I'm trying to troubleshoot an issue with the BMM150 on a recent build of 16 PCBS. All 16 PCBs have the same problem, the X axis is overflowing while the Y axis, Z axis, and temperature read correctly. Specifically register 0x42 reads 0x01 and register 0x43 reads 0x80 at all times, regardless of sensor orientation. These boards are the second revision of a PCB for which software was developed and remains unchanged with resepct to the sensor. On the down rev boards all three axes read correctly.
The footprint geometry and pin assignments are identical between the revisions. Because we can communicate with the chip and read valid data out for all but 1 axis I don't suspect a hardware design problem. For this design the DRDY and INT pins are disconnected.
The parts assembed onto these PCBs were manufactured in 2019 and have the code:
What is the self-test result? Can you inspect the part under a microscope and confirm that the silicon is not chipped ?
Since BMM150 is in a WLCSP, the most common failure mode of BMM150 is mishandling that can chip the silicon and damage the sensor.
As you can see from the bottom-side picture below, the X and Y axis are very close to the edge of the part (this is why the balls are not centered, but offset)
The self test bits report 0x01 for all three axes suggesting there is nothing wrong internally.
I did inspect all of the boards under a microscope for any chips or damage to the package. Yes, I understand that CSP parts are delicate. Per the datasheet this part was not undefilled. It just went through reflow and underwent a class 3 inspection. The parts were also x-rayed by the assembly house and no shorts or assembly defects were seen.
I never suspected chip damage since all 16 boards exhibit the same behavior - overflow on the X-axis.
I started suspecting over-temp during reflow but the assembly house provided the temperature profile (attached). From the temp curve it looks like the over didn't reach the maximum reflow temp suggested by the datasheet.
I also double checked the panel and footprint for issues. I specifically checked the panel for shorts and opens and there were none. The two left-inside pads were also not connected as they should be (DRY and INT are floating).
I'd like to remove some of these parts and send them in for inspection if that is possible.
I'm located in North America (WA state). I'll just use the contact form, thank you for posting the link.
This is not a new design but rather a revision of an existing design. The revision was made to fix a single footprint error and all of the changes were minor. The loctation of all the chips is the same as the down rev. The footprint geometry for the BMM150 is the same.
There are no high current traces or magnets anywhere on the board on in the environment in which it is being evaluated. The x-axis overflow is static regardless of the orientation of the board. The y and z axes respond normally and respond to a magnet brought into the vicinity of the board as the board is rotated.