Pixhawk 3.3FW

SkyHunter Mini

Maiden flight, all in autotune mode, on second wide RIGHT hand circle (1.5minutes into flight), pixhawk/plane BRISKLY rolls LEFT so far it's inverted and is does not responding to controls..

FULL right aileron the whole time, several seconds later pixhawk picks up my turn RIGHT command and starts turning right side up.

By this time it was too late and I was in the trees. Propeller and motor dont stop, they powered through the whole thing. Which tells me the pixhawk was somewhat alive, just not responsive.

All I have to analyze is telemetry logs, and i've spend hours pouring through it with no answers.. Im hoping someone else could take a look.

GPS stats are good

Battery is healthy

EKF flags 165 and 831???

no failsafes that i've found

@line ~8.16e6 in .tlog is approximately when things go bad

the pixhawk starts doing its own thing and doesn't datalog/teletrylog what is actually happening. data is no longer coming in via telemtry.... either telemetry connection lost OR pixhawk is rebooting

 

Any thoughts? .tlog attached. Any help is hugely appreciated, I cant solve this crash!!!

2015-09-06 18-17-56.tlog

2015-09-06 AT Crash1.tlog

2015-09-07 AT Crash2.tlog

You need to be a member of diydrones to add comments!

Join diydrones

Email me when people reply –

Replies

  • Developer

    Hi Nick,

    The SYS_STATUS.load value is from 0 to 1000, so a value of 115 means 11.5% CPU load. That is normal for plane on a pixhawk.

    Can you check carefully on the microSD card to see if there are .bin dataflash logs on it? A DF log would be much more useful than the tlog.

    So far the only thing I've noticed is that your model seems to be quite tail heavy. The trim servo position on the elevator is 1488, but you are flying around in level flight with the elevator at around 1380. That means that there is quite a lot of commanded down elevator to keep you level. That implies the plane is tail heavy (ie. the center of gravity is quite a long way off). Can you please check the CoG carefully?

    Ideally you would flick to manual mode for a couple of seconds when you are at altitude and see what happens. I suspect you'll find the plane noses up sharply when you do that. Make sure you have plenty of recovery altitude if you try that, as the plane may be very unstable.

    Apart from the fact that the crashes happened in autotune mode I don't see anything in the log to indicate that it was autotune itself that caused the crash. Autotune doesn't actually do very much - it very slowly changes the P gain on the roll and pitch controllers, but that can't suddenly make a plane invert like this. So right now my best guess is that the plane is tail heavy and stalled (as the old saying goes, a nose heavy plane flies badly, a tail heavy plane flies once).

    Do you have the rudder servos fitted? If so, have you carefully checked the rudder direction for stabilisation?

    Cheers, Tridge

    • I'd like to emphasize the double check for the rudder direction.

      I was flying a newish plane last weekend with basically its first autotune and had stuffed up the rudder reversal. In an untuned first flight in autotune it was not obvious that this was reversed (but in RTL it was really obvious). Behavior was very similar to yours but I had quite a bit more height and it sort of fluttered after initial roll. Stabilise didn't seem to help but I only had it on for a second (or less). Once in manual it flew out fine but I did have a few seconds of height to orient and pull out. This happened twice in a single flight. Ran out of battery and frequent family points before testing correct rudder reversal at the field.

      This weekends flight was much better. I didn't do much tuning due to the wind but I would have expected to see this in the time that I was up. Testing RTL did an amazing job for the amount of wind and lack of tune.

      Daniel.

  • And it just happened again. Patched it together and went back out.

    I've narrowed the cause down to be the AutoTune feature. This time I launched in stabalize, flew around and switched to autotune, momentarily after, wings inverted and spiral nose dove into the ground....

    why is autotune flipping me over upside down?

    is 115 a normal value for sys_status.load!? more than 100 seems not right...??? sys load goes over 100 when in autotune, 85 in stabilize.

    log and two videos to come

    • Had a very similar thing happen to me over the weekend. APM 2.6, soon as i went to autotune it did something similar, went to manual but not response. Will post logs and video up later.
  • This is specifically what I'm referring to.

    You can see my bank to the right in both CH1 in and CH1 servo out, all is well there...

    But then you notice attitude.roll go to -3 (to the left... uncommanded) this must have been when it was completely inverted to the left. I didn't hit anything, the video proves it was a clean barrel roll. Even more interesting is that everything just stops after that.... no data.

    eventually down the road the pixhawk restarts it's self???

    3702087495?profile=original

This reply was deleted.

Activity