BQ27Z746: Golden Image and very high chip temperature

Hello Ketil,

Sorry for the confusion on the question but are you saying that the current was high going thur the fets and this would cause a short circuit mode within the battery gauge thus you attached R202 to fix the issue? Regardless of using R202 or the FET your current sensing is done thru your sense resistor. Without looking at your entire schematic, data memory  parameters, or a log during a event than I would suspect the cause of the over heating being from the FETs, sense resistor themselves and going beyond the max battery capacity of 32000mAh within the system. 

Thank you,
Alan

Hi Alan,

Thank you for your response. 

Yes the HOCD (Over current in Discharge Protection) bit was set and the FET's was turned off, because of to high current in the sense resistor. The OCD voltage value was changed to -40mV. but it did not help. We needed to get the system to work and decided to short circuit the FETS with 0 ohm resistor and then everything works fine.  We then downloaded a new GoldenImage from the CPU.  To get the Remaining Capacity updated we sent the reset command, but many of the devises then lost contact with the bq27z746 after this command. It looks like the bq27z746 got in a "latch up" mode and the circuit got very hot.  The only way to fix this problem was to decharge the battery to <2V.  Then the bq27z746 got normal temperature and worked fine again.  This happens on many devises.  We know that this was triggered by the reset command.  

Do you have any comments of this problem?

We will use the FET's again for the next production batch, (made changes the motor control to limit maximum current), but we need to known what to do after downloading a new GoldenImage. 

Do we need to sendt a command after downloading a new GoldenImage?

Best regards

Ketil

Hello Ketil,

We are working on your response and should have a answer for you within the week.

Thank you,
Alan

Hi, we've found new relevant information in this case. We got several devices the last few days getting in the same "latch-state" without us running reset BMS commands. For all the devices we have no way to exit this state with code or self-recovery. It always needs to cool down for several hours without charging before getting out of this state. We also obeserve for all devices that during this cooldown period, the device is pulling a lot of power.

Our CEO is very concerned, since we have several hundred devices out there that might have to be recalled due to this error. This could render our company bankrupt :( We really need your assistance on this issue, so we kindly ask for your quick response.

Hi Alan, have you find out anything about this problem?  We got 3 failing units today. BQ27Z746 got very hot after plugging the 5V charging cable and stopped working. We use bq25071 for the charging unit. Normally we need to wait for the battery to discharge below 2V before the bq27z746 will work normally again.

Hello Ketil,

We've received this trend to an expert to better support your needs, they should reach out with a solution when possible.

Thank you,
Alan

Hello Ketil,

When the reset command is sent, is there still a short across the FETs?

Regards,

Adrian

Hello Adrian

I think the problem is related to a missing serial resistor to the BAT input pin. The BAT pin is connected direly to the battery. When the circuit get very hot the current is going into the BAT pin.  Can charging ON/OFF noise or FET ON/OFF make noise to the BAT pin, that make the bq27z746 enter a kind of "latch-up" mode?  We have not added a capacitor from the input to the output FETs. Should we also do that to prevent to get this error?