We are working on the evaluation and testing of the ADIS16480 in our set-up and we encountered a certain behaviour that we didn't know why it happens.
The issue is in the gyro readings mainly, when the sensor is placed in a vertical way (y-axis is pointing down), the gyro readings in this direction changes very heavily (in the range of 6 deg/sec in about 30 minutes), on the other hand, other sensors perform well and their output is in the range from -0.8 deg/sec to 0.8 deg/sec. All of this while the set-up is stationary and there is not motion applied to the sensor.
When we put the sensor in a horizontal (normal) orientation, where the z-axis is facing up, we don't observe this issue. We can get values from the 3 gyros in the range -0.8 deg/sec to 0.8 deg/sec.
We are still debugging and seeing if we need to add certain configuration (like tare command, or resetting the device) in order to achieve same behaviour in all possible poses. So, we though you might help us and give us some pointers to things that we need to take care of, and the cause of this observed phenomena.
The plots attached are for a small test for a few minutes. The sampling rate is maximum sampling rate (2460 Hz), and the SPI communication frequency is 1 MHz. The EKF_CNFG register value is the default, we tried changing between the inertial frame and the body frame and observed the same behaviour.
When placing the sensor in a horizontal position (Z-Axis pointing up) the gyro output in case of stationary set-up gives the values in the range of -0.8 to 0.8 in all three axes.
Tare command was not performed in this experiment, should it affect the bias-instability of the gyros, or just affect the position and orientation of the reference frame.