Hi,
I've flown my fixed wing with APM:Plane firmware successfully for over 30 missions. Just earlier this week I experienced a crash that has rattled my confidence. I have some theories but hopefully you can help me reconstruct what happened by looking at some of the data. I feel like someone at NTSB trying to analyze data from a black box and have spent countless hours pouring over it. I've looked at numerous wiki posts on ardupilot.com, but have been unable to uncover the exact cause.
Right now my strongest theory is that a high voltage transmission line caused some sort of interference. You'll see below why I believe this is the cause, but hopefully someone can educate me based on the data from the telemetry log. Unfortunately I was unable to obtain the dataflash log.
This is actually a mission I've flown a few times before to help a friend get some aerial imagery of his property. Two weeks ago I flew this mission at 170m and he asked if I could fly a little lower. So I flew the exact same mission at 120m instead. The screen below shows the flight path in Google Earth. I drew in the white line to represent the transmission line. The flight path goes from left to right so you'll see that the line was flown over a couple times before the crash. The crash is on the right just south of the transmission line.
Now, check out the photo from my UAV right before the crash:
https://www.dropbox.com/s/8oreumxh3qn04qo/IMG_2066.JPG
At the top of the image you'll see the transmission line and that the UAV is high above it. This image is taken heading from north to south. As seen in the Google Earth photo above, the UAV is heading north-south when it crashed in the field. Here are a couple more photos in the sequence where the crash begins. All photos are taken on a 3s interval:
https://www.dropbox.com/s/6clhfaixvzw6zrn/IMG_2067.JPG
https://www.dropbox.com/s/rv5kttl53j4xo9k/IMG_2068.JPG
You can see that it all happens right after passing over the transmission lines. I'm not sure if that's coincidental, but if I knew how to look for the right information in the logs it may be apparent that there's interference. I've gone through the steps below but nothing conclusive jumps out at me:
http://copter.ardupilot.com/wiki/common-diagnosing-problems-using-logs/
Finally, here is the tlog:
https://www.dropbox.com/s/7p5upg72fi2y6fz/2014-07-03%2009-03-56.tlog
and the same log converted to text:
https://www.dropbox.com/s/ijksjavsqxzfx8t/2014-07-03%2009-03-56_tlog_to_txt.txt
I'd greatly appreciate any insight that you may have to offer.
Many thanks,
Dennis
Replies
hi Deniis, i am subscriber on your channel youtube, can you help me, for analisys my tlog ?
wp 1 = 15 meter
wp 2 = 15 meter
wp 3 = 15 meter
wp 4 = 15 meter
wp 5 = 10 meter
first my quadcopter is take off manual until 23 meter, and then i switch to loiter, everything looks good, and i switch into auto mode, first way point reached, bot on wp 2 copter is fall down...
have you any suggested for this cases, please look my attached tlog
2016-07-16-09-19-15.tlog
Hi, I don't know much about the logs from APM as I am just setting up my first APM at the moment as I usualy use eagle tree.
As for the clevis breaking I strongly don't believe it was the cause of the crash as I have a wing I use for FPV and I had a clevis pop off mid flight and the wing went into a death spiral as only one elevon was working and didn't lose altitude in the level position.
The reason I know the clevis came lose mid flight was that I had a Mobius camera mounted sideways under the wing for look at wing flex under high g moves and caught the clevis coming lose on the camera (clevis didn't break it just popped off).
Hope this helps.
I wish i could help, but it has been a few years since I looked at tlogs. Maybe I will try anyway just to see if I can find anything for you.I am pretty disappointed that you have got got any replies yet.
Here is your YouTube video about the crash in case the extra info helps anybody find something useful in the logs:
https://www.youtube.com/watch?v=oOwx7G9reNM
Hey Derek, thank you for the response and watching the vid. It's much appreciated. I wish I could have obtained the dataflash log, but it was long gone when I got back to the house and hooked up. One interesting bit of info is that I flew this same mission yesterday, with a new clevis, and no problems whatsoever. I'm still sticking with the theory that the cheap nylon clevis broke midair and APM did it's best to get me back on the ground with minimal damage.
I agree. I think the clevis is the cause too. I am trying to view the tlog slow enough to see if the HUD or map show any signs of this. So far, I am seeing the link line in front of the UAV icon on the map side seem to show that the plane was trying to turn a hard right but could not.
Unfortunately I cannot find what the pink like should represent in Mission Planner. It is not documented (as is always an issue with Mission planner, in my experience). I can find what all the other lines there do from another person that asked here in the forums. (green is nav path, black is course over ground, red is current heading, orange is path to next wp)
Investigating issues like this is one of the reasons I have not flown my AMP quadcopter is over a year :(
So, if you move your mouse around the map on Mission Planner you can get it to update the lines on the map faster. Looks like my theory was likely just due to a mission planner issues with displaying the lines. Watching it slower it looks okay.
I see no huge lean to either side on the graphing of the Pitch, Roll and Yaw.
I see a clear from in Pitch.
I see a clear drop in Altitude.
I see no issue with GPS satellites (as you said in your video).
I see no issues with the Mavlink connection to telemetry/Mission Planner.
Good stuff, Derek! I looked at this same data and couldn't make any progress on detecting a failure. Which still leads me to believe it wasn't APM-related. Sure wish I had those dataflash logs. I appreciate you taking the time to dig into this. Just a side note that I was able to fly this same mission over the weekend and had no problems.