This is a threshold-based tilt detection.
The Z-axis output measured at 80° is stored as a reference, and the current Z-axis value is compared against it during operation.
When the device was operated in a high-humidity environment, a significant shift in the Z-axis output was observed.
As a result, tilt was detected at angles lower than expected(e.g., around 75°).
After returning the device to a low-humidity environment, the output shift remained.
However, after leaving the device overnight at approximately 60 °C, the output behavior recovered and tilt detection again occurred around 80°.
This behavior suggests that exposure to high humidity caused a persistent change affecting the Z-axis output, which remained even after humidity was reduced and was only resolved after thermal exposure.
If anyone has insight into physical or electrical mechanisms that could explain such humidity-related output shifts in the ADXL343—whether related to the sensor itself, packaging, or external conditions—I would appreciate your thoughts or experience.