BQ40Z50-R2: The learning cycle cannot be completed using the CHEMID we applied for.

Hello,

Can you please share the .gg file of the gauge settings as well at this time?

Regards,

Anthony

Hi Anthony,

Attached the gg file for you review!learning Cycle 20250516.gg.csv

Hello Rain,

We have received your update and we are working throughs working on your response.

Thank you,
Alan 

Hi Rain,

Just to confirm, what is the cell orientation being used here? Is it a 1S pack?

Regards,

Anthony

Hi Anthony—thank you for the update! I can confirm this is a 1S4P pack. In case we can provide any additional information, please let us know.

Hi Jeff,

Thank you for confirming, I believe the issue here is coming from the DOD0 measurement taken after the discharge to empty. For the gauge to complete the initial Qmax update, the distance of the DOD0 measurements taken before and after the charge period must be at least 90% of the total range. In the case of the 7963 ID, you can see that at its highest, the DOD0 measurement after the discharge to empty and before the charge is 13200, and after the charge it is 0.

The total range of DOD0 is from 0 to 16384, so taking this into account:

(13200-0)/16384 = 0.80566 -> 80.566%

I believe in this case, the cell will need to be more deeply discharged during the discharge to empty phase of the cycle to increase the DOD0 measurement at this point. If you look into the log for the 7236 ID, it meets the 90% condition at this point.

Regards,

Anthony

Hi Anthony—thank you for the update; this is a great find. We will repeat the learning cycle according to your instructions and share the outcome.

Hi Jeff,

Thanks for the confirmation, please share the results when available.

Regards,

Anthony

Hi Anthony,

Thanks for your help! It still failed after retesting. The UVP voltage we set is 2.7V. When discharging to this voltage, the battery has basically lost its capacity. The fail log of the re-test is attached. Please help analyze the possible causes of this problem again. Thank you!20250610 Log.rar

Hi Rain, Jeff,

Looking into the difference between the two chemIDs, I believe I see where the issue is here. During the relax point after the initial discharge, the voltage rebounds pretty sharply back up to to ~3200mV, where in the OCV table for chemID #7236, this is still near the end of the OCV table, classifying it has a higher DOD point. However, in the chemID #7963 OCV Table, 3200mV is much earlier in the table, making the DOD much less at this point. Since this is occurring during the discharge to empty step, the discharge rate doesn't matter as much as the second discharge since resistance updates are not occurring (which requires at least a C/10 rate, preferably a C/5 rate). 

Based on this information, I would like to try reducing the discharge rate in this first step to try and combat the voltage rebound at the end of the discharge, which I believe could give us a higher DOD point when the cell is relaxed. Can you please try this test with a 1.5A or 2A discharge during this discharge to empty step? We do not have to repeat the entire test; once the discharge has completed and the cell has relaxed, we can check the DOD0_1 value to see if the rest of the test is going to be successful.

Regards,

Anthony

Hi Anthony,

Thanks for your help! This issue seems to be close to the root cause. We are retesting as per your suggestion. We will update the test results later. 

Rain

Hi Anthony,

Attached the re-test log for your review! After changed the discharger current to 1.5A, DOD0_1 value is 9560. It still seems a bit high. Can it be determined from this value whether the learning cycle can succeed? Thanks!20250617Log.rar

Hi Rain,

Looking into this log file, it seems like the DOD0 hasn't updated during the relax, which is most likely why it is still high. The [REST] bit has been set, meaning the gauge has begun taking OCV measurements for the DOD update, but the VOK bit has not been cleared, meaning it has not qualified the OCV measurement for the Qmax update. The gauge should do this automatically after 5 hours of relax, which it looks like has not passed in the log file. Can you please confirm whether the gauge completes this after 5 hours?

Regards,

Anthony

Hi Anthony,

Thanks for your update! Previously, we only focused on the [REST] bit. Once this bit was set, it would start charging. We didn't confirm the status of VOK and didn't wait for 5 hours in this time. We are already retesting and will wait for 5 hours after the discharge ends before starting to charge. The verification results are expected to be available on 6/20.

Best Regards,

Rain

Hi Anthony,

I'm apologize for the delay feedback. We verified the learning cycle again. After discharging and waiting for 5 hours, we saw that the VOK bit changed, but it still failed. The attachment is the test log. Is this problem related to the cell? Is it necessary to re-send the cells or packs to Texas for testing? This problem has affected our trial production plan. Thanks!0618.rar

Hi Rain,

Sorry for the delay, looking into the log file, it seems like VOK still doesn't become set prior to the discharge. I believe this should be set automatically when the GAUGE_EN command is sent. Can you confirm that you are seeing the same thing from your setup?

Regards,

Anthony

Hi Anthony,

Thanks for your patient explanation and high efficiency feedback. We observed carefully. When the battery was fully discharged and left to stand for 5 hours, the VOK did not change. But as soon as we inserted the adapter, it immediately became 1. Regarding this issue, is it possible that the cell capacity used for verifying the Chemistry ID is different from the cell capacity we produced, resulting in incomplete discharge and causing the test failure? Since we haven't retained the cells for authenticating the Chemistry ID, could we arrange for the cells currently in use to be retested for you?