BME680 Hardware Related Issue

[adamalfath]

I'm designing a custom board with BME680, 6 in total, unfortunately all of them are experiencing the same problem. The I2C bus voltage level is dropped to around 0.77V, consistent on all boards. When I remove the sensor, the bus voltage back to VDD, and communication with several other I2C devices on the board is working normally. I'm using 100nF decoupling on both VDD & VDDIO, and 10k pull-up on the I2C bus to 3.3V. Also I think I tried my best to maintain the recommended reflow profile and ESD handling procedure.

What could possibly be likely the main factor that makes this sensor fail? electrical? environment condition (storage/shipping)? something on the soldering process or maybe a bad batch from the distributor?

[BSTRobin]

Hi adamalfath,

Could we see your BME680 schematic?

[adamalfath]

Here's my schematic & board layout

and a couple of information that may be useful:

• All the power & ground via is connected to the respective inner plane.
• All the power rail is present as it should be.
• The MCU (I2C master) is powered with an always-on +3V3 rail while the sensor is powered with +3V3' rail (with an apostrophe) generated by the 2nd LDO which the EN pin is toggled by MCU GPIO. No transient observed during switching ON/OFF operation.
• I misplace one decoupling capacitor that should be on the VDD(IO) rail to the CSB pin, but that should not be a problem.
• There's 10k pull-up to +3V3' on SDA and SCL near the MCU.
• All the I2C communication is working normally when the BME680 is removed from the board.

[BSTRobin]

Hi adamalfath,

If the actual I2C external pull-up resistor has been connected, there is no problem from the schematic diagram.

[adamalfath]

Yes, I'm too sure this problem is not caused by the wrong design/schematic. It must be something to do with the part itself, whether it defected from the beginning or broken in handling/soldering.

I'm also interested in this internal I/O diagram:

Can it be caused by this busted ESD diode? Because that 0.7XX Volt on I2C bus is like the line is shorted and the value close to the diode forward voltage. Maybe Bosch technical team ever find a problem with similar behavior like this?

My goal here is to understand how this sensor failed and learn to prevent it from happening again.