BQ35100: Problems reading SOH in EOS mode

Thank you for your answer, it seems that the problem is solved. Now I'm doing a long term test with a fresh battery in order to find out if the SOH will be inline with the obtained results.

Sometimes I also have another problem while reading the Measured Z and Scaled R using my MCU. Sometimes  Measured Z and Scaled R are equal to 255 and I've to perform multiple readings (two or three) in order to obtain acceptable values. Note that when even Z and R are equal to 255, I’m able to read correctly the value of the SOH.

Thank you!

Hello Eric,

Yes, I'm reading the correct length and the output values seems now correct. It seems that problem was related to the rest time.

Thank you Eric.

Hello Onyx,

In my referred “long term test”, It is suppose that for every discharge the SOH theoretically drops about 0,15%.

As you suggested I’ve programmed the cell terminate voltage to 900mV but SOH percentage dropped 5% in each reading, stabilizing in 85%.

At this point of the experiment, the cell I’m testing should be around 57%.

Attached to this message there is plot showing the Measured Z, Scaled R and SOH corresponding to 278 discharge cycles. I hope that this might help to figure out the problem.

Thank you.

Eric,

I'm using ChemID = 0x619 (EVE ER14505-SPC1520).

Thank you,
Sam

Sam,

Every time you issue the "New Battery" command it will reset the pulse counter back to 0 meaning scaledR will not retain the learned scale value and you will revert to using MeasuredZ again. 

I suggest for your test after you have gone through your 25 pulses write the Data Memory parameter [Gas Gauging][Design][State of health] back to 100 without issuing the new batt command. 

Thanks,

Eric Vos

Hello Eric,

I am not issuing a New Battery command after the learning pulses. I've only done that when I installed a new battery.

Nevertheless, I'm running a new test where I put the SOH value back to 100% right after my first 25 pulses (note that the New Batt R Scale Delay = 25).

But one thing I still not understanding is the behavior of the Scaled R. On the first 25 readings Scaled R follows Measured Z, after that, Scaled R dropped to 511, this happened consistently on three tests. Since that in the previous post you said that Scaled R should roughly follow Measured Z what could explain the sudden drop Scaled R? (and stays that way forever).

Am I configuring New Batt R Scale Delay correctly? I’ve chosen 25 because it is after 25 discharges that Measured Z stabilizes.

Thank you.

Regards,

Sam

Eric,

I'm starting to doubt if my selected chem ID is right.

I am using one cell (EVE ER14505):
www.tme.eu/.../EVE-ER14505_S.pdf

I've selected chem ID 0x619 where the corresponding bqstudio description is : ER14505-SPC1520 (5200mAh).

I assumed that "ER14505-SPC1520 (5200mAh)" is similar to my battery. I've just adjusted the "Gas Gauging -> Cell Design Capacity mAh" value to 2000mAh.  And after some research I've now realized that that chem ID is not the correct one. The most similar chem ID that I've found was the 0x614 from Tadiran.

Sorry for my mistake, but now, what is the most suitable chemID for my cell ? 


Regards,
Sam

Good morning Eric,

This weekend I’ve made another set of discharges that are equivalent to consumption of 19% of the battery. This time I’ve used the chem ID 0x614 which I think to be most similar to the cell I’m using (EVE 14550 datasheet).

As recommended, I’ve installed the new cell, sent the “NEW_BATTERY” command and performed 25 discharge cycles (New Batt R Scale Delay = 25). Right after that, I’ve forced the SOH to 100% without issuing “NEW_BATTERY”. As the plot bellow shows, the Measured Z and Scaled R seems now to have consistent values but the SOH is dropping 1% after each reading (SOH max delta = 1%). I still have the “Cell terminate voltage”= 900mV (the problem is not caused by this).

The fact that I do not have a usable SOH can be related to chemID that might not me suitable for my battery?

Regards,
Sam

Sam,

From your images attached it appears scaledR and measuredZ are always the same value. This doesnt make sense to me since on the learning pulse the scale factor should adjust ScaledR away from measuredZ to better correlate the cell to the ChemID.  When you forced SOH to 100% the only thing you should have changed is [State of Health] at address 0x4209 back to 100%. 

I also do not think ID 619 is a good match. Your cell calculated an impedance around 6.6K while ID619 is sub-1K. 

In addition, we will be unable to match the "expected" line you have shown. The only way we can estimate SOH is based on the Impedance table. Based on the data you have shown the impedance is very flat, giving the gauge very little to differentiate between. Please continue the test to empty logging the measuredZ so that we can see they type of slope you encounter with your cell and setup

Thanks,

Eric Vos 

Sam,

Your impedance is registering as 6000, while our impedance table is sub 1K. Since your scaleR is not being brought down to within the table you are effectively registering a "Dead" cell. If you were to change [SOH max Delta] = 100 then SOH would drop to 0% on the first reading. 

Please issue the new battery command again and run for the 25 pulses. we should see ScaledR adjust way downward. After you see this happen then revert SOH back to 100% (not using command). ScaledR should still be low. Then continue the test

Let me know any questions

Thanks,

Eric Vos

Massimiliano,

Something that is not concidered is the temperature aspect. The TI gauge will take the measuredZ and compare it to the Temperature compensated Ra Table. What is your temperature reading? 

Thanks,

Eric Vos

Hello,

The temperature I'm reading is around 26 degrees +/- 2.

I'm using the internal temperature of the bq35100.

Thank you in advance,

Sam