We are working on a Battery Management system which its battery pack is installed with BQ20Z75BTR-V160 gas gauge. The communication lines (i.e. SMBD and SMBC bus) circuitry is as per page 15 application schematics of bq20Z75 datasheet.
On the other hand we have a host controller (microcontroller or uC) that locating on a separate board. The host is powered by a 3.3V source; uC communicates with the above battery system via I2C interface with pull up resistor of 1.5kΩ. There is a TVS sort of device (B72500D50A60) protecting uC’s Vcc line.
Our problem is our uC is sometime gets killed while the battery is hot plugged into host’s controller board.
Based on our oscilloscope waveform we tends to believe that the killing source is originated from battery pack’s (hence BQ20Z75’s) SMBC line. This is said so because of the following evidences:
- There is an overvoltage event on 3.3V line (Fig. 1(a) Ch3 (pink curve)) which was push up to 4.8V peak. After peaking 3.3V stay low (<3V) which indicates that the uC was killed.
- In Fig.1 (a) during the overvoltage event SMBC, SMBD, and 3.3V lines are all push far above their nominal (3.3V).
- SMBC is blame to be source of the problem because:
- The first overvoltage spike take place at SMBC.
- The overvoltage amplitude is highest (as high as the battery voltage VBAT=15V) at SMBC as compared with others.
- From Fig. 2(a) the overvoltage event of SMBC is not immediately take place but about 20ms after the battery (VBAT) was plugged in.
Meantime we kind of believe that BQ20Z75’s SMBC line is sort of latchup during the hot-plug action which might create an unnecessary path that channeling between VBAT and SMBC; such excessive voltage overstressing and killing our uC.