Other Parts Discussed in Thread: BQ76930, EV2400, , BQSTUDIO, BQ78350
Tool/software:
Dear TI Support,
I am currently working on a fuel gauge implementation using the BQ78350R1 in combination with the BQ76930 AFE. I am performing a full learning cycle following the SLUA903 procedure to achieve accurate SoC tracking.
My system uses a 6-cell (6S) Li-ion battery configuration.
Here’s what I have observed:
- The battery was fully charged and the FC flag was properly set.
- The individual cell voltages at full charge were approximately:
- Cell 1: 4187 mV
- Cell 2: 4185 mV
- Cell 3: 4169 mV
- Cell 4: 4170 mV
- Cell 5: 4180 mV
- Cell 6: 4175 mV
- The gauge shows a valid FCC value (e.g., 4144 mAh) under Data Memory > Gas Gauging.
- The discharge cycle was completed until EDV2 was reached, and both TD and EDV2 flags were set.
- A relax period of over 60 minutes was performed after discharge, with current below the Quit Current threshold.
- However, the Update Status remains at 0x00, and the FCCX flag is still active, even though the Cycle Count has incremented to 1 and a valid FCC value appears to have been calculated.
**Important additional notes:**
- Between each charge and discharge cycle, I issued a software reset using the ManufacturerAccess command `0x0041`.
- I also pressed the BOOT button on the evaluation board multiple times throughout the learning cycle process.
- I am now wondering whether these actions may have disrupted or invalidated the learning process.
**Test setup details:**
- I am using the official BQ76930EVM evaluation module together with the EV2400 interface.
- Communication is handled via BQStudio (latest version) on a Windows 10 PC.
- Actual 6S Li-ion cells are connected to the BQ76930EVM (no simulator).
- A precision power supply and electronic load are used for controlled charge/discharge.
- I monitored all relevant flags (FC, TD, EDV2, FCCX, Update Status) during the process.
**My main questions are:**
1. Shouldn’t the Update Status move to 0x06 if FCC and Qmax have been updated?
2. Could any condition (such as the PRES pin, insufficient relax duration, or configuration setting) prevent the gauge from completing the learning cycle?
3. Would software resets and pressing the BOOT button during the process invalidate the learning cycle?
4. Is this behavior expected in a 6S configuration with the BQ76930?
If needed, I can also provide the exported `.gg.csv` file from BQStudio for further analysis.
I want to ensure that the gauge has fully completed a valid learning cycle and that SoC estimation will be reliable in operation.
Thank you in advance for your support.
Best regards,
