Auto mode commands not followed

I have build a quad with a pixhawk 2. Have flown a number of successfull missions taking crop pictures. Without changing firmware or hardware one day on auto take off the quad just went sideways and crashed. After that I have been struggling for weeks now and cannot find out what the problem is. Have updated to latest firmware. In stabilize mode everything is working fine. Auto missions is completely unpredictable. Sometimes it would go to 2nd waypoint and then just start climbing, one time up to 350m even though geofence is set at 140m. This Saterday I went to airfield. Setup short auto flight and the first mission completed perfect from auto take-off to auto land. Then repeating the exact same mission the quad completed the mission to last waypoint and got stuck. After that same mission but after first waypoint it went up to 90m instead of the programmed 40m. Completed rest of waypoints and got stuck on last one again. Each time I had to switch to manual mode to land. Next time same mission again went up to over 90m on waypoint 2 instead of the programmed 40m run to last waypoint and then started climbing to 100m. I switched to RTL and it just kept climbing. Had to switch to manual to land. Does this make sense to anybody?

Probably a curse, was that near a pet cemetery?
(Unless you ask for something else than crazy guessing, post data flash logs.)

2018-03-11 08-28-58.tlog (329.8 KB)
2018-03-11 08-00-01.tlog (516.9 KB)

Here is 2 of the tlog files!AszASo_V73Jdh0J_QSdlNufbNfX_

Here is a link to some more log files.

are those on the microsoft server the dataflash logs ?
can you share using something that do not try to make me register with microsoft ? ( got rid of their mess/spyware since the 90’s) :slight_smile:

Hi, try this link its to a dropbox folder

which one is the event, you have some from February, some from March ?

looked at *32.28.bin

you have way too much vibrations in Z axis, also, you should not ignore that many EKF errors and warnings.
Fix vibrations. - balance propellers, fasten things, and/or mount AP better.

Props are balanced but I will check again. AP is mounted on the standard 4 rubber mountings. I will check out forums about what to do about EKF issues. All that said I don’t understand how the quad could do at least 5 fully automated flights without any single issue and then suddenly everything change. The flights that was successful were all field survey missions that scanned about 6-7 hectares each. After that I programed just a short circular flight to demonstrate it to a client and that was when the problems started. I originally thought I messed up in setting up the auto flight but after review it looked like everything was right. I then upgraded to the latest version of firmware thinking that was maybe the issue and after that did not work I wiped and reloaded the firmware. I will go and review the z-axis vibrations in the auto flights that was successful and compare it to these. Thanks for your input