Agro drone trying to flip during takeoff

ArduCopter Copter 4.3
1

Hi, dear community, I’ve got a problem and can`t figure out what is actually wrong. My agro drone hexacopter is trying to flip during takeoff. I tried to take off in stab mode, motors X8 are calibrated, props directions are correct, gps fix is ok, battery 12S is ok, sticks are calibrated, the fc is cube orange, it flew well previously on different drones. I attach the log file, will appreciate your help!
https://drive.google.com/file/d/1-RjuH8omaktqbpAvHbWKGu4H_LkKvnII/view?usp=drive_link
All calibrations - compass, IMU, level were done succesfully

2

99% of the time the motor order or prop direction is wrong. Check it again it is ABC, not 123

3

Thanks for your reply! I ordered motors according to wiki Hexa X frame, when I start motor test A-B-C-D-E-F they rotate 5-1-4-6-2-3. actually from the first to the last if to count them clockwise direction. I can`t find any explanation what motor numbers these letters have to match.

4

And …

  1. Does the prop rotation direction match the wiki?
  2. Will the props produce lift when rotated in that particular direction?
5

1 and 2 - yes, checked several times
Can power 5v not provided to servo rail affect pwm output to esc?

6

I changed the fc, and finally took off. But the attitude was unstable, especially in loiter mode, sudden movements in roll axis, I hardly managed to land it. I attach the log, would appreciate your advise. There was in-flight mag yaw re-allighnment
https://drive.google.com/file/d/1rAFyVviaYlKiwJ_gNGkuAiWghftRcewz/view?usp=sharing

7

The in flight yaw re-alignment is a normal message. No worries there.

Try reducing your rate PID values by 50% and trying again in stabilized mode. Pitch doesn’t look too bad, but roll has some big oscilations. See if you can hover for a minute then land it. Don’t try Loiter or RTL or auto, until the tuning is under control.

Hold on… Your voltages are all over the map. Make sure your battery voltage is calibrated, and it’s a really good idea to get current monitoring set up. Your voltage starts at 4.2V/cell but drops quickly below 3.2. And then it bounces as the flight goes on. Hard for me to say without current monitoring to validate, but given the duration of the flight I would say this battery is either bad, or significantly under spec’d.

Vibrations look good.

Compass is okay enough for now. Once you get it flying there may be some benefit to running MagFit and improving the calibration. But for the most improvement you’ll need current monitoring working and calibrated.

1 Like
8

Thanks for your reply Allister. The battery voltage pin is set up, the battery is 12S 60000mah and seems to be ok, but i noticed large voltage drop. The log shows voltage drop up to 10Volts, may it affect stable power supply of the fc?
I never tuned pids manually except initial built-in autotune in the mission planner, and the drones actually flew normally. What exactly roll oscillations you mentioned mean?
What PIDS do I have to reduce?

9

I found out that the voltage is actually good, it drops only 5v during flight, and is displayed correctly both in the mission planner data tab and on osd picture, as well as measured by voltmeter…but! in data flash logs it is displayed large pressure drop, 20vots! i wonder what is the reason for such strange behavior…the voltage is being monitored through the 3rd pin using voltage divider 1/20
The fw is the 4.4.0 the fc is Cube orange

10

The data flash log is probably recording at a higher rate than you see in the OSD, so you may not see the drops in voltage on the OSD.

https://ardupilot.org/copter/docs/initial-tuning-flight.html#initial-aircraft-tune

Per the wiki: Reduce the following by 50%



    ATC_RAT_PIT_P

    ATC_RAT_PIT_I

    ATC_RAT_PIT_D

    ATC_RAT_RLL_P

    ATC_RAT_RLL_I

    ATC_RAT_RLL_D
11

Allister I did as you wrote, actually it became worse, I performed a small crash. The logs are attached. can’t find the answer what is actually wrong with the roll axis
https://drive.google.com/file/d/1T9dXfx_nu9EX2OnjSZ_J41l2YSsPeX2P/view?usp=sharing

12

I still think there’s a power issue.

I also noticed that there were no pitch or roll inputs from you during the flight. In stabilized mode it is normal for the pilot to need to input something to keep the drone in position. If you did the test outside, and there was any wind (a breath) you need to make corrections for it.

1 Like
13

I found out that the ground of the voltage divider was not linked to the ground of the Cube power supply, cause its mounted on a different ubec. The problem with the voltage drop in logs is solved. The problem with roll instability still remains. Allister does the fc have to make the dron stable in the roll and pitch axis during hovering while there is no roll or pitch input?

14

I changed PID values the way you recommended, and tried to take off , the roll instability became even worse than was before. Will apreciate your help. https://drive.google.com/file/d/1cgkpL_plwXZJ41dozQgIupkzmzXRglvZ/view?usp=sharing

15

I think you will have to start back at default PID’s. A large craft like that typically needs higher than default PID values. Set them back, set these and make a simple hover flight for ~1 minute:
INS_ACCEL_FILTER,10
INS_LOG_BAT_MASK,1
INS_LOG_BAT_OPT,4
PSC_ACCZ_I,0.4
PSC_ACCZ_P,0.2

if you are sure the motor order and direction are right you need to get it in the air promptly in Stabilize mode to collect some useful data to review.

2 Likes
16

Dave, I corrected pids as you wrote in the list, increased roll and pitch pids and made 2 flights in loiter mode, because I was afraid to fly in stab, having not enough space for instant movements. The attitude became better but the drone still performs instant roll movements on the right side. Pitch seems to be ok. I attach 2 logs - with pitch and roll 0.2, and with pitch 0.2 and roll 0.27 Thanks a lot for your advise!
https://drive.google.com/file/d/1olVj4puFS4yCESerwy7VBv_Xsv9yoJFI/view?usp=sharing
https://drive.google.com/file/d/1fJlGg-DIt6TnpEwE-QjwrmZtO4eqwG3K/view?usp=sharing

1 Like
17

The problem was with the 1st motor. I repaced it and the drone flew perfectly. The rcout value for the 1st motor had much oscillations, seems like the ESC was damaged. The final PI values are 0.2 for roll and 0.16 for pitch
Thanks for the advise!