FINALLY fix mag rotation issues. #7366
Conversation
|
@matanhavi @Tubme please test this PR thoroughly with all the problem cases you found. @LorenzMeier I also realise that we can cut our loop times for voted_sensors_update in half (because I made small optimisations over the last few PRs which remove bad assumptions preventing this previously). We are currently super-wasteful with nested loops in the code. Something to remember after this fix goes in. |
|
@bresch please test this :) |
|
I tested with the same configuration, and the PR fixes the issue. |
| @@ -66,7 +66,7 @@ VotedSensorsUpdate::VotedSensorsUpdate(const Parameters ¶meters, bool hil_en | |||
| // initialise the corrections | |||
| memset(&_corrections, 0, sizeof(_corrections)); | |||
|
|
|||
| for (unsigned i = 0; i < 3; i++) { | |||
| for (unsigned i = 0; i < math::max(ACCEL_COUNT_MAX, GYRO_COUNT_MAX); i++) { | |||
There was a problem hiding this comment.
This is wrong: i stands for the axis index, not sensor index.
* sensors : second cut at fixing mag calibration - remove old problematic code * sensors : use more intuitive naming for loop variables
|
This does not fix our issue. On the X2.1 the external mag0 was 180 degrees offset (the GPS + Mag physically are oriented toward the front). Even more, The CAL_MAG0_ROT was at Internal Mag (-1) and I could not change it! I tried a couple of times to give Yaw 180 degrees without success. I tried forcefully to change it with param 4 but again it did not store my choice. It seems like the QGC accepts the change but then it stays back to Internal Mag. I know there are 3 internal mags in X2.1 and 1 external. I don't see a MAG3 in Parameters... Forgot to mention we flashed yesterdays master... |
Fixes #7358.