I wonder what happened here:
The flight was right after a perfect one:
with no parameters changed, as far as I remember. This was in May 11 - it took over a month to find the airplane and some more time to recover it ;)
The plane had an APM2, with ArduPlane V2.33. The flight plan in the crash flight was flown before and was safe wrt terrain obstacles (here: Flight plan). The plane crashed 2/3 of the way from WP #3 to #4.
Has anyone seen this kind of erratic flight (altitude and speed) before? The crash occured because the plane flew much lower than in the plan and with too little throttle. Just a few secs before the crash it appears to make a last-ditch attempt to recover altitude but too late.
It was 28 m up in that tree, and was recoverd by shooting golf balls at it from at compressed air cannon :( The plane was shot to pieces, but amazingly, the APM2 and the GoPro are still alive. The APM2 delivers plausible values from all (IMU, mag, pressure, airspeed) sensors.
I will dig out the flight log as soon as I find out how.. (does not seem to work in Mission Planner).
Nice vid. Lots of odd behavior it seems. Wish I knew enough to offer some practical advice, but better luck next flight.
Start watching vid at 6 minutes, crash is at 6:30 in.
From Leimbach to Rote Fabrik, that's quite ambitious, considering that you have to keep visual contact with your drone :)
I lost an EasyStar a few months ago, with similar symptoms. You can read about it here: Hardware or software glitch? - ArduPlane 2.28. It happend at Pfäffikersee, next to the Badi. I have read here different reports about similar (?) malfunctions. This has led me to believe that APM is just not totally ready for prime time, unless you can afford multiple backup and safety features.
Actually, I don't care about legislation too much. But the plane crashed about 2 meters from the rear of a parked car. I don't want to imagine the consequences...
BTW, what kind of radio technology do you use to maintain (telemetry) connection beyond line of sight?
Just for completeness:
The XBees were Series 1. Series 2 may be just as good range wise but have the disadvantage that the are "paired" using their serial numbers. In series 1 you just set IDs as you please, and pair on those. This way one can use one ID for "ground station" and another for "aircraft", and then use any station with any aircraft (as long as you fly only one at a time).
The ground station antenna is a biquad on a 125x125mm reflector. Designs can be found on the net, for WLAN range extension etc. The range is like 4 or 8 times that of my Futatba R/C. It is possible to send commands like flight mode to the drone via Mission Planner/Mavlink, so I have some autopilot control beyond RC range. On my unfortunate mission however, it failed. Later I found that part of the antenna connection had broken.
I will build a phase and impedance matched quad bi quad antenna next, and see how far I can get a reliable link with that... also I am working on a box with an LCD display that sniffs Mavlink and reports the last known position, course, speed, altitude etc (yes I know one exists already),
Kind regards, Søren