BQ78350-R1A: OCDL how does't work

some extra info,

All bit in the Enabled Protection A, B  and C are set except for the RSVD bits

Hello Nick,

I 'm confussed in the setup of my exisiting design.

I try to build a BMS with the BQ76200 and the 78350.

For LiFePO4 4S: 10A Max chartge 14.6 /10A and  max discharge 100A

Uptil now I'm not able to get the gauging working properly.

I qas startingwith a CHip version 78350-R1 and downloaded the firmware R2 in trhe device.

Then I start with a chemistry file ID 0478 and saw that all voltage levels are to hig so I addepted them to there range 3750mV

Asl changed the sense resistor settings to 0,1milli-ohm and with that I changed the IPSCL1and IPSCL0 in tje SBSconfiguration  both to 1.

can you have a look to the configuration file and help me out to get things working?

See  the attatched files below

regarding the OCDL I also not see it wqorking jet although i have read the TRM but this is mebe asl orelated due to the settings i have done uptil now.

kind regards

Rens

BUILD_1503_30-11-2020test settings.zip

Hello Nick,

I 'm confussed in the setup of my exisiting design.

I try to build a BMS with the BQ76200 and the 78350.

For LiFePO4 4S: 10A Max chartge 14.6 /10A and  max discharge 100A

Uptil now I'm not able to get the gauging working properly.

I qas startingwith a CHip version 78350-R1 and downloaded the firmware R2 in trhe device.

Then I start with a chemistry file ID 0478 and saw that all voltage levels are to hig so I addepted them to there range 3750mV

Asl changed the sense resistor settings to 0,1milli-ohm and with that I changed the IPSCL1and IPSCL0 in tje SBSconfiguration  both to 1.

can you have a look to the configuration file and help me out to get things working?

See  the attatched files below

regarding the OCDL I also not see it wqorking jet although i have read the TRM but this is mebe asl orelated due to the settings i have done uptil now.

kind regards

Rens

5277.BUILD_1503_30-11-2020test settings.zip

In the image if BQstudio I'm charging the battery with 4Amps the meter give in my opinion the correct value. the voltage level during charging is at that moment 13558mVolt  that is normal. But the reading at the registers charging voltage is 9600mV and that is aslo strange . the remainnig capacitiy is 96% and is not correct. it is  less then that. After a time I get an OC and after a period it's stops charging although It have never reached it's end point in my cass it should end around the voltage level 14,6Volt.

kind regards

Hi Rens,

Have you followed GPCCEDV ? This is required for best performance.

Best regards.

Hello Nick,

No i did not. I am  working on making log files for it  right now I have the first to files at 20 degree with to currents 3Amps and 10 Amps.

What I was thinking that when i u se one of the already known chemistry (ID-0478) file for LIFePO4 I wasd ecpecting a first starup that already give som more or less good data. but chargnig is working but dischaging totaly not,

The files at low temperture i can put it in the refrigerator at 5 degree but for the high temprature I have to see how to give around 40 degree.

kind regards
Rens

Hi Rens,

Can you upload your .gg and log files showing behavior.

Thanks,

Hello Nick,

See attached zip file

in there the settings file with chemistry ID 0478 started and two log files and the spec of the used LIFePO4 cell.

The pack is build with 4S9P.

There  log file 10Amps have the corrct temprature from the battery cells (sensor in the middel of the pack)

the 3Amp log file had the tempsenor on my desk and is not representing the cell temptrature.

Kind regards

Rens

Emergostar_TI_BMS files.zip

kind regards

Rens

Hi Rens,

I'm not sure if your OCDL question was answered yet. An OCD fault will recover if the recovery conditions are met (described in section 3.5 of the TRM). Each time the OCD protection is triggered again (OCD recovers and then new OCD is triggered), the OCDL counter increments. Once it reaches the counter limit (OCD has occurred and recovered multiple times), then the OCDL protection will trigger.

OCDL must also be enabled in the Enabled Protections registers. From your .gg file, it looks like this bit is not set.

I checked one of the log files and the discharge is getting stopped due to the CUV protection - this is set to 2400mV in your .gg file. It is not starting from fully charged - your battery should be charged to 3.65V according to its datasheet. The log files for the GPCCEDV tool should be a full discharge from fully charged.

Best regards,

Matt

Hello Matt,

You are right the question regarding the OCDL is not answered yet. I have looked again to this meganism and could not get it working as described. For this I enabeld the on the OCDL in the Enabled Protection B register. Then for testing I changed the Protections register OCD Threshold to -50mA (in my case this is 5Amps), Delay 6 sec, Recovery Threshold st to 20mA (in my case 2Amps) and recovery delay to 5sec. for the OCDL I have the Latch Limit at 5, counter Dec Delay at 10sec and Reset at 15 seconds.
when I start testing with an overcurrent of 7 Amps with a duration of 3 seconds  I see that after 5 cycles the the OCDL  is not active d DSG  FET stays on. Then I apply the same 7Amps to the system for a duration of 7 seconds and thje DSG FET is disabled (As described in chapter 3.5 during Trirp) but then The DSG FET stay disabled until I give reset or apply charge current. I hoped or expected that the system would count 5 times over current and the stay in permanent bock(DSG FET Disable).
I see in the manualat chapter 3.5  that Recovery  is only posible when Current should met the recovery Threshold. there for I changed the OCD Recovery Thresholddfrom +20mA to -20mA but then if I apply a continues load of 7 Amps the system is switch of the DSG fetand the current drops to 0 the recovery becoms active and tjhe system switch on again but that in a loop switch on/off for ever.
No I'm still not clear about OCDL.

In the gg.file  Ihave seen that My Design voltage was 3600mV and should be 3200mV, I have updated that and was expecting that it Have invluence on the Gauging,but it did not.

Then the second question regarding the gauging, It's indeed switch at CUV 2400mV.  
I have charged the battery completly stuffed to 3750Volt with a LiFePO4 charger. then It stops an I let it rest for15 minutes to some hours and the voltage drops between 3650mV to 3383mV depending of the duration of rest period.The gauge meter is showing complete full and Remainig capacity is showing 102mAh(in my case 10Amps). Learnig Full Charge Capacity is showing now fater 9 complete cycles 104mAh . When I then apply a load to the (In have set the OCD back to 1000mA, in my case is that 100A).the Ramaing Capacity still drops in letsay 30 seconds to almost 0 and after1 minute it is zero, but  the battery is still able to deliver energy of almost 10Ah.

Regarding your instruction to folow the  GPCCEDV I had not there for I started to create the data with BQstudio interfacing my circuit and measured six cycles . then I creded the file GPCPackaged.zip with zeven files in itan d uploaded it to "GPC Tool - Upload" . But I receive mutiple times the statment:

" 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"

I have checked the files but In my opinion the files are correct can you also look to that ?

In a video instruction on youtube from TI (watch?v=WAcxrGDb68g) there is talked abouut A QEN bit to enable a learning cycle i canm't find this bit in the BQ78350, is there an other way to set it in learning mode?

is my gauging not working because of current scaling i used? In the SBS configuration IPSCL0 set to 1, due to the high current I use and small sensing resitor of 0,1mohm.

Uptil now it's a hard way to get the device up and running correctly. I hope you can help me out to get it's operation correctly.
The first thing we would to see operate correctly is the gauging.
We want to finisch this first desing as soon as possible.

3835.GPCPackaged.zip

the last gg.file also included.

dump_1033_8-12-2020.gg.csv

Kind regards

Rens

Hi Rens,

I'm not sure you are interpreting OCDL correctly. It is not just a longer version of OCD - it is a way to prevent getting stuck in an endless loop where the OCD is triggering and recovering. It can be used to keep the DSG FET disabled for a longer time period. For example, if there is a weak short on the battery the OCDL protection can prevent continuously cycling into and out of OCD.

I took a quick look at the files that were submitted to the GPCCEDV tool. I think one issue is that the voltage is for one cell, but the config file specifies 4 cells. I think you either need to use the total pack voltage in the log files or change the config file to 1 cell. Also, I noticed one of the files has a bad temperature reading for one row (lowtemp_highrate line 106). I suggest quickly plotting each of the log files and remove any discontinuities that may cause issues with the tool. 

I took a quick look at the .gg.csv file. The EDV_CMP bit is set to zero which means it will use the fixed EDV values. When the cell voltage falls below Fixed EDV2, the SOC will immediately jump down to 7%. EDV1 will change the SOC to 3%. After you successfully generate the CEDV parameters, you can update the CEDV parameters in data flash and then set the EDV_CMP bit to '1'. This should help the gauging accuracy.