We have designed the LIS3DH onto a custom PCBA, and we are seeing some oddities with the sensor readings post manufacturing. The PCBA goes into a two part injection molded enclosure made of ABS, and the two halves are ultrasonically welded together. After this process, we are seeing what appears to be an offset zero issue along two of the axes. The entire range of gravity vectors is as expected at ~18.9m/s^2. However, positive Z facing toward the earth yields a gravity vector magnitude of ~13.8m/s^2 and negative Z down reads ~5.1m/s^2.
Has anyone else seen something like this or have ideas for how to fix this?
