Subject of the issue

I'm finding that the Compass (yaw) is 90 deg off from what I'm expecting with respect to the roll and pitch directions.

I'm using SparkFun Example11_DMP_Bias_Save_Restore_ESP32 that I have modified slightly for the ESP8266. This example takes the quaternion data and calculates roll, pitch and yaw. The angles are all good, except for the zero reference for the yaw angle. That seems to be rotated 90 deg.

If I set the board flat with 0 yaw, the "nose" of the board will point north. If I lift the nose up I would expect the pitch angle to change, but the roll changes.

issue #114 @texasfunambule may have made a comment about this > The quat9 example seems to work fine (except maybe for compass alignement)

Could the quaternion components be mixed in the roll, pitch, yaw equations?? That is beyond my pay grade.