Want to report some observations after configuring several Pixhawks and Pixhawk4 mini with GPS of type M8n (Pixhawk4 mini with their standard M8n and Pixhawk 2.4.8 with Drotek M8n) and M9n (Pixhawk 2.4.8 with Drotek M9n XL). Firmware versions used being 4.1 beta 6 and 4.1 final:
In the above cases the autopilot was either placed flat (orientation 0) or on the right edge (orientation roll 90). In no tests the external compass was recognised with orientation 0, even though the compass was placed fairly flat, and pointing straight ahead. I did not make note, but one typical orientation was YAW 270. In all tests the compass calibration was executed after the initial calibration of the sensors (level, right, left and on). Test driving the rovers (boats) lead to multiple issues afterwards.
Compass calibration performed using Mission Planner.
I found I had to disable the automatic detection of the external compass calibration orientation BEFORE performing compass calibration. Issues during test driving then non-existing.
I also notice that following the regular flow of calibrating the sensors using Mission Planner was never successful when orientation was not set as a parameter value BEFORE doing the calibration if the AHRS orientation was anything else but flat (the roll 90 case above): The initial level/right/left is always successful. But the final “place vehicle level and click…” always seems to end with an error. But if I set AHRS orientation FIRST (like roll 90 in my case) as a parameter change and THEN do this calibration, this final leveling of the vehicle is always successful.
I would assume this is not as it should be. I cannot say wether this is an issue with Mission Planner or the latest Ardurover version. But I just checked with a company that stays on version 4.0 for now (limited to the compass observation), and they do not experience this. So my assumption is that the 4.1 version might have an issue.