bq40z60 - Chemistry ID selection

Fred,

We have an online tool to help try to find a match to your cell. We can characterize the cell, if you cannot find a match.

Tom

Thanks Tom!

I suppose I can use the bq40z60 Eval Board to acquire the required data sets with bqStudio even if the board requires a minimum of 2 cells in series. Do I see this right or will I need an Eval board that supports 1 cell to acquire the data?

Fred

Fred

There is no issue with the number of series cells. You just have to use a one parallel cell configuration.

Tom

Hi Tom,

I used the online chemistry selection tool to characterize my LiMn Batteries I am using with the bq40z60.

I got the following reply from the web site:

Chemistry ID selection tool, rev=2.21        
        
Configuration used in present fit:        
ProcessingType=2        
NumCellSeries=2        
ElapsedTimeColumn=0        
VoltageColumn=1        
CurrentColumn=2        
TemperatureColumn=3        
        
Best chemical ID : 270    Best chemical ID max. deviation, % : 3.54    
Summary of all IDs with max. DOD deviation below 3%        
        
Chem ID    max DOD error, %    Max R deviation, ratio
0    0    0
        
Warning: Deviation is above recommended level. New chem ID needs to be released for this cell. Please contact your TI representative to send the cell to Dallas for characterization   

The question is now how can I get a suitable chemistry IT for the batteries?

I attached the files which I submitted to determine the chem ID. Perhaps you can look at it and assist.
The characteristics has been recorded with bqStudio and the bq40z60Eval board following the instructions in slva725.pdf.

Thanks,

Fred

Battery_2S-report.zip

Battery_2S.zip

I worked with your data and could not improve the results. ChemID 270 is the best option, but it is just over the limit of being acceptable. You can continue with this ChemID or you can submit cells to TI for characterization. Please provide an email address, if you want to pursue the characterization route and I can send information on the submittal process.

Hi Tom,

Thanks for your reply!

I spend the last few days trying to perform a learning cycle using chemistry ID 270 but without any success although I followed the instruction in SLUA597 as closely as I possibly could. The last try just failed while I am typing this response. I am using the automated cycle using 2 relays.

I can start the cycle with a fully discharged 2S battery so the discharge cycle takes only a few seconds to reach the termination voltage. During this process, I get already a bright read error message "No acknowledgement from device" which is not true as the device is fully responding to all commands and this appears only when starting the learning cycle. With some luck, it will relax the battery now for 6 hours but in most cases, the following happens anywhere withing the 6 hours waiting time:

Mon Aug 29 11:16:11 ACST 2016: Learning cycle started...
Mon Aug 29 11:16:11 ACST 2016: Start task: Discharge ( TerminateVoltage = 6000 mV )
Mon Aug 29 11:19:15 ACST 2016: End task: Discharge completed in 00:03:04
Mon Aug 29 11:19:15 ACST 2016: Start task: Relax ( Duration = up to 360 minutes )
Mon Aug 29 12:21:25 ACST 2016: Error: [QEN: Not Set] [RUP_DIS / RDIS: Not Clear]  
Mon Aug 29 12:21:25 ACST 2016: Learning cycle cancelled
Mon Aug 29 12:21:29 ACST 2016: Error: [QEN: Not Set] [RUP_DIS / RDIS: Not Clear]  
Mon Aug 29 12:21:29 ACST 2016: Learning cycle cancelled
Mon Aug 29 12:21:34 ACST 2016: Error: [QEN: Not Set] [RUP_DIS / RDIS: Not Clear] 

If I enable logging of the registers and data flash, it is guaranteed to fail as shown above. Without logging, there is a significantly higher chance to reach the end of the 6 hours without error.

This is a screenshot of the bits after the failure (scan enabled):

If I reach the end of the of the 6 Hours cycle, if failed the first few days with "QEN" not set until I realized I have to press the "Gauge_EN" button after around 5 hours to manually enable the Gauge. I actually thought the automated cycle is doing this for me.

At some point and many, many hours of trying later, I managed once that it started with the charging cycle (after pressing Gauge_EN)! The Battery charged at C/5 until it reached the the taper current I set in the charge setup of the learning cycle. The relay turned the AC supply off instantly after the charging current dropped below the 60mA I entered. This turned out to be wrong as it takes longer for the FC flag to be set, even if the taper current in the advanced charge algorithm was set to a higher level of 70mA. I prevented another possible failure by quickly bypassing manually the relay and allowing the charger to complete charging and set the FC flag.

The automated learning cycle detected the FC Flag and advanced to the 2 hours relaxation time before discharge.

About 5 minutes before the 2 hours of relaxation period expired I was sure it was going to timeout with an error message if nothing quickly happens. I decided to select the "Registers" tab and activated "Scan". Seconds later the state of the bit refreshed and the learning cycle progressed from relaxation state to discharge state. Does the learning cycle not poll all required status bit and requires them to be updated by the Scan function in order to progress through the calibration states?

The battery was now discharged with constant current of C/10 by a electronic load. After the battery reached around 7.6V (3.8V/cell) the gauge suddenly showed 0% and the FD flag was set. I do not know why.

close to the end of the discharge cycle but with the battery voltage still at approximately 6.5V, the bqStudio lost communication with the bq40z60Eval board. the discharging thresholds was st to 6000mV, CUV = 2500mV, CUVC = 2400mV, Shutdown Voltage 2600mV, PF Shutdown Voltage 2500mV. All protection features were disabled. I cannot see a reason why the device would turn itself off before reaching the 6000mV termination voltage setting of the learning cycle. I could discharge the battery to 5900mV when I recorded the battery characteristics to determine the chemistry ID.

Here are my current register settings:

I attached the logs of a learning cycle which failed during the initial relaxation period for your reference.

learnFail.zip

This turned out to be a very time consuming and frustrating process and even if I can run a successful learning cycle and achieve an acceptable precision, I am still worried how to introduce this into serial production as I would most likely need to run this learning cycle on each production unit. Currently  I am only using the bq40z60Eval, EV2300 and bqStudio 1.3.52. Just the 2 battery cells are ours.

I hope you can help me there, Tom.

Thanks,

Fred

 

Hi Tom,

I invested another day and night to do a fully manual learning cycle without the bqStudio tool which is obviously not working.
I just optically checked the flag registers and manually connected/disconnected power supply and load.

It appears the cycle succeeded and I have a full log which I have attached for verification.

I made a copy of the Flash Firmware image and the data Flash as I suppose this is needed to replicate the settings and calibration values on other batteries. Is this correct?

From this point onward things went wrong...

I changed in Advanced Charge Algorithm -> Termination Config -> Charge Term Taper Current from 60mA to 50mA.
ACFET, CHR, DGS, PRES flags in Operation Status A turned green. There was no way to enable them anymore.
Further investigation showed that the Settings -> Configuration -> DA Configuration was cleared 0x00.
Same with Settings -> Configuration -> Temperature Enable, which was also set to 0x00.
Any attempt to write the correct values in those registers failed. After pressing Read All, they reversed to zero.

I disconnected the cells and power supply to force a reset. But no success. I could not update those 2 the registers anymore.

I decided to program the Flash firmware image, I saved before, back into the device.

Now I could manage to write the correct configuration in the DA Configuration and Temperature Enable register.
The ACFET, CHR, DSG, and PRES flags changed now to red an I could charge and discharge again.

Why did this happen???

Now I intended to verify the precision of the previously completed learning cycle but the gauge showed only 0% and this did not change at all.

How can I get this working again? I assumed when I program the previous flash Firmware and Data Memory, the calibration parameters are restored.

Also, how can I get the batteries characterized?

Thanks again.

Fred

Attachments:

Log, Data, Firmware after Learning Cycle completed:

AfterLearning.zip

Log, Data after a complete restart and firmware re-programming of the bq40z60Eval:

AfterRestore.zip

Hi Tom,

Sorry but I got another issue I cannot explain.

After the last incident I described in my previous post, I get now the FUSE flag set just before finishing charging.
Interestingly, that all worked a few days ago and never had issues with this until I got the problem in my previous post(s) which I can also not explain.

Here is the log file showing when the flag is set:

Fuse.log.zip

Here is how it looks like:

Thanks again and please have also a look at my previous posts. Perhaps you can see what is going on.

Fred

Update to my previous 2 post for clarification:

Don't use Automated learning cycle in bqStudio 1.3.52. It does not work even if it can control the 2 relays via EV2300.
The learning cycle and data logging works if it is done manually by checking the status flags and enabling the power supply or load at the correct time.

In the last post, one cell was most likely damaged as it charged over 4.3V and triggered the 2nd-level protection which set the Fuse pin high.
The same cell was also discharged below the set minimum cell voltage, even the stack voltage was within specs,  which caused the bq40z60 to turn off and communication with bqStudio got lost.

Fred