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.

BQ25792: Rapid VBAT OVP events following battery disconnection

Part Number: BQ25792
Other Parts Discussed in Thread: BQ40Z50, , BQ25798

Hi there—our platform uses the BQ25792 to charge a pack which is managed by the BQ40Z50. During various fault scenarios, the BQ40Z50 may open the charge FET while charging remains active, resulting in the following sequence:

  1. BQ40Z50 charge FET is disabled, effectively disconnecting the battery.
    1. BQ25792 triggers a VBAT OVP event, and disables the boost converter.
    2. VSYS falls to the cell voltage minus the drop of the charge FET body diode.
  2. BQ25792 re-enables the boost converter and triggers another VBAT OVP event, over and over.
  3. BQ40Z50 charge FET is re-enabled automatically to protect its body diode.
  4. BQ40Z50 charge FET is disabled again.

This sequence is depicted in the following plot of VBAT / VSYS:


Under normal circumstances, our applications processor detects (1) over I2C and deasserts the /CE pin of the BQ25792 as quickly as possible. In this case, VSYS is regulated to VSYSMIN (3.5 V). However, this intervention is not immediate, and our post-regulation stage driven by the BQ25792 briefly experiences the low voltage seen during step (1-b). Depending on the magnitude of this voltage, it may trigger an under-voltage condition and cause our platform to shut down.

It would be preferred if the BQ25792 can automatically output VSYSMIN in this condition, as opposed to relying on the host to disable the charger as fast as possible. After browsing other E2E threads, I noted the pin-compatible BQ25798 has been recommended in response to this problem.

However, I noted page 33 of the BQ25798 datasheet has the same warning as the BQ25792 datasheet:

No battery or battery removed while charge enabled and TS pin function is disabled - The charger continuously tries to charge the BAT capacitance, typically resulting in the BAT voltage alternating between a low level and BATOVP fault. The SYS voltage follows the battery voltage up, potentially reaching SYSOVP fault. In order to determine if a battery is attached, the host must periodically disable charge, force IBAT discharge current and then read the ADC BAT pin voltage. Alternatively, the host can monitor the INT pin for rapid interrupts and then read the charge status bits for fast (​<1 s for typical BAT pin capacitance) toggling between charging, taper and termination.

My question is—can the BQ25798 solve the problem I have described here, and if so, what is its behavior during this scenario?

Thank you in advance for your support; in case I can clarify any of my observations, please let me know.

  • Hi Jeff,

    Both BQ25792 and BQ25798 have BATOVP events when battery is removed and charge is still enabled (by CE bit or pin and no TS fault due to missing thermistor).  Unlike the BQ25792, the BQ25798's converter does not shutdown for 30ms due to the BATOVP event.  This shutdown results in SYS voltage collapse if there is a load on SYS.

    The only other choice is to prevent the OVP event by adding at least 200uF capacitor to BAT/BATP.  You might be able to reduce that capacitance slightly for your specific VBUS and BATREG but most likely not below 100uF.  This is after derating for temperature and applied voltage.10x10uF ceramic caps will not be enough due to their derating issues.

    Regards,

    Jeff

  • Hi Jeff—thank you for your prompt support!

    Just so I understand—is this to say that the BQ25798 buck/boost continues to run during a VBAT OVP event, and we will continue to see a steady VREG + 200–400 mV on VSYS in the scenario I have described? If so, what protects the battery during OVP events in which it remains connected—perhaps the BATFET opens during this case?

    Or, are you saying that the BQ25798 buck/boost will still shut off during a VBAT OVP event, albeit at a longer delay (30 ms)? When I zoom into my capture, it looks like the BQ25792 buck/boost is trying to restart every 10–15 ms.

  • Hi Jeff,

    Yes.  SYS will have a triangular ripple between VBATOVP and SYSMIN.  The charger's VBATOVP threshold is 103% x BATREG.  Your battery pack protector (I am assuming you are not trying to charge bare cells) has a slightly higher setting for over voltage and will open the pack charge FET at that voltage. The 798's converter does not shut off but the 792 converter does shut off.

    Regards,

    Jefff

  • Hi Jeff—thank you for the update; I am aligned with you. We will continue to study this problem, and consider the BQ25798 as a replacement.