Hi,
We had a crash last week when flying with AP 3.7 on a pixhawk 1. I’ve looked into the logs and think that I have found a sign of what caused the crash, so now I hope that someone perhaps can provide me with some insights on why/how this happened and how it can be fixed (or if it has already been fixed).
Here is the log: https://www.dropbox.com/s/pvt77vke4yjttpi/225.BIN?dl=0
A note on the setup: We are testing some guidance controllers from an external single board computer that sends desired roll/pitch/throttle to arduplane, using GUIDED mode, but I believe (or hope?) that this is unrelated to the crash.
The UAV (Skywalker X8) was flying straight, when it suddenly started rolling to the right. The pilot tried to take control of the UAV, by going into MANUAL mode, but felt that this had no effect on the UAV (he then flipped the mode switch back and forth, in an effort to regain control of the UAV, which is reflected in the rapid mode changes towards the end of the log). Fortunately, the crash was quite soft, and the only thing that needed replacement was a bad servo.
From the logs I see that roll/pitch follows the desired roll/pitch up to about line 857500 in the log, but then they separate. My interpretation is that the roll/pitch grows in magnitude (for some unknown reason) and that the desired roll/pitch then try to compensate for the growing error in the path following.
After the crash, the UAV started responding again after the serial between the pixhawk and the single board computer was unplugged. Also the faulty servo moved. When the UAV was put in STABILIZE and tilted from side to side, the response on the ailevons was extremely low compared to normal (a factor of 5 lower, or so).
What could be the cause of this all? Is it as simple as a servo being faulty, or are there any other possible explanations? Could it in some way be related to “jamming” of the serial link from our single board computer to the pixhawk?
Appreciate any input and thoughts you may have!
Kristoffer