I have two quads with the same problem and entirely different hardware.
F450, 2830's, Full size APM, 10x4, BEC, 3.15, X, baro covered, (ESC, Radio, Acel and Compass cal'ed)
Blackout mini H, 1806, min Apm, 5x3, BEC, 3.15.X, baro covered,(ESC, Radio, Acel and Compass cal'ed)
Logging full, but happens with all logs off.
In hover after random time period quad just falls suddenly.
Studying the logs looks to me that:
Vibrations within spec
hdop 1.8/8 at illustrated event
acel ok, compass ok
Approaching an event, throttle-in constant in hover, throttle out normal
Motor 1 is turned off
Motor 2 goes full
Quad flips.
Graph of one event attached and logs with multiple repeatable events.
Thanks for your time.
Replies
Your experience is why I decided not to fly my big birds until I got all the kinks out. I got a blackout 220mm mini quad frame, after the F450 crashed, where the props are 50 cents. etc. to act as a test bed. I am really glad I did because it has taken some major punishment.
Just for completeness in this record. I use optoisolated esc's that don't have bec on them and I have a big cap on the 5v line output of the ubec to act as a buffer to prevent brown outs. Which makes the power issues seen by the fc more strange.
The really hard part is not knowing if the issues are operator error or flaws in the design of the fc. I will readily admit that there are probably more opportunities for operator/build errors than coding errors. But when I had a real job my worst nightmare was when after troubleshooting for weeks Microsoft admitted they had a bug. Then there was nothing I could do to fix the server.
Just took all apart. vcc on the board at the pm connector is 5.2v. So what ever is happening to vcc it is happening on the board. Out of four boards they are all surprisingly different measuring across to the input rail. Some as low as 4v and most around 4.7v That is with everything disconnected. Just power to the pm pins and measuring at the input rail for the receiver.
I have decided for my own peace of mind I am going to separate my APM hardware from Arducopter. I am going to replace my APM with a Virtual Robotix Vr Micro Brain without changing anything else. That way I can know for sure whether there is something wrong with my build or start to troubleshoot whether it is a problem with the APM hardware or Arducopter firmware, since the VIrtual Robotix runs Arducopter but is an entirely different hardware design.
Cheers
I replaced my FC with another APM (not the micro brain) and the problem has gone away.
Interesting. So which brands were you using, and what are you using now?
The Vcc thing in particular is very odd, if it's supplying 5.2V but the board is actually measuring low 4's. Has there been any circuit hacking going on?
The vcc thing is very weird Robert. The sleeper thing is that it works and flyies but seems a good candidate for the sudden failure when the voltage drops low enough to cause logic levels to get confused. I am going to replace the ubec in case it is acting strange under load. But the drop across the board is true when powered up in a static lab situation and that doesn't seem right regardless. I need to get out all the boards and compare them. I have samples of all of them.
Cheers
Just for completeness I am posting the logs from my F450. Again completely different hardware but as the image of the motors graph it looks amazingly similar in failure.
No vibration problems, no compass problems, no vcc problems. crash anyway.
Cheers I am outa here...
2014-08-08 11-04-22.log.xml
F450vsBlackout_Failure.JPG
Mike,
I think that's the .xml file that's sometimes created by the MP for some reason. It's the .log or .bin file that's most useful.
Video of Blackout failing.
http://youtu.be/En7AFry3Ao4
Sorry good old windows was hiding the extensions and did not see the xml part. Here is the correct file.
I have flown the blackout enough times now that it seems the fc was defective, because I don't see the failure, but I don't see those same problems in the F450 logs.
It is not important to me to figure out the F450 since I will move on to building a bigger quad so don't spend time on this for me.
All the best
2014-08-08 11-04-22.log
-
1
-
2
of 2 Next