BQ20Z95: IIC BUS data issue

Hello,

We use BQ20Z95 as battery gas gauge IC. IIC BUS as data interface.

Between MCU and battery, we use TAC9617A as IIC repeater.

We monitor the temperature data of battery by MCU with IIC bus, we find there are some extremely low(-273.05, -272.55) and high(317, 5380.85, 5329.45) temperature data during monitor process, although the rate  less one in a thousand.

Whatever we change pullup resistor value, the extremely temperature always exist.

Could you help analyze why this issue happen and give a hardware solution?

Thanks.

Cherry Zhou：

Hello,

We've got the issue and escalated to E2E for more help as the link below, you could take a look at it:

e2e.ti.com/…/bq20z95-iic-bus-data-issue

,

Cherry Zhou：

Hi,

We would recommend that you check communication by removing the TCA9617A. This will allow you to verify gauge operation. If you still see intermittent high/low temperature, then we would check the solder quality of the thermistors.

We strongly recommend that you refer to TCA9617A datasheet and follow all the design guidelines in it. Also make sure that the MCU communication code handles clock stretching and handles errors gracefully.

,

Chambin Zhang：

Hi,

I jumped TCA9617 to connect MCU and battery, the issue still happened. so TCA9617 itself didn't lead to the issue.

I changed the two other batteries, the issue still happened again. that mean battery solder may not the issue reason.

I found the issue temperature keep are -273.05 and -272.55, no other issue temperature after I reduced the pullup resistor value.

I guess issue may cause by battery itself or MCU IIC PEC (packet error checking) function.

How to verify these two possibility?

Thanks.

,

Cherry Zhou：

Hi,

We've gave some updates to the expert and please see the link on E2E below in case needed:

e2e.ti.com/…/bq20z95-iic-bus-data-issue

,

Cherry Zhou：

Did you check solder on thermistor leads?

As you mentioned, it is most likely MCU code issue. The easy way to verify is to use bqStudio to log the data and see if the error can be reproduced. If it cannot be reproduced, then the MCU code will need to be checked.

,

Chambin Zhang：

Hi Cherry,

How to use bqStudio to log the data? we have EV2300 monitor the battery.

But we only found normal temperature as follow picture, no issue temperature record. is there others way to log temperature?

Thanks.

,

Cherry Zhou：

Hi,

On the Registers tab, there is a "Start logging" button. Choose a file and let it run for a few hours. Then you can check the log file temperature column data.

,

Chambin Zhang：

Hi Cheery,

I used EV2300 monitor the battery discharging and record the log file. Here I found follow issue in the log file.

1. 

about he "Possible Error-See error log" information. where can I see the detailed error information or error reason?

2.  you can see a time interval in the below picture, I found several place data discontinuity in the log file. what's happen? is it a issue or normal situation?

Thanks.

,

Cherry Zhou：

Hello,

We will consult with our expert and get back to you early next week.

Thanks.

,

Cherry Zhou：

Hi,

bqStudio has a scan interval of 4 seconds for registers by default. Looks like this was changed.

Hammering the gauge with communication can leave the gauge with not having enough time to perform simulations. We will need more data to know if the gauge entered an unresponsive state where there is discontinuity in the log. Can you provide more details on what bqStudio settings/selections were changed while taking the log?

Infrequent communication errors are normal. These must be retried in the real system. You can get details of communication in another file that is in the same location as the log file. It should have "err" in the file name.

,

Chambin Zhang：

Hi Cherry,

Please see bqStudio setting information, we only monitor real-time temperature.

Set scan cycle is 1 second, log cycle is 1 second.

Please see our log file there is a error record

My question is what's this message meaning? is there no data be measured？ what information be sent to out MCU at this situation?

I saw in the log file, there are "Sample" number and "Elapsed(s)" number, at the beginning, the Elapsed(s) number = Sample number, but as the time fly, the Elapsed number become much bigger than Sample number, could you explain what's reason and what happened?

Thanks.

,

Cherry Zhou：

Hi,

We've created a new thread regarding your follow-ups since the original post had been long time ago, you could also take a look at the link below:

https://e2e.ti.com/support/power-management-group/power-management/f/power-management-forum/1134321/bq20z95-iic-bus-data-issue

,

Cherry Zhou：

Hi,

Could you send us the entire log file? We also are curious why samples are taking 1s in the beginning and what appears to be around 1.4s as the file goes on.

,

Chambin Zhang：

Hi Cherry,

FYI.

Please note: I changed the time interval from 3 second cycle to 1 second cycle at first file at the monitor beginning stage.

907 22 03.log9141102.log

Thanks.

,

Cherry Zhou：

Hi,

Thanks for your info and please expect the response.