I have had four ADIS16480 modules in ice-deployed buoys in the central Arctic measuring wave-induced motion. I found that I was getting a lot of magnetic interference from a pulse-width modulated solar panel / battery charge controller whos wires have to pass close to the IMU, so I switched off the magnetometers using EKF_CNFG = 1 (giving up heading as I had another source of heading data), and the reported Euler angles improved greatly, since the low frequency magnetometer data that had this interference was no longer entering the Euler angle solution through the Kalman filter. Yaw just wanders around, which is OK since I am primarily interested in tilts. However, during yaw transitions between -180 to 180 degrees, sometimes, but not always, the the pitch and roll Euler angles occasionally will wander through +/- 90 degrees. Is there a better way to fully disable the magnetometer data from entering the Euler angle calculation?