• State Resolved
  • Locked Locked
  • Replies 6 replies
  • Answers 1 answer
  • Subscribers 63 subscribers
  • Views 382 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.

BQ76952: Cannot balance cell 1 with default VC0 configuration

Christian Pavkovich
Prodigy 30 points
Part Number: BQ76952

Hello,

Issue up front: the default recommended connection for VC0 does not work for external balancing of cell 1.

I am having an issue with cell balancing on a design using the BQ76952. The design has external balancing as described in the SLUAA81A App Note. I mistakenly used 1k ohm filter resistors instead of the 100 ohm recommended values as I was following the SLUA420A App Note. That being said, I don't believe it is contributing to the issue I am having.

The cell balancing circuit for C1 was powered while the rest of the design was not powered by providing a voltage only across V1 and V0. The VC1 and VC0 nodes were shorted together to mimic the internal FET closing on the AFE (which was unpowered at the time). It was observed that the voltage Vgs was lower than expected. After a time it was discovered there was an additional element to the divider in the form of the resistor from VC0 to GND that is called out on page 66, 67, and 75 of the BQ datasheet.

This resistor as stated makes it much more difficult to balance the lowest cell during operation, and is strictly different than all other cells in this way, as imbalanced cells above cell 1 do not have an additional element in the divider. Is the resistor to GND necessary as there is already a filter resistor from VC0 to C0? As far as I can tell I am mimicking the design as expressed in the datasheet, but please correct me if I am wrong. GND is BATT-, C0 and BATT- are shorted externally on the harness battery-side.

I do not understand how the balancing is supposed to function if this is the required configuration.

  • 0
    TI__Mastermind 42060 points

    Hello Christian,

    That resistor (R101) seems to be actually redundant. This in essence would just be in parallel to R58. You can remove R101 since you already have R58.

    C0 should essentially be referenced to BATT-, so you already have this resistor (R101) in place connected to C0/BATT-. You do say you have it connected externally, which is okay. I would recommend to add a Schottky diode and Zener on the VC0 pin directly to protect it against voltage transients.

    I'd also recommend reading our Cell Balancing With BQ769x2 Battery Monitors application note to understand how the balancing can be configured with the BQ76952.

    Now, unrelated to that, 1-kOhm exceeds our recommended values, this may cause cell voltage measurements to be less accurate. Our recommended max of 100-Ohm ensure datasheet accuracy specifications.

    Best Regards,

    Luis Hernandez Salomon

  • 0 in reply to Luis H. S.
    Prodigy 30 points

    Luis, thank you for your reply. Are you familiar with the internal pin structure of the BQ76952? I removed the resistor R101, however with the harness connected (this shorts C0 to GND externally) I am reading a resistance value of 527 ohms from VC0 to GND with the AFE unpowered. I would expect this to be 1k ohm since the only connection is now R58 from VC0 to C0 - which is ultimately GND on the harness side.

    Do you have an idea where the unexpected resistance value is coming from?

  • 0 in reply to Christian Pavkovich
    Prodigy 30 points

    I realize I'm using GND instead of BATT-, I should have said BATT- for each instance.

  • 0 in reply to Christian Pavkovich
    TI__Mastermind 42060 points

    Hello Christian,

    The pin should be Hi-Z when unpowered if I remember correctly, I measured on one of our internal boards from VC0 to BATT- and measured the resistance accurately (~100-Ohms, since this is what we recommend)

    I'd check everything connected in that path to see if there's any unintended resistance somewhere. For there to be ~500-Ohms, there may be a 1-kOhm parallel resistance to the VC0 1-kOhm resistor.

    If you'd like, you can send me the schematic on private messages so I can review it and see if I notice anything.

    Best Regards,

    Luis Hernandez Salomon

  • 0 in reply to Luis H. S.
    Prodigy 30 points

    We were able to remove R101 on a different board and attempt the same test which has worked hardware-wise. It's looking like the unexpected additional path to BATT- might be an internal AFE issue (damage, etc.).

    The way I am understanding it, the datasheet is recommending a similar filter from VC0 to BATT- if your bottom-of-stack connection to GND is made on the PCB side, but since our connection C0 to GND/BATT- is made on the harness side this extra resistor is unnecessary. 

    This connection, for example:

    The only remaining question I have at this time is: is there an explanation you have for the unexpected resistance to BATT- we can measure from VC0 to BATT-? All connections besides the AFE have been removed and it is being measured at the AFE pins. Is there a fault that could cause this, maybe overvoltage?

    Regardless, we have gotten balancing to work on a different board and you were able to clear up some confusion I had with the AFE.

  • +1 in reply to Christian Pavkovich
    TI__Mastermind 42060 points

    Hello Christian,

    The image you show only has a single filter resistor, you had two in parallel.

    This resistor should essentially connect only to the BATT- sensing side (where you want to measure the voltage). 

    Since ground and VC0 are not on the PCB, I'd recommend to add the Schottky diode and Zener to the VC0 pin to protect it from voltage transients (that may occur during high loads due to parasitic inductances in the traces/wires).

    Christian Pavkovich said:
    is there an explanation you have for the unexpected resistance to BATT- we can measure from VC0 to BATT-? All connections besides the AFE have been removed and it is being measured at the AFE pins. Is there a fault that could cause this, maybe overvoltage?

    If the part was completely powered off and R101 was removed, there shouldn't be any additional resistance in the path. The part may have been damaged as you mentioned, since you tested on a different device and it seemed to measure okay. The only time it may measure differently may be during cell balancing, since VC0 is pulled high during this time.

    Hope this helped clarified things, and let me know if you have any more questions Slight smile.

    Best Regards,

    Luis Hernandez Salomon