Recently we are doing a project that detects the motion of an object that is being dropped on the ground or table. However, we found that very often ADXL362 cannot detect the high acceleration value when the impact occurs. For example, if we drop a gadget with ADXL362 built inside from 30 cm above the table, the acceleration value should be about 8g or even higher at the moment when the gadget touches the table surface. However, our experiment results show that 8 out of 10, the acceleration values detected by ADXL362 are below 5g. The acceleration values are shown in the figure below, where black line is the root sum squared of the three axes values, and R,G,B lines are for x, y, z axis respectively. Sampling rate is set to be 200Hz. Maximum acceleration value is set to be 8g. You can see that the acceleration value is near zero during the free fall, and then the impact occurs. An acceleration value of 3g is detected. About 8 out of 10, the detected values are low.
If we drop the gadget on a soft surface, such as a stack of tissue papers, the acceleration value is about 7g. The data is shown below.
I wonder if this low acceleration values are caused by low sampling rate or other reasons. We tried other company's accelerometer also. This phenomenon also occurs but not so often. For example, Botch AN1057 cannot detect correct impact acceleration value 3 times out of 10, using the sampling rate of 125Hz. But ADXL cannot detect correct impact acceleration value 8 times out of 10, using the sampling rate of 200Hz. So the sampling rate may not be the root cause. So why is it not working? Do we need to change some settings?