BQ34Z100-G1: SOC drops to zero

Hello Wyatt,

I'm Erica and I'm working with Matteo on the same project.

We are running learning cycle following your suggestion but we cannot complete it because it never update learning status to 5. We tryed more than 3 complete cycles with the same result, waiting for relax for more than 2 hours after FC detected and more than 5 hours after complete discharge.

I add configuration file and logging here.

210104 configurazione e log.zip

Do you have some advice to us please?

Thanks,

Erica

Hello Wyatt,

thanks for your reply.

The design capacity programmed before learning cycle is 22000mAh (line 49 in the file start210104_autoapprendimento.g in my previous e-mail).

We've done learning cycle exacly as described in TRM Appendix D and we tryed more than one time: we cannot reach Update Staus 05. The gauge detect FC (you can see in the logging file) and we let it relax for more than 2 hours with no load.

Thanks,

Erica

Hi Erica,

This is a common issue with many related threads on this forum.

Try searching the forum for bq34z100-g1 update status = 5

Best regards,

Hi Nick,

thanks for your reply.

I'm reading all the post I can find about this issue and ckecking every parameters but I cannot find something wrong in the configuration:

design capacity is correct, the calibration was done, FC is -1 (and gauge always detect FC), Max Error 100%, taper currents is larger than the Chg/Dsg Current Threshold, quit current have a low value (40mA), relax time is longer than 5 hours for both charge and discharge (with no load connected - measured current on gauge is 0), current is C/5 for both charge and discharge, temperature are within 10/40°C range, RSOC_HOLD is disabled, the cell is properly selected because we find the exact model number in your library..

Our battery pack is NIMH, 24V 22Ah.

Moreovers we have already done the learnign cycle on other battery packs previously, with the same component Bq34z100-G1 succesfully.

Do you have any other idea on parameters to be checked?

Could you have a look to the files we shared?

Thanks,

erica

Hello Wyatt,

as wrote yesyerday, I've checked all the common issue in the TRM but cannot find any wrong configuration in our set data.

Design capacity is correct, the calibration was done, FC is -1 (and gauge always detect FC), Max Error 100%, taper currents is larger than the Chg/Dsg Current Threshold, quit current have a low value (40mA), relax time is longer than 5 hours for both charge and discharge (with no load connected - measured current on gauge is 0), current is C/5 for both charge and discharge, temperature are within 10/40°C range, RSOC_HOLD is disabled, the cell is properly selected because we find the exact model number in your library.. and so on..

The FC is NOT always set; when we do charging phase, the FC turn red in the right moment (at the end of charge).

So there are no wrong configuration form your opinion in our file?

Thanks,

erica

Hello Wyatt,

if the Design Capacity is 22000mAh, the configuration of QMax 0 at the beginning of learning cycle must be the same? Is it right or not?

Thanks,

Erica

Hello Wyatt,

yes we've done exacly like SLUA925. 

The VOK and FIRSTDOD remains high and it cannot update learning stus up to 05: the Quit Current is 40mA, we relaxed pack for more than 5 hours for both charge and discharge, the Terminate Voltage is 1000mV/cell and FC is correctly detected (with deltaT/delta t).

The Passed Q is always more than 100% of Design Capacity: maybe we have to settle Design Capacity higher? The Q MAX CELL 0 value we settled is 22000mAh exactly like Design Capacity, Passed Q afetr charge is more than 23400mAh. Shoul we change Design Capacity to 24000mAh?

Design Energy parameter is used or not for NIMH chemistry?

Also one note (I don't know if useful): during relaxing the voltage of full charge drops from nearly 30V during charge to nearly 28V when relaxed and from 20V cut-off to nearly 24V when relaxed. Should we change something in your opinion?

Battery pack is 24V 22Ah (minimum) and 24Ah (typical), 20s2p 

Thanks,

Erica

Hello Wyatt,

today I try again changing Design Capacity and QMax0 to 23500mAh even, if we consider the real capacity of the battery pack, it is about 22000mAh (discharge cycle from full charge state measures nearly 22000mAh in our testing machines). Of course Passed Charge to achieve a full charge is more than 22Ah, nearly 23.5Ah.

From manufacturer data sheet, typical capacity of each cell is 12Ah and minimum 11Ah, so (for 2p battery pack) a variation form 22Ah to 23.5Ah is reasonable. Does it create a problem on Bq learning cycle?

Can you confirm that QMax 0 must be settle at the same value as Design capacity before starting learning cycle? 

Moreover, which are the OCV range expected during relax for NIMH?

My Update status this morning is still 04....

Thanks,

Erica

Hello Erica,

A small variation shouldn't be an issue, as long as you get 90% change and the passed charge isn't greater than the design capacity.

The OCV depends on the cells you're using.

Can you share the .gg file you have now after the edits made?

Sincerely,

Wyatt Keller

Hello Wyatt,

the Passed Charge that Bq measures is greater than Design Capacity.

Is it mandatory that Passed Charge value result lower than Design Capacity? If you confirm this, I must change again the Design Capacity value (during today cycle the Passed Charge is still greater than Design Capacity).

Thanks,

Erica

Hello Wyatt,

here also the last .gg file from auto export.

69start210112_autoapprendimento.gg.csv

Thanks,

Erica

Hello Erica,

Yes I will have to confirm if it is tomorrow, but I believe the requirement is it must be less than design capacity or within a certain percentage of it.

Sincerely,

Wyatt Keller

Hello Wyatt, 

the Qmax cell 0 value is the same as Design Capacity, I confirm. We settle the maximum value rated for the cells.

We've done the cycle again (it's nearly the 20th complte cycle with charge and discharge) but nothing changes, the Update status is still 04 and don't change to 05.

Did you find something wrong in our .gg file?

Can you kindly check the OCV voltage for the cell code 6118? Seems that Bq cannot detect relax mode and don't take any measurement.

Thanks,

erica

Hello Erica,

Let me get someone else on the team to take a look at this, sometimes these chemistries are difficult to configure correctly.

Sincerely,

Wyatt Keller

Design capacity setting looks ok as your discharge is more then 90% of discharge capacity. But gauge also checks

if you have more then 90% DOD delta between first and final DOD points. It is possible that it is not satisfied if battery is either not

fully charged or fully discharged compared to OCV table. If this is the case, you can still create golden image by submitting the log file and GG file
to gauge parameter calculator golden image tool. This can be accesses here:
Please follow the instruction manual on preparing the submited files. It is importaint that GG file is exported right after programing chem ID, so Ra values are fresh unchanged chem ID values.

Hello Yevgen,

I've some question about "Simple Guide to GPC Golden GG Maker Tool" before to start a new cycle:

- can I use this tool also for 24V battery pack? (I've seen the standard config. is only for 1 cell)

- have I to do IT_ENABLE before start charging?

- have I to wait till FC detected as in the learning cycle? We have CC charging method and we have to stop manually the charge phase when FC is detcted.

For my knowledge:

- why I cannot reach more than 90% DOD even if battery pack is rightly fully charged and fully discharged?

- which are the OCV value you consider for our cell 6118 model?

Thanks,

Erica

Hello Yevgen,

thanks for your feedback, I'm doing today an other cycle (charge-relax-discharge-relax) as required by GPC Golden GG Maker Tool and I'll update result tomorrow morning.

With reference to your last 2 answers, I'm sure that battery pack is full charged: we are testing it with our machines and we can collect data of charged capacity, dV/dt end of charge and also dT/dt, everything is in line with manufacturer data and FL is also detected by Bq in the right moment. This is why I'm 100% sure that battery pack cannot be >10% undercharged.

It is not a problem related to full charge but seems that BQ cannot detect relax mode: can you inform about OCV expected for NIMH 6118 cell model?

Thanks,

Erica

Hello Erica,

It looks like the log provided goes from around 1450mv to 900mv, but the chem ID goes from 1425mV to 1100mV, can you share the GPCCHEM report and submitted data? This may cause an error during the learning cycle because the gauge has no reference for voltages lower than 1100mV.

Sincerely,

Wyatt Keller

Hello Wyatt,

why the chem ID goes only to 1100mV? All manufacturers indicate NIMH cut-off voltage about 1000mV to 900mV/cell if you want to discharge to empty.

We've done several times the upload of log, config and gg file to GPC tool (as required from Yevgen); we received automatic reply with unknown error.

Here is the zip file with requested files inside, can you kindly check please?

8053.gpc.zip

Thanks,

Erica

Hello Erica,

What was the original reason to use chemID 6118? Its always best to run the GPCCHEM tool to select an ID.

You don't need to include the chemID in the config file. The only thing I can see that looks incorrect is the current pulse before the discharge. What is the error message you receive? 

Sincerely,

Wyatt Keller

Hello Wyatt,

we use chemID 6118 because it is exactly the cell model we are using for our battery pack!! I know from the manufacturer that this pretty new cell was included in the Bq library thanks to a specific test made in cooperation between Ti and manufacturer.

Why do you think isn't it correct? Is it there any problem with chemID 6118?

We have used a specific chemID (without using GPCCHEM tool) to run other learning cycle for other battery packs without any problem, of course only if the exact model of cell used for the battery pack was included in your library.

I've included the chemID in the config file because it is clearly requested in the "Simple Guide to GPC Golden GG Maker Tool" at page 2 (SLUUBC9).

The current pulse before discharging was only an operator mistake, we took off the pulse from file and upload again the .zip file. I've received again this message error:

Error: Computation was not successful, but the specific reason was not identified.
       Please check that all files are prepared according to the documentation.
       The support personnel have been informed

Can you check please? I need urgently to solve the problem, if there is a problem with selected chemID, please kindly let me know.

7144.test.zip

Thanks,

Erica

Hello Erica,

Just wanted to confirm that's how you selected it.

It does look like there may have been an issue with characterizing this cell, because the range does not match the cell manufacturers spec.

I'll contact our chemID team about re-characterizing if what I see is correct.

Sincerely,

Wyatt Keller

Hello Wyatt,

do you have any news for us?

Thanks,

erica

Hello Erica,

This cell will need to be run again. It was last run in October 2017 and needs to be updated.

I sent you a request on the E2E forum to take it offline.

Sincerely,

Wyatt Keller

Hi Wyatt,

This is Paul from Eberspaecher Vecture inc. in Canada. We have the related trouble with BQ34z100-R1 with the SOC drop (during discharging) much faster than normal.

We designed a 8S1P BMS with LFP battery (3750mAh).. When we charge it with 7.5A (2C), the SOC and RM look good. then we discharge it with 1.9A, the SOC and RM drop much faster than normal. I have tried to change many related settings in configuration, but we still got similar result. I attached the log file.Golden EN IT 17FEB21-2.log

Hi Wyatt, I lost the connection at previous email.

Thanks,

Paul

Hello Paul,

Could you make a new post with your questions? This post as been taken offline and we resolving the issue there.

Sincerely,

Wyatt Keller