0 in reply to Christian Moyer
    Prodigy 10 points

    Yes, here you go.

    The 2.5mR resistor is for current sensing elsewhere.

    LEDs are connected from VREF to the STAT lines with a 200R resistor.

  • 0 in reply to Chris Tanner
    TI__Mastermind 35396 points

    Hello Chris,

    Chris Tanner said:
    Hi, I have had this two-layer PCB layout running fine for a year or so, but the most recent batch of PCBs has a problem where it goes into an unstable mode:

    Have you taken the IC off a working board and placed in on a non-working board, to confirm issue follows IC or board? Is there any difference in this latest batch of boards?

    Best Regards,

    Christian.

  • 0 in reply to Christian Moyer
    Prodigy 10 points

    Hi Christian

    I don't have the facilities to remove and refit this IC unfortunately. I have ascertained some more information:

    - Placing a finger on the PCB in *just* the right place, somewhere over the IC and nearby components, restores stability, and I get the "charge complete" LED coming on. I can't repeat this predictably, and don't think it would be useful to spend the time doing so! It's a capacitive fix; there are no loose joints.

    - I had one previous batch PCB which was not showing problems, but now is ... which led me to think that perhaps there was a voltage spike from a power supply in place of a solar panel which has damaged the IC. But I've tried one of the new batch from the new PCB stack, which has never had a power supply attached, and it exhibits the same problem.

    So in summary it appears that all of the new boards so far have the problem, and it's possible to get old boards to fall into the same problem.

    The new boards are the same schematic apart from I have replaced a coin cell CR2032 (for an RTC) with a LP2985-33DBVR regulator, so we can use the primary battery for maintaining RTC. The instability remains even when I isolate this section of the circuit, so I'm pretty sure it's not that.

    The PCB layout for the charging components (and everything nearby) is the same as per the previous version.

  • 0 in reply to Christian Moyer
    Prodigy 10 points

    Potentially another clue. The original first-batch board has gone back to being stable. Temperatures and other parameters are the same as yesterday. TC pin is 2.1V on both the stable and unstable boards.

    When I put the battery under load, the new boards go into stable charging.

    Battery open circuit is approx 16.2v
    Battery in unstable mode is approx 16.0v, with a pulse of ~17.1v every 2 seconds 
    Battery under load [stable charging] rises to 16.4v and continues to charge while under load

    Attached stable and unstable waveforms on BTST in case that's useful.

     

  • 0 in reply to Chris Tanner
    Prodigy 10 points

    The 17.1v gave me a clue which may point at the problem, the new batch has a 5% tolerance on a resistor which makes up the charge-complete set point on VFB and I think it's interacting with a cut-out in the BMS which is built in to the battery pack. Previous batch had a 1% tolerance. Will report back.

  • 0 in reply to Chris Tanner
    TI__Mastermind 35396 points

    Hello Chris,

    Chris Tanner said:
    The 17.1v gave me a clue which may point at the problem, the new batch has a 5% tolerance on a resistor which makes up the charge-complete set point on VFB and I think it's interacting with a cut-out in the BMS which is built in to the battery pack. Previous batch had a 1% tolerance. Will report back.

    Thank you for the update, please let me know if changing this resistor resolves the issue.

    Best Regards,

    Chistian

  • 0 in reply to Christian Moyer
    Prodigy 10 points

    I have confirmed that all the resistors in the potential divider are less than 1% off target value. However, when in circuit ... 16.88V on the battery gives only 2.03V on VFB, which is causing the IC to try and continue charging, up to the 17.1V point where some protection kicks in, or something else causes instability.

    So I shorted out one of the 10k resistors (R36) to give a pot divider of 690k/100k for VFB and now the system is unstable with the peak charge pulse at 16.8v, which suggests it's not related to any BMS which is in the battery pack.

    Changing R15 (100k) to 110k stabilises the circuit with STAT2 going low (charge complete) at 15.5V which would be expected if 2.1V on VFB is the set point. But clearly not enough to charge the battery to full.

  • 0 in reply to Chris Tanner
    TI__Mastermind 35396 points

    Hello Chris,

    Chris Tanner said:
    So I shorted out one of the 10k resistors (R36) to give a pot divider of 690k/100k for VFB and now the system is unstable with the peak charge pulse at 16.8v, which suggests it's not related to any BMS which is in the battery pack.

    Do you see any changes, if you add a 100nF decoupling cap near the FB pin?

    Best Regards,

    Christian

  • 0 in reply to Christian Moyer
    Prodigy 10 points

    The decoupling cap appears to make the "charge complete" light come on, but charging does not happen even with a low battery.

    I have seen significant success now over a few of the new batch simply by replacing the low-tolerance resistor with a more accurate one, so it *seems* to be down to that. I will keep monitoring and trying various batteries and chargers / solar panels and report back if the resistor fix is not the final solution. Thanks for your help.