I just updated my companion software to version 0.0.30 and my Ardusub to version 4.0.3. I am attempting to calibrate my accelerometer by following the steps in the sensor calibration on the bluerobotics website and am getting a failed calibration each time. I am running version 4.0 of QGroundcontrol. Can you help me out so I can get me sub back in operation?
1 Like
Hi @AMIEngineers,
Assuming you’re using our standard electronics enclosure, are you calibrating with the Roll90, or are you specifying some other flight controller (Pixhawk) orientation? Using QGC 4.0.5, an accelerometer calibration for Roll90 normally looks like the image in our docs, with a picture of a plane showing which way to orient the vehicle.
I believe the word-based instructions only show up if you specify a different flight controller orientation. If the actual orientation of your controller doesn’t match the orientation you’re calibrating to then that can cause the calibration to fail (because it detects gravity in an invalid direction).
Thanks for your response Eliot. I was calibrating with the Roll90. The calibration steps did have the images that you posted. The word-based instructions appeared after the calibration steps were complete. I attempted six times without any luck. I even tried alternating the port and starboard orientation to see if that made any difference, but could not get a successful calibration.
I ended up disconnecting the power from the BlueROV2 and restarting my laptop. Lo and behold, when I powered up the ROV and restarted QGroundcontrol, I had the green light next to the accelerometer. I was then able to recalibrate compass, level, and pressure. I went through the pre-dive checks and all appears to be functioning correctly. Still unsure why I was unable to achieve a successful accelerometer calibration prior to rebooting. Any thoughts?
I will be diving today. I will let you know how it goes.
1 Like
Ahh ok, fair enough.
I don’t really have a better answer for you here than “electronics and software can have issues sometimes”. My best guess is that something unexpected happened during the first calibration that caused the sensor or ArduSub to go into some sort of fault state that it wasn’t able to automatically recover from, and powering down and restarting then brought the system back into a known state, which allowed the expected communication with and calibration of the accelerometer and other sensors to occur.
Good to hear - hope it stays that way 
Hope the dive went well!