BQ35100: SOH in EOS-Mode drops after 20days from 100 to 0 in 5hours

Hi,

Our device expert would look into this and get back to you shortly within a day or two.

thanks

Onyx

Teufel,

Are the plots you are showing your data from your run? Are you able to send the entire log file? 

You are getting good measuredZ and ScaledR results. Please set termination voltage to 900mV. There is an issue with some old ChemID's where this was a needed change. I thought they were all fixed, but maybe one was missed. Please send the SREC file exported from your device so i can confirm all settings. 

Thanks,

Eric Vos

Hey Eric,

yes the plot is showing my collected data.

Attached you can find the .srec-file and the logfile. Please note that .srec-files are not allowed to upload, so i added an .123.

After the SOH dropped zero, I let the test run continue.

The logfile (bat.log) is a. csv-file with a tabulator as seperater. If you're checking the file you might wonder about the three drops in scaled r and messuared z. At this three times I switched the bq35100EVM-795 from my host mcu to bqStudio and backwards. But I didn't change anything in the configuration of the BQ35100.

Thanks in advance for your help!

191125.srec.123

6646.bat.log

Our experts will help on this next week. Closing this out for now.

t_teufel,

I am looking into this, but i am needing help internally because i can not explain the sudden drop as it is. Everything from my side looks ok. Did you do as i recommended and change the "cell termination voltage" to 900mV? 

Also did you swap to a fresh battery 1/2 through the data log? 

Thanks,

Eric Vos

I changed the cell termination voltage to 900mV, but I didn't start a new test run. I want to discharge the battery completly before starting a new run.

To the second question: No, I didn't swap to a fresh battery. As I mentioned before I switched at this time the bq35100EVM-795 from my host mcu to bqStudio and backwards. I switched to the bqStudio to read out the configuration and LTFlash.

Hello Teufel,

Please let us know what the new test results are by switching termination voltage to 900 mV.

Thanks!

We bought a second bq35100EVM-795 and I build up a second testrun with 900mV termination voltage.

But if the SOH drops to zero during one day, like in the first run, we will see after 20 days. Around the 20th of December.

Till now it all seems the same as at the first run. The SOH stays stable at 100%. Theoretically the SOH should drop by 2% a day.

Attached you can find the logfile of the second run.

2625.bat.log

This is the graph of my second test run. SOH stays at 100% instead of dropping by 2% a day (till now around 14%).

t_teufel,

I am getting a chance to look into this again and wanted to clairify a few things. 

1) in the first run SOH was decreasing straight to 0% 

2) in the 2nd run you modified TermV to 900mV now SOH is staying at 100%? 

For case #2 this looks to be the correct behavior. Your scaledR is coming out to the <2500, Looking at the impedance table for the ChemID you are using this still corresponds to fully charged. 

For this gauge the main purpose is to get an end of life warning. SOH is an add on, but it isn't as accurate as we would like. The more and more empty the cell gets the better the accuracy should get. Since these cells voltage and impedance is very flat at the beginning of like i don't expect we get decent accuracy until after 50% life is expelled. 

Thanks,

Eric Vos

Right, at the second run I changed the TermV to 900mV. But this affected the result not really.

The SOH of the 2nd run was at 100% for the first 12 days and drop to 0% during 4 days. And please remember I'am discharging the battery by 2% a day.

When the SOH droped to 0% the measured Z is around 8,5Ohm and the scaled R is around 5,5Ohm. Compared to the Ra Table is a discharged SAFT LS14500 at around 28Ohm. How does this fit together?

Back to my 1st run which is still running:

Measured Z (32,8Ohm) and the scaled R (16,4Ohm) are stable since a week. The voltage during a discharge puls drops to 2,7V but the EOS Bit is not set.

Here I can't understand why the Measured Z is stable at 32,8Ohm. When I'm calculating internal resistor by hand (Difference between Voltage after the puls and Voltage during the Puls multiplied with the puls current) the result is around 57,2Ohm.

SecondRun.log

FirstRun.log

t_teufel,

Starting with the first log. SR stops being updated. The only reason for this would be the EOS has been detected and set. So EOS was reached. I do not see it in the log to confirm.  

For the 2nd run. Your EOS smoothing engine also running. Please lower your smoothing start threshold to 900mV as well. You do not need to restart the test, just manually update [State of Health] under Gas gauging tab back to 100% and continue running the test. 

Eric, thanks for your help! Normally I'am checking the EOS-Bit and if it's set it would be logged.

I stopped both running tests and started new ones. In both configurations I changed "Cell Terminate Voltage" and "EOS SOH smooth Start Voltage" to 900mV.

Afterwards I pluged in a new Batterie and changed to a customized MCU to discharge the batteries and doing the EOS measurements. During the three learning pulses the SOH drops 2% each learning pulse. After this three pulses the SOH stabilized and I changed it by hand to 100%.

Now the batteries are discharged by 2% per day. Let's see if it works well now.