Lane switch on compass_learn=3 flight

I decided to look at the log file of the flight yesterday when I did the in-flight compass calibration.

Nothing much interesting except there was one EKF3 lane switch.

The largest innovations are in SV - velocity related. Pretty interesting since I was just hovering around in my back yard.

In case anyone wants to see if it offers anything valuable to the 4.2.0-rc1 debugging, I uploaded the log file.

1 Like

I think this relates back to our previous conversation about COMPASS_LEARN vs MagFit.

You clearly achieved better results with the MagFit results, so why fly with the COMPASS_LEARN parameters?

Thank you Yuri for keeping up with my posts.

It relates to past conversations about COMPASS_LEARN vs MagFit only in this was the flight where I conducted the COMPASS_LEARN compass calibration. I posted this because I wanted to provide whatever support I could contribute to the release candidate testing.

By the way - I never said I didn’t fly with MagFit calibration. I’m merely reporting my observations.

As info - to understand the differences better, you’d have to run MagFit again after making the MagFit calibration changes. When I’ve done this in the past, the “before” charts always show some need for additional changes in calibration. I can imagine a few reasons why this might be true - but what I imagine doesn’t matter.

My question about the COMPASS_LEARN vs MagFit is in what way are they different?

My questions about this log is why was there a lane switch and was SV the cause.

Another anecdotal observation - this flight seemed to take quite a bit longer than previous COMPASS_LEARN flights with previous releases of firmware. Perhaps even twice as long.

1 Like

Thanks for the testing, I’ll pass this onto the other devs too (probably primarily @tridge) to see if they have any comments.