Fly-away in manual mode

I have a fly-away several seconds after takeoff that resulted in a bad nose landing.

Took off and had control for a few seconds, then my taranis reported “rssi level critical” and the plane slowly turned and then nosed in. Never changed a mode, and cannot confirm a brownout. Flew in same field last week fine.

3DR Products
-pixhawk
-lea-6h
-digital airspeed
-v2 telemetry 915mhz

FrSky
-x9d taranis
-x8r rx

3.0.3 firmware

Logs and graphical data can be seen here
http://www.droneshare.com/mission/14481

And here is the taranis log
https://www.dropbox.com/s/roedesms1ayug1q/SurvAir_1-2014-08-16.csv

Looking at the log it goes into RTL for no reason, doesn’t even declare a mode change (RC9 is the mode channel). The log also abruptly ends at the RTL event but the kmz shows several more seconds up till the impact.

Obvious question is the transmitter battery charged? Its not a flyaway BTW its a crash :wink: Did the receiver disconnect at all?? The aircraft was going very quick, what is it?

Yes the transmitter is charged. Shows 12.4v right now, 3s lipo. And it’s a penguin v2. The rx connector is still seated firmly on both the rx and the pixhawk. Powers on and works too, even though it is missing one antenna now.

It entered RTL due to your low battery failsafe. (line 284197 and 8)
It cannot be trusted to fly autonomously, because your compass is not realiable at all, too much influence by current/bad calibration.
You could have re-taken by switching to any mode, then back to stabilize.

That makes sense, but that would only have been due to instantaneous voltage drops from throttling up. I have a turnigy supercap but never installed it. Needless to say it is going in when spare parts arrive to rebuild, and I am turning off that voltage fs as well as current fs.

Thank you Andre.

I still do not know why the graph ends at the RTL event though.

your final baro.alt record is 2.277814m , that is during a steep descend at , on line 294521
it seems to me like it descended 10m in about 600ms before final log lines,

  • so ending at 2.2m (provided that terrain had the same elevation at crash site, is not unusual.) - there is no sign of trouble with VCC.
    You can safely assume that the log ends because it did not were written to SD yet, or file FAT table was not updated with that change yet, when you got a brownout or loss of power due to crash.