I experienced my 2nd crash with my APM2.5 last evening. The first crash destroyed a super sky surfer then last night destroyed an X8 Wing. The exact same scenario. Took off in manual mode (never switched out of it), flew one circle, during the second circle the APM took over, banked full right, went to full throttle and dove into the ground at full speed destroying the plane. I was using telemetry at the time and have logs on my laptop but not sure how to read them. I've attached them to this thread.
Haven't looked at the logs yet, but have you had successful flights with APM2.5? ie: you tested stabilization throws on the ground to make sure they move right, and have a proper failsafe set up on your Tx/Rx?
Off hand it sounds like you hit a failsafe, and didn't have it set for RTL or something more manageable.
What Tx/Rx are you using?
The only APM2.5 flights were documented in my first post. However, I did have an APM2 that worked fine. I did the ground test to ensure that stabilization functioned properly based on orientation of the aircraft. I did have failsafe setup for throttle monitoring and tested this on the ground and it worked. I was using an OrangeRX 9ch receiver from HobbyKing and a Spektrum DX8 transmitter as well as a OrangeRX satellite receiver attached to the 9ch and an additional UBEC in addition to the BEC that's part of the ESC. I had 10,000 mah on board via 2 5000 mah 3s packs and my flight was less than a minute long so I'm ruling out depleted batteries. All that said, I was nowhere near out of range for the RX, less than 50 yards away when this happened.
Did you have radio telemetry set up? Were you using the latest mission planner with the failsafe screen? Did you validate that the plane switched to RTL mode on signal loss?
Yes, Yes & Yes. Mission planner 1.2.12, APM 2.65.
I think you are going into failsafe (RTL). I'd set the plane (no prop) somewhere down the street, and walk/drive away with the radio and laptop and see if it doesn't go into RTL mode right away.
That's worth a try. I'll have to build up a mock plane to try this since the X8 was completely destroyed.
Can't see much at all on the tlog, doesn't actually look like a flight, can you get the last .log file off the APM with the mission planner?
Attached is what I was able to download from the APM.
Under what conditions should the APM swtich out of manual mode on it's own?
The last reasonable size log is 2012-09-27 12-36 19.log, Roll is the red line, does this look like your flight and crash? It's hard to tell what was the flight with the crash except this one where roll is fairly extreme right at the end. Otherwise this log doesn't seem to show much else, maybe one of the devs can help.
As far as I know should never switch out of manual mode if not commanded to do so, maybe APM was not have involved at all and it was something else like a range issue or interference or connection break problem.
Additional note - I'm aware of the issue with the 2.5 and have my APM in the official case offered by diydrones.
Sorry about your loss. I know it doesn't help you but this reminds me of the point why one may consider APM1 be the more appropriate autopilot for planes due to its hardware multiplexer. I'm fairly sure that this crash wouldn't have happened in that case. I think it has been discussed earlier that there are also external RC multiplexer available that could at least help in early stages (every plane setup is different).
On a more productive note: I haven't looked at your logs but two things that would come to my mind are the (relatively new) throttle failsafe and maybe a brownout of your board. Which part of your setup (UBEC or internal ESC voltage regulator) powers your APM?
My ESC was supplying power to the APM while my UBEC was supplying power to my Rx - which would also be applying power to the APM.