Uncontrolled climbing (throttle down)

Hi Mates!

A dream of my childhood has finally come true. About a month ago I finished building my first drone. It’s a quadcopter based on DJI F450 frame with original Pixhawk (ArduCopter v3.2) flight controller + external GPS, with Open RLS 433Mhz (openLRSng v3.7.4) Rx and Tx, and Mobius action camera. I used HK ESCs 15-18A for hacker Style 20-26M motors, and Turnigy 9x as a controller.

So, I’m a newbie in RC, but I have a great desire to learn. I had just 3 flight session so far.

The first one completed in 10 seconds with a crash due to wrong programmed ESCs. Two of them were set Brake ON, another two - Brake OFF, what caused a flip right before landing. As a result 4 broken propellers and pulled out GPS mast.

The second flight session was rather successful. But I was scared by the first accident, so I tried to be careful as much as possible. I flew few meters above the ground. Noticed that pitch stick is inverted. Gave a try to different flight modes, but the copter wasn’t too responsive in modes other than Stabilize, + Simple, + Super Simple, so I switched them immediately off (AltHold, Loiter, etc).

The third session consisted of 2 flights.
After getting familiar with Loiter mode on the Internet, I decided to start a flight within it. Right after the arming, I pulled Throttle stick slowly up. The copter started taking off. After a while it seemed to be too high for me, so I started pulling Throttle stick slowly down to the very bottom position (and hold it in that position almost till the end). The drone didn’t listen to my commands. I started panicking - my hundreds of dollars are flying away from me, and I can do nothing. I pulled Land mode switch - no reaction at all. Then I noticed that it didn’t listen to the Throttle stick, but somehow reacts on pitch, roll and yaw controls. Probably the height was very high, and wind was really crazy at that altitude, so the copter wasn’t very responsive, but I managed to hold it more or less above me. After a minute of such crazy dances, I believe , it received my Throttle down command and turned motors off… fell down from about 25 meters… No serious damage though.
The problem here is not a fact of the crash, but uncontrolled climbing that caused a crash then.
You can see on the first graph (attached) that ThrottleIn was going down, while ThrottleOut kept on raising. When ThrOut followed ThrIn, it was rather smooth, then suddenly started fluctuating - that’s the moment of a kind of “glitch”.
Below you can a video from the copter, log file and a graph are attached to the post:
youtu.be/RxCGasv5uvc

Then I decided to fly again, but thousand times more careful. I thought that then I could manage to avoid such a wild behaviour again. I flew a dozen of seconds right above the ground and then BZzzzzzz… an it’s again in the clouds. I tried to follow the same strategy, but I wasn’t so lucky to “catch the signal”, so the best I managed to do is to direct it to the trees. One propeller and GPS mast were broken. Motors full of dirt.
Unfortunately, I don’t have video of that flight, but some a graph and logs are also attached.

I’ve spent a day analysing the log files, but couldn’t find anything specific.

Could you please take a look on what I wrote here and help me, if you see the reason of the issue? I’m really scared now to have any flight, until I’m 100% sure that the issue is resolved.

Looking forward to your assistance. Thanks in advance!

Logs:
2015-01-25 15-25-09 26.bin.log
2015-01-25 15-25-31 27.bin.log

How have you mounted the APM to the frame? your Z axis vibrations are VERY high and could have caused your problems. Also looks like you have an under specced battery as the voltage is being pulled down to 9v very early in the flight.

Just found two more interesting graphs, where ThrIn and ThrOut are synced at the beginning. After a while ThrIn goes down, but ThrOut starts fluctuating, but steel goes up, chasing the DAlt (desired alt). Why does the copter choose so wired desired alt???

Thanks for a quick response!
Pixhawk is mounted inside of F450 (to the bottom plate) using adhesive foam, that was in a Pixhawk kit.
What do you mean by “under specced”? It’s a Turnigy 3000mah 3S 30-40C.

your first log is fail safe city, voltage and radio. You have a couple of RTLs which will make it climb as it tried to come home.

MarkM is right, your battery voltage drops very low.

edit…both logs show lots of fail safe action

mission planner auto analysis of first file

Test: Autotune = UNKNOWN - No ATUN log data
Test: Balance/Twist = GOOD -
Test: Brownout = GOOD -
Test: Compass = FAIL - Large compass off params (X:-38.00, Y:-54.00, Z:248.00)
Large compass offset in MAG data:
Z: 248.00
mag_field interference within limits (13.84%)

Test: Dupe Log Data = GOOD -
Test: Empty = GOOD -
Test: Event/Failsafe = FAIL - ERRs found: CRASH GPS_GLITCH FLT_MODE FS_THR
Test: GPS = FAIL - GPS glitch errors found (1)
Min satellites: 0, Max HDop: 2.14
Test: IMU Mismatch = FAIL - Check vibration or accelerometer calibration. (Mismatch: 4.76, WARN: 0.75, FAIL: 1.50)
Test: Parameters = GOOD -
Test: PM = GOOD -
Test: Pitch/Roll = FAIL - Roll (164.21, line 7094) > maximum lean angle (45.00)
Test: Thrust = GOOD -
Test: VCC = GOOD -

file 2

Log File D:/Users/Stu-Desktop/Downloads/2015-01-25 15-25-31 27.bin.log
Size (kb) 2418.5078125
No of lines 33383
Duration 0:02:22
Vehicletype ArduCopter
Firmware Version V3.2
Firmware Hash c8e0f3e1
Hardware Type
Free Mem 0
Skipped Lines 0

Test: Autotune = UNKNOWN - No ATUN log data
Test: Balance/Twist = GOOD -
Test: Brownout = GOOD -
Test: Compass = FAIL - Large compass off params (X:-38.00, Y:-54.00, Z:248.00)
Large compass offset in MAG data:
Z: 248.00
Large change in mag_field (44.84%)

Test: Dupe Log Data = GOOD -
Test: Empty = GOOD -
Test: Event/Failsafe = FAIL - ERRs found: FS_BATT FLT_MODE FS_THR
Test: GPS = GOOD -
Test: IMU Mismatch = FAIL - Check vibration or accelerometer calibration. (Mismatch: 4.88, WARN: 0.75, FAIL: 1.50)
Test: Parameters = GOOD -
Test: PM = GOOD -
Test: Pitch/Roll = FAIL - Roll (-86.28, line 21424) > maximum lean angle (45.00)
Test: Thrust = GOOD -
Test: VCC = GOOD -

Your copter is unsafe to fly at the moment in my opinion

Hi all,

I did some more tests, and sorted out that the copter flies well in Stabilize mode.
But after manual or automatic (e.g. due to Failsafe) switching to any AltHold mode (AltHold itself, Loiter, RTL, Land, etc.) it starts rapid climbing.

Does it make sense? Any ideas, any guesses?

You need to improve the mounting and/or balance your props and motors as your z vibrations are too high.

Thank you, MarkM! I will definitely try to change the way the copter is constructed and adjust controller’s vibration dampening. I’m aware of props balancing as well.
But what did you mean by motor balancing? What can I do about the motors?