[drones-discuss] Auto: Strange loss of altitude and crash

96 views
Skip to first unread message

Maurice Barnes

unread,
Apr 18, 2014, 10:06:29 PM4/18/14
to drones-...@googlegroups.com
On a typical mission today with an altitude set to 50m after 2nd waypoint the copter kept reducing altitude until it eventually crashed into a tree. I am not able to determine what caused it so I am attaching my logs as well as my waypoint file. The only thing different about this mission was that I was testing a speed of 1800 cm/s. Can someone analyse and say what could have caused it.

Regards,
Maurice

P.S. Using the latest Arducopter from Master and the latest MP. I also have video of the flight I could upload if that will help
Polo_Circle_high_speed.txt
2014-04-18 16-49-08.bin
2014-04-18 16-49-08.kmz
2014-04-18 16-49-08.log

Jonathan Challinger

unread,
Apr 18, 2014, 10:13:04 PM4/18/14
to drones-...@googlegroups.com
Looks like you told it to go too fast. I know we used to have some protection from overspeed causing descent, but I'm not sure if we still do. Throttle goes to and stays at 100%.


--
You received this message because you are subscribed to the Google Groups "drones-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to drones-discus...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Randy Mackay

unread,
Apr 18, 2014, 10:41:41 PM4/18/14
to drones-...@googlegroups.com

 

     In AC3.0.1 ~ AC3.1.3 we have protection in Loiter mode but not in Auto.  Basically a constraint was in the wrong place which allowed roll/pitch to overpower throttle.  It’s fixed in master and will go out with AC3.2-rc1 which could be as soon as the end of April.

 

     Too little too late for Maurice perhaps so sorry about that!

 

-Randy

Maurice Barnes

unread,
Apr 18, 2014, 10:51:57 PM4/18/14
to drones-...@googlegroups.com
Hey no prob.. No one was harmed in the process and I am glad I can contribute to making it more robust. I have at least six test cases each day spanning various aspects of the system so I am bound to stumble upon something. I will test the fix in a more controlled environment :-)

Maurice

Maurice Barnes

unread,
Apr 19, 2014, 8:01:07 AM4/19/14
to drones-...@googlegroups.com
Randy,

Can I retest this at the same speed?

Maurice



On Fri, Apr 18, 2014 at 9:41 PM, Randy Mackay <rmac...@yahoo.com> wrote:

Robert Lefebvre

unread,
Apr 19, 2014, 8:19:13 AM4/19/14
to drones-discuss
Ah, good to know.  I always thought this protection was in place.  My mapping copter is a bit portly and doesn't have a lot of reserve power.  Actually hoping to do some missions with it today.

Randy Mackay

unread,
Apr 19, 2014, 9:55:43 AM4/19/14
to drones-...@googlegroups.com

Maurice,

 

     Feel free to test master but please remember that it’s still under development so changes are going in every day and I can’t promise it will fly ok.  I’ve been flying it and so have a few others and it’s generally ok though.

-Randy

Maurice Barnes

unread,
Apr 20, 2014, 11:36:59 AM4/20/14
to drones-...@googlegroups.com

Randy,

Did some more tests and the problem still exists. . I started at 50m and it dived to 20m before I recovered manually with a rtl. Speed was set to 1800 cm/s  with accel at 480 cm/s/s

A point to note is that the loss of altitude is more pronounced when flying into wind.  I will also admit it was a bit gusty probably 20km or so.  I can post logs later.

Regards,
Maurice

Randy Mackay

unread,
Apr 20, 2014, 8:11:36 PM4/20/14
to drones-...@googlegroups.com

Maurice,

 

     Ok, thanks for testing.  With the logs I’ll be able to dig into it further.  I can also try it in the simulator although not immediately.

Maurice Barnes

unread,
Apr 20, 2014, 10:56:26 PM4/20/14
to drones-...@googlegroups.com
Randy,

See logs attached. If there is anything else you need please let me know. The incident occurs between 4925 and 7677 (approx).

Regards,
Maurice

2014-04-20 10-01-51.log
Reply all
Reply to author
Forward
0 new messages