ADXL355 adxl355_get_raw_fifo_data() perfectly zero acquisition

Hi, 

I'm experiencing perfectly zero acquisitions for raw_x, raw_y or raw_z using adxl355_get_fifo_data() with an ADXL355 accelerometer: I guess it is possible to acquire perfectly zero, but it seems strange for me. 
It happens randomly, but, trying to measure it, I can say around 30/40 samples in an acquisition window of 30 seconds at 125Hz of sample frequency.

Is that possible?

Thanks,
Fabio