• State
  • Replies 4 replies
  • Subscribers 62 subscribers
  • Views 59 views
  • Users 0 members are here
Support feedback
Options
Options
Related

BQ25756: BQ25756: Failing to Enable Charge from Battery Overdischarge

Jaden Reimer
Prodigy 10 points
Part Number: BQ25756


Hi TI support,

I'm working with a BQ25756 charge controller design that enables charge automatically in hardware (no FW involvement required) so a battery can recover from overdischarge by solar power.

I'm troubleshooting a design bug impacting 10% of my assembled boards. 

Nominal boards: when I apply solar input voltage with a overdischarged or open-circuit battery, the regulated battery output (Vbat) quickly rises above VBAT_SHORT=3V and enters pre-charge. Vbat enters a nominal regulation loop.

Failing boards: Vbat never rises above 1.5V (always less than VBAT_SHORT) and fails to automatically enter a charging state. It seems to never exit trickle charging. When I kick-start it by applying 4V to Vbat the charge control enables and operates independent of FW.

Notes:

  • Nominal and failing units both charge nominally (as enabled by hardware) when battery voltage is present or when firmware is in the loop controlling the charger.
  • Vbat resistance is similar across units.
  • Failing units with only a discharged battery connected float to 1.53V on Vbat. Nominal units float to only 53mV.

Questions for your team:

  1. What is the controller's nominal behavior in trickle charge (specifically w.r.t switching operation and the value/control loop for IBAT_SHORT)?
  2. What might explain Vbat failing to rise when solar input voltage is applied?
  3. What could explain why this issue varies board to board?

Please let me know any additional information you need to troubleshoot this issue. Thank you for your support.

  • 0
    TI__Genius 9710 points

    Hello,

    1. The charger will charge at around 150 mA for trickle charge.
    2. If the trickle charge current is too low or the leakage on the output is too high (or a combination of the 2) which prevents the output voltage from rising.
    3. The variations in the trickle charge or the output leakage could explain variation.

    Is it possible to force the battery protection FETs open? This could bypass the trickle charge phase and enter the precharge phase.

    Best,
    Michael

  • 0 in reply to Michael Bradbourne
    Prodigy 10 points

    Thanks for the response. Unfortunately the BMS FETs can only be re-enabled in this configuration from a "charger present" condition, meaning the charger needs to present a voltage to the battery terminals before the BMS will re-engage the discharge path.

    New findings:

    I swapped the charge controller chips between a board that was properly charging and one failing to charge a battery out of the BMS over-discharge state:

    • The board that failed to auto-charge (with battery terminals open circuit) at 20V solar input now charges. Trickle charge to pre-charge transition confirmed by scope capture.

    • The board that auto-charged (with battery open circuit) at 20V solar input now fails to charge (charge failure follows the chip). I confirmed failure to exit trickle charge by scope capture. The high-side input (buck mode) gate pulses are all short (~40ns) and it only attempts trickle charge with short pulses for 400us before stopping switching entirely. Other failing boards attempted to charge for much longer.

    My latest hypotheses on what is going on here:

    • Chip to chip leakage variation or some other parameters are responsible for the difference in behavior. (With 1V applied to the FB pin, board unpowered, a failing board = 1.6-1.8uA draw, passing board = 3uA draw). FB divider resistances measure the same between boards.

    Could you please share your comments on these findings and request any other information that would help troubleshoot this issue?

    It's entirely possible there is still a PCBA-level explanation here, but that doesn't explain why a failing board is now successfully charging with a different BQ25756 chip. I also reflowed a failing board to isolate solder reflow assembly issues as a factor (still failed afterward).

    I will continue to trace down where the trickle charge current is going in the system since it isn't floating Vbat up as expected.

  • 0 in reply to Jaden Reimer
    TI__Genius 9710 points

    Hello Jaden,

    This is interesting behavior. I do not know where on the IC the current would be leaking out. The device should regulate the current from the sense resistor, and there shouldn't be any pin connections on following the sense resistor in the power path.

    Are you using hardware to set charge current or register settings?

    Are you able to use the registers to lower the charge current and then disable precharge to get the battery protection FETs open at first?

    Best,
    Michael

  • 0 in reply to Michael Bradbourne
    Prodigy 10 points

    Hi Michael, thanks again for the input. I set R_ILIM_HIZ = 2k and R_ICHG = 5k with ACP/N shunt = 5mOhm and SRP/N shunt = 5mOhm. If I calculated correctly, this should result in default HW limits of 10A for input and output current. These are the charger limits on startup in a discharged battery condition since the system (and I2C control interface) is not yet powered on to configure FW application limits.

    As for register configuration, per above I need to assume the system (I2C) side is unpowered. I'm relying solely on charge controller HW configuration to resume charging and recover system power.

    I will continue to update with any new findings. In the meantime please share any other theories/findings on your end.