BQ28Z610-DEVICE-FW: Learning cycle completed but there's errer in generating Golden Image File

Hello Alan,

I hope you're doing well.

Regarding my application, I’m using this IC for battery management, and my goal is to keep the FETs always ON. When a charger is connected, the system should enter charge mode, and when power is removed, it should switch instantly to discharge mode—similar to how batteries operate in laptops.

Before starting the learning cycle, I attempted to manually send the FET_EN command to turn on the FETs. The IC functioned correctly in both charge and discharge modes. However, cell balancing was not active at that stage, which I understand is expected prior to completing the learning cycle.

My main concern is:
In the final application, after completing the learning cycle, do I still need to manually send commands (FET_EN, CHG, or DSG) to turn on the FETs, or will the IC automatically manage FET control based on its internal logic?

My expectation is:
Once the learning cycle is complete and the FETs are enabled initially, the IC should autonomously control charging/discharging and balancing, while I only read status and SoC information from the gauge.

I’ve completed the learning cycle successfully, and as described in TI’s documentation, cell balancing has been activated, which confirms the learning process was successful.

However, I encountered an error when attempting to generate the Golden Image files. Could you please help me review the error I faced during this step?

Thank you for your support.

Hi Alan,

Thank you for sharing that information.

Regarding the operation mode, can you clarify when the FETs should work automatically? In what scenarios does this occur, or do I need to send a command first?

I've sent two commands (0x20 for DSG_FET_TOGGLE and 0x1F for CHG_FET_TOGGLE), and now the gauge is functioning as expected. Is this the correct approach?

Additionally, do you have any ideas on how to resolve the issue with generating the golden image files?

Thanks!

Hello alan,

Thank you for the clarification — the operation is now clear. Could you please specify which parameter thresholds I should set to ensure the FETs close automatically in both charge and discharge modes?

Regarding the golden image, I followed the same steps as shown in the picture, but I encountered the following error message:

"Golden Image: Reading the data memory or flash image has failed. The golden image files could not be created."

Could you advise on how to resolve this issue?

Thank you again for your support.

Hello Tarek,

We will continue to work on your question, today is a national holiday, Memorial Day.

Thank you,
Alan

Hello Alan,

Thank you again for your valuable input.

I reviewed the TRM and proceeded to enable automatic FET control by sending FET_EN and GAUGE_EN commands. I also adjusted the Chg Current Threshold and Dsg Current Threshold to 10 mA, then to 20 mA.

Here’s a summary of my observations:

• XCHG works as expected and activates automatically when the charger is connected.

• XDSG does not activate automatically when the charger is unplugged, even with an electronic load drawing more than the DSG threshold.

• When I disconnect the charger, the IC powers off completely and BQStudio disconnects.

• I also tried a 150Ω load resistor. Same result: unplugging the charger shuts off the IC and disables discharge.

• When both charger and load resistor are connected, XDSG activates once the current drops below the charge threshold and the device enters Relax mode.

• However, once this happens, the IC begins restarting every ~10 seconds. If I unplug the charger during this period, discharge stays active only until the next reset, after which the IC shuts off again.

It seems the IC cannot stay powered from the battery alone in this condition, causing unstable discharge behavior.

Do you have any insight into what might be causing the resets or how to maintain stable discharge after unplugging the charger? If needed, I’d be glad to arrange a quick call to review this together.

screenshot below - when Discharge mode active

Hello Alan,

Thanks for your support.

I’ve attached the schematic showing the BQ28Z610DRZR circuit.
The battery pack is a 2S1P Li-ion configuration (7.4V nominal, 8.4V max, 3000mAh).

Let me know if you need anything else.

Extracted pages from 3_Extension Tinker V1.0.pdf

Hello Alan,

Thank you for your review and feedback.

I’ve removed the FET and resistor (Q8 and R66) between the battery and the gauge, and replaced them with a direct solder bridge. Originally, the FET was added for reverse polarity protection during battery insertion, but I’ve now eliminated it to prioritize stable operation of the IC.

However, I’m still observing the same issue — the gauge powers down every ~10 seconds.

One more thing I’ve consistently noticed: when I connect the gauge IC via I2C and open BQStudio, the software doesn’t detect the device unless I also connect the charger. This happens even though the I2C bus driver is powered from an external source.

Please let me know your thoughts on this behavior or if you'd recommend any further checks.

Hello Tarek,

Your update has been received and we are working on your response.

Thank you,
Alan

Hello Alan,

Yes, I confirm that I’m still observing the same issue with the same firmware setup. When I connect the gauge IC to the EV2400 and power the I2C bus externally, BQStudio doesn’t detect the IC until the charger circuit is connected — only then does the IC power on and BQStudio detects it automatically.

For the charger circuit, I’m using MCP73842T-840I from Microchip. I had a detailed discussion with their support team, and after several tests, we confirmed that the charger circuit is functioning correctly. I’ll share the schematics with you for reference.

At this point, my main concern is understanding why the gauge IC isn't powering from the battery directly. I’ve measured VC1 and VC2, and the readings are within expected values.

Charger schematics: 

Hello Tarek,

Due to workload we are asking more time to work through e2e.

Thank you,
Alan

Hello Alan,

Thank you for the update — no problem at all. I understand the workload and appreciate your continued support.

Looking forward to your feedback whenever you're able to get back to it.

Best regards,
Tarek

Hello Alan,

I’ve observed something important that helped identify the root cause.

When the battery is inserted, the gauge IC does not power on by itself from the battery, as it previously did. I checked the PBI (Power Backup Input) pin — a capacitor is connected as recommended, but the voltage at this pin reads 0V.

To investigate further, I supplied power directly from the battery to the PBI pin, which successfully powered on the gauge IC and allowed BQStudio to detect it. However, after about 9 seconds, the IC powered off even though PBI remained connected. It stayed off for around 3–4 seconds, then powered back on — and this cycle repeated continuously.

This behavior matched my earlier test results with an electronic load and charger, where the IC entered discharge mode, then cycled off and on at the same interval — even though no power was applied to PBI during that test.

I also noticed further instability: in one test, the gauge IC did not power on or get detected by BQStudio when the charger was connected. Probing the 10Ω series resistor connected to PIN 8 (PACK) showed a voltage drop of 0.2V to 0.45V, indicating 20–45 mA of current. During this period, the IC became noticeably hot. As the datasheet notes, the internal 1.8V LDO has a short-circuit protection limit of 20–50 mA, which leads me to believe this IC may be damaged.

The good news is that I tested another board from the same production batch, and it operates normally — powering up directly from the battery and recognized by BQStudio without any external supply. I will proceed with repeating the learning cycle on this new board.

Since the issue appears to be due to IC damage, I will hold off on sharing the .log file for now.

Best regards,

Hello Alan,

Thank you for the support.

I’ve replaced the battery and successfully completed the first learning cycle to identify the best CHEM ID. I submitted the data to the GPC online tool and received a report with the recommended CHEM ID. However, I couldn’t find this ID listed in the Chemistry tab in BQStudio.

I saw that there’s an option to program the gauge IC using a GPCRB file, but I’m not sure how to generate it. Could you please advise on how to create the GPCRB file or how to program the gauge IC with the recommended CHEM ID?

Appreciate your guidance.

8468.GPCPackaged-report.zip

Hello Alan,

I hope you had a great weekend. Do you have any updates on my previous question related to CHEMID and programming the GPCRB file?

Thank you for your support

Hi Alan,

I’ve successfully performed the ChemID setup on a new board, and everything was proceeding smoothly during the learning cycle. I configured the data memory properly and began the full charge phase. Once the battery was fully charged and the relaxation phase completed — with Update Status and Qmax updating correctly — the system was ready for the discharge phase. However, as soon as I unplugged the charger to begin discharging, the fuel gauge IC lost power again.

I suspect this issue may be related to changes in the Bit Registers. Initially, the FETs behaved as expected after sending the FET_EN command early on. After the full charge and relaxation stage, the RDIS bit cleared, OCV was captured, and the system was ready to update the resistance table during discharge. Because the IC lost power at this point, I had to manually disable FET_EN and send the DSG_FET_TOGGLE command to turn on the DSG FET. This worked, and I confirmed that the Ra Table was being updated .

However, once the discharge reached the termination voltage, the DSG FET turned off and the IC lost power again. As a result, the final relaxation stage couldn’t be completed.

I also noticed something that might help diagnose the issue. Initially, when the IC was functioning properly, I measured the voltage on the PBI pin (Pin 10), and it matched the voltage on VC2 (Pin 11). But when the power loss occurred, the voltage on PBI dropped to 0V or a floating value (around 0.5V). This suggests that the internal connection between VC2 and PBI might be affected — potentially due to register configuration.

Could you please help me resolve this issue? It’s been quite time-consuming and is preventing successful completion of the learning cycle.

Thanks again for your continued support,

Hi Alan,

Thank you so much for your reply and for sharing the links.

In the GPC report I received, the recommended ChemID (with the lowest deviation) wasn’t available in the chemistry list, as you mentioned. However, the report also included a list of alternative ChemIDs with acceptable deviation (less than 3%).

I selected one of those ChemIDs and successfully programmed the fuel gauge with it.

Thanks again for your support — this issue about CHEM ID is resolved.

Best regards,
Tarek