Accelerometer zero-g level change over time

[PierreGremaud]

Hi,

I am using IIS2ICLX MEMS accelerometer in an IOT device.

The IIS2ICLX is soldered on a PCB that sends data to internet every 10min.

To measure the zero-g level of both x and y axis, I suspend the device from the ceiling so that the PCB (and the accelerometer) is perfectly horizontal and and vibration-free.

I noticed that the zero-g level changes if I briefly touch the device, and wait for the device to return to its horizontal position (the temperature of the device does not change).

Does anyone has an explanation ? 

I there a way to minimize this effect by using two redondant accelerometers, with a 90 degre rotation between them ?

Thanks for your help.

[Federica Bossi]

Hi @PierreGremaud ,

The most likely explanation is mechanical stress or slight PCB deformation caused by touching the device. Even if temperature stays constant, a small change in board strain or residual tilt can shift the accelerometer zero-g level.

To minimize this effect, make sure the PCB is mounted as rigidly as possible, avoid stressing the board near the sensor, and wait for the device to fully settle before taking measurements. Using two accelerometers at 90° will not really remove this effect, since both sensors would still be affected by the same mechanical stress.