Everything was perfect with my octa quad and pixhawk and arducopter 3.1.2 since this big crash in auto mode
Few seconds after takeoff, the pitch increased to 45° and speed 25 m/s ! but loiter speed and nav speed were set at 12m/s
[attachment=0]2014-02-24_151216.jpg[/attachment]
I’ve done several times this same mission with apm 2.6 and pixhawk with firmware before 3.1.2 and no problems
According to the logs it's not an hardware problem or vibration problem
If someone could watch at the logs, It could be very helpful for me
I hope to regain confidence in the pixhawk
thanks
the logs are here: [dropbox.com/s/nyp1czm2f6jwh ... drones.rar](https://www.dropbox.com/s/nyp1czm2f6jwhz6/crash%20logs%20diydrones.rar)
Judging by your .log file, you were in mode 0 all the time, stabilize, and pitch got very high, but that was also the desired pitch. To me, it looks like you flew it at time of crash - no sign of anything other than expected result. - you did not yaw around, (and considering no log of RCin), I can ony guess of pitch channel being reversed, or your brain being offline for a moment
(no offence intended) - but you were in “stabilize” - a very simple mode where no GPS, or magnetometer, or preset speeds apply - also - you thought you were in some auto mode which indicates that your situation awareness were imperfect.
I was in Auto mode from 10:49 (tlogs) or 4100 (time number on .log)
You are right, almost all the flight was in stabilize but the last minute was in Auto mode (takeoff in Auto mode until the crash)
ooops - my bad.
I used APMplanner to view the logs, and there’s apparently a bug that does not plot mode change.
yes, line 4056 confirms that you switched to auto…
your WPNAV_SPEED is 1200 (12m/s) , about 43km/h.
I guess your setup is unable to maintain altitude at that speed (required pitch) - and Pixhawk did what it was supposed to do,trying to achive the set groundspeed.
One can argue that altitude should have priority.
(BTW: as you see, loiter speed does not apply to auto mode.)
My setup is 400W (MN3110 T-motors) x 8 motors for 3.5kg copter with batteries
I used 12m/s for NavSpeed in all my missions and no problems with more than 50 flights
and the speed in auto mode was more than 20 m/s before falling so the 12m/s were reached
at 25m/s (90km/h) the copter hit a tree and is almost 60% destroyed …
Do you think that there is a link with ROI setup in mission ?
Well, the data speaks for itself:
As pitch increases to achieve the speed, before it’s able to reach max speed, altitude drops - Arducopter even maxes out ThrOut.
It means, at this angle, (used to achieve your desired max speed, and your speed increase) - even “full power” could not maintain altitude.
maybe you had heavier battery or payload than the other day, maybe headwind. but the facts are there
(remember that the target speed is groundspeed, if you had tailwind, much less pitch would be needed to achieve it than headwind)
-please use new forum at ardupilot.com - this old mess of a forum can’t handle wide pictures or filenames with “:” in time.
I had same playload than before (Gopro + 2 batteries)
But that day there was little headwind no more than 10km/h
I’ve already done this same mission with a strong wind (40km/h) with no problems
[color=#800040]My problem is that I don’t understand why the copter didn’t get the 50m altitude before going to the next waypoint ??[/color]
[attachment=0]2014-02-26_063131.jpg[/attachment]
For each mission I use this first command to get enough altitude and then move the copter horizontally
And that’s what the copter did but not this time
Hi Stephane, I had a look at your logs. Something strange happened that I can’t figure out. It seems like the Autopilot had it’s directions reversed, and it went the wrong way, but I have no idea how this could have happened.
I’m going to try to get a few other guys to have a look.
Thanks Rob
Could it be a compass issue ?
I remember that the copter was directed to the south when I first put it on the ground before takeoff
Is there a way to compare internal and external compass logs ?
Before take off I got a bad compass health message, I unplugged and plugged the battery and the message disappear
But the copter was ok in stabilize mode and what is the link with the fact that the copter didn’t climb at 50m after takeoff ? (does it needs compass to do that climb?)
[quote=“stephane71”]Thanks Rob
Could it be a compass issue ?
I remember that the copter was directed to the south when I first put it on the ground before takeoff
Is there a way to compare internal and external compass logs ?
Before take off I got a bad compass health message, I unplugged and plugged the battery and the message disappear
But the copter was ok in stabilize mode and what is the link with the fact that the copter didn’t climb at 50m after takeoff ? (does it needs compass to do that climb?)[/quote]
Your heading was north the entire flight. The GPS confirms that this is the problem, showing acceleration to the north while the accelerometers rotated by the attitude solution show acceleration to the south, as shown in attached graph. Check compass orientation. Note that this could cause crashes even in stabilize or acro, as the AHRS relies partly on compass.
Edit: Looks like it was climbing just fine to me. Just not holding position, which yes, it does need the compass to do.
Edit2: I know randy wrote a guide somewhere for checking compass orientation. I’ll ask him about it. Simple way is to point it north and check the heading on the hud. Probably won’t help you determine if the compass is upside-down, though.
i was just reading what your did and it sounded very similar to what mine did. here is a link to my thread on the log forums, i have video there showing the crash, perhaps you can look at the video and see if yours did the same thing, it would be very interesting to see if it did. http://ardupilot.com/forum/viewtopic.php?f=80&t=6358
Who labelled this as Solved? Doesn’t really seem we have root cause yet.
I’m wondering if this could be a situation where, there was an external mag in use, oriented in a different way from the internal. The the external failed or something, and the Pixhawk started using the internal one, without any indication to the operator. But with the orientation being set for the external compass, could cause this failure.
