BQ34Z100-G1: learning cycle settings: Application settings or datasheet settings?

Hello,

You should always run the learning cycle on the full battery voltage range according to the manufacturer, then after the learning cycle you can change the values to your application.

Sincerely,

Wyatt Keller  

That is how we ran the learning cycle. But: after the settings are changed, a full discharge will anly allow 12 grid points to be taken and at the end of the discharge, SOC is at 5%... 

We are on big 105Ah packs

Hi Yunfeng,

In your first post, you said settings are changed to application values after the learning cycle. You should run another learning cycle after changing settings to application values.

Did you also changed Design Capacity to actual value?

Thanks abraham!

No, the design capacity value remains the datasheet value. 

Could you please explaine why I shall run another learning cycle with the new values? Also, why shouldn't have run that first learning cycle with the application values,

Thaks angain

Hi Yunfeng,

Standard procedure is changing the values to application values first, then running a learning cycle.

But since you run first cycle already without changing the values to application values, another cycle is needed.

Don't forget to change Design Capacity to actual capacity before running the learning cycle. 

Is this standart procedure written anywhere? Because (maybe I never understood) your mates seem to say otherwise? Like above: " You should always run the learning cycle on the full battery voltage range according to the manufacturer, then after the learning cycle you can change the values to your application." 

It is an answer I have seen several times...

• Could you please explaine why? This way, I would understant.
• Also, where is the design cap taken in account in the algorithme? I haven't seen it in SLUA450...

Thank you for your precise answers

Hi,

What does "datasheet settings" mean? Gauge datasheet setting value? Battery cell datasheet value?

For learning cycle, refer to app note:www.ti.com/.../slua903.pdf

It means battery cell datasheet value. I have rade that documents quite a bit thank you. 

Could you answer to my question above?

Hi Yunfeng,

Please follow the app note slua903 as standard learning cycle procedure.

Yunfeng TANG said:

• Could you please explaine why? This way, I would understant.
• Also, where is the design cap taken in account in the algorithme? I haven't seen it in SLUA450...

The purpose of golden learning cycle is to update Qmax and Ra table, so charge voltage and term voltage should be set to the max and min values of the cell, so that when battery is charged from term voltage to charge voltage, gauge can capture Qmax.

Design Capacity is used to determine Qmax update in a learning cycle, as stated in slua903, Qmax can be updated after capacity change is higher than 90% of Design Capacity.

Best Regards,
Abraham

Thank you abraham!

So what shall I do with my original post:

Hello, 

We ran a learning cycle using datasheet settings (successfull), But our application settings are different. We wrote the application settings after doing the learning cycle.

Application: 

• Charge voltage=57.6V
• Term voltage=46.8V

Datasheet settings: 

• charge voltage=58.4V
• Term voltage=40V

What could be the consequences? 

• The grid point never exceed 12, is that the reason? 
• when doing a full discharge with High rate (using application voltage as term voltage), We have 5% SOC remaining. After a long relax period, Soc changes from 5% to 0

I Can't see any sign of good accuracy in this, could you advise me?

thanks, 

Hi Yunfeng,

Please export and upload .gg file from the gauge and learning cycle log, I need to take a look at those.

Best Regards,
Abraham

8662.gg.csv

Thank you!

Hi Yunfeng,

Please upload log file also (open bqStudio, hit Start Log, name the log and Save, after the cycle is finished, hit Stop Log to stop).

Best Regards,
Abraham

8270.discharge high and low rate.log

It is a lof of discharge with high and low rate. It seems fcc nevers updates during the discharge...

Hi Yunfeng,

I don't see any 5% SOC remaining after discharge in this log, but it seems Qmax is not updated properly. Is the battery fully charged during the learning cycle (FC = 1)? Another thing needs attention is charge and discharge current should be in C/5 ~ C/10 range when doing a leaning cycle.

Sample 10855, passes from 6% around the ned of the discharge, then the low voltage value forces SOC to 0 I guess.

During the learning cycle, the battery was fully charge FC=1 and discharge current was between C/5 and C/10. For the charge current, I am not sure. When you say qmax is probably not updated properly, Which value of Q do you speak about? qmax dod0? qmax passed Q? Qstart? Qmax passed Q seems ok looking at the referenced discharge machine...

Do you have a log of learning cycle?

DODatEOC, Qstart, DOD0 and QmaxDOD0 values are not updating correctly.

Nop I'm afraid I didn't take any. But the update status followed without any problems until 5 and I had to do another charge discharge to get update status to 6, but at the end I managed to finish the cycle. How do you see that they are not updating correctly?

Those values are all 0 in the log.

Did you run a Chem ID match using GPC tool?www.ti.com/.../slva725a.pdf

yes we did and the report showed a chem ID of 13.78% deviation. (the log sent yesterday is with this chem ID).

Meanwhile (after the log you have seen), we found out that the chem list had been updated and found the exact chem ID of our cells. 

So I am now using it, but I haven't done another learning cycle (I asked if I should but still haven't received any reply).

The log bellow shows a discharge with the exact chemical ID. It ends at 4% not 0, I don't understant why. The .gg I sent you yesterday is correct for this log.

6644.Discharge high and low rate.log

Thanks abraham,

we now have the chem ID of our cells, but how to run another learning cycle? the update status value won't go back to 4...

I asked in this post https://e2e.ti.com/support/power-management-group/power-management/f/power-management-forum/1002816/bq34z100-g1-run-a-second-learning-cycle

still hasn't got any answer

thanks!