BQ35100: Unreliable SOH in EOS mode

thameera hettiwatta

Part Number: BQ35100
Other Parts Discussed in Thread: BQSTUDIO, CC1310

Hello,

I'm using BQ35100 EVM board in EOS mode with Tadiran TL-5903 (Chem_ID: 0614) and facing the issue of unreliable SOH reading which keeps on dropping on every cycle until 0%.

I have set up everything according to Technical reference manual and other documents related to BQ35100.

Calibrated Voltage, CC gain & delta using bqstudio.

Program cell chemistry.

Write design capacity(2400mAh), design voltage(3.6V), terminate voltage(900mV), cell count(1).

Set EOS mode and togged LIFETIME_EN.

When setup with my device, I use i2c on cc1310 to communicate with BQ35100 as follow;

1. GE HIGH

2. GAUGE_START

3. ENABLE LOAD (100mA for 2s)

4. GAUGE_STOP

5. WAIT FOR GAUGE CALCULATIONS

6. Read SOH, Measured Z, Scaled R, control.

7. GE LOW

The test setup has a capacitor battery(HPC1550) connected(in parallel to load) to pack- and bat+ of BQ35100 EVM board.

LiSOCl2 battery is connected to bat+ and bat- of BQ35100 EVM board.

I have found the same issue has arise to many others as well.

https://e2e.ti.com/support/power-management/f/196/t/708442?tisearch=e2e-sitesearch&keymatch=bq35100%2525252525252520eos

https://e2e.ti.com/support/power-management/f/196/t/833628?tisearch=e2e-sitesearch&keymatch=bq35100%2520eos

https://e2e.ti.com/support/power-management/f/196/p/799572/2990904?tisearch=e2e-sitesearch&keymatch=bq35100%2520eos#pi320995=1

As mentioned in these forums, i tried the modifications and still the SOH drops in Max Delta.

Following are the tests performed with the changes mentioned in the above forums.

Setting the Cell Terminate Voltage to 900mV.

Turn on the gauge after the battery get relaxed to overcome the SOH_MERIT, SOH_ERR, or EOS_BAD_OCV flags.

SOH Max Delta is set to 10.

I have set the New Batt R Scale Delay to 10, where i did 10 cycles as follow.

Then i set the SOH in [Gas Gauging][Design][State of health] to 100% without giving the NEW_BATTERY command. And performed the test again as below.

But, still the SOH keeps reducing in each and every cycle by Max Delta.

Thank you.

Thameera Lakshan.

over 6 years ago

0 Batt over 6 years ago

TI__Mastermind 42535 points

Hi Thameera,

Please follow instructions in the TRM for EOS mode. Your initial EOS learning determines the accuracy. If for whatever reason, your battery was degraded then your capacity and SOH will be affected.

0 thameera hettiwatta over 6 years ago in reply to Batt

Prodigy 100 points

Hello Batt,

The previous test was carried out using a new battery. I've refereed the TRM 5.3 for EOS mode before doing the test. Can you please provide more information about EOS mode. The details in TRM 5.3 is not sufficient to understand completely about EOS mode.

Thank you,

Thameera Lakshan

0 Batt over 5 years ago in reply to thameera hettiwatta

TI__Mastermind 42535 points

Hi Thameera,

I'm not sure what information you are looking for, however the bq35100 TRM has information about the EOS mode. You can also look up other posts in this forum on the topic for support.

0 thameera hettiwatta over 5 years ago in reply to Batt

Prodigy 100 points

Hey Batt,

I'm still having trouble in figuring out whats the issue causing the SOH_MERIT. Can you provide me a solution to stop the constant drop of SOH, if SOH_MERIT occur the SOH will remain as previous without dropping by Max Delta. As the TRM says, the EOS smoothing will try to smooth out the SOH. But the default value of eos start smooth voltage is 2.8V, which is very low and a state where LiSOCl2 become unusable. Does this cause the constant drop of SOH?

The TRM does not define about R-Short Trend Filter and R-Long Trend Filter, in 5.3.2 EOS detection. Can you please brief about it?

Thank you,

Thameera.

0 Vos over 5 years ago in reply to thameera hettiwatta

TI__Genius 10370 points

Thameera,

The short and long filters are used to identify when the impedance of the cell has reached the exponential growth part of the curve near the end. It is the main algorithm for lithium thionyl chloride cells. when the filters deviate enough, we know the rate of impedance change is growing enough we are near the End Of Service.

The SOH algorithm running in EOS mode is not a highly accurate reading. SOH is determined by comparing the ScaledR (measuredZ * scale factor learned on learning pulse) of the reading against the Ra table generated from our chemID process. The Ra table is comprised of 14 grid points through the state of health and are not evenly spaced. For ID614 the Ra table ranges from 6762 to 8873 which is a very narrow range. If your scaledR gets calculated larger than 8.8K the gauge computes that as SOH should be around 0%.

I suspect this ChemID will need to be remade or re-tested with a different load rate to help increase the resolution. However i do not think this will help you.

Going back to your data provided i see some major concerns. Your measured impedance is getting calculated as 51K. This value is way to large to make sense. This tells me you are most likely registering current but no real voltage drop. The first thing to make sure is you are getting about a 100mV drop minimum while the load is applied. You mentioned there is a cap in parallel to the cell. This cap will keep the voltage elevated and not allow a true reading of the battery voltage under load. For this you will either need to send in 10 samples of (Bat + Cap) and have a new chemID generated with this configuration, or add a FET in your system to disable the cap while doing a "gauge measurement pulse"

SOH_MERIT is most likely coming on because of the voltage readings the gauge is receiving. Please try removing the Cap and seeing if this resolves this issue.

Power management

Power management forum

BQ35100: Unreliable SOH in EOS mode