Hi,

I've seen multiple threads about this topic and have tried most (if not all) documented fixes for this problem.

I build a quadcopter using these parts:

Pixhawk Advanced Autopilot
Ublox Neo-7M GPS with Compass (not connected)

DT750 Hextronic Outrunner 750kv
RotorBits QuadCopter Kit With Modular Assembly System

Props 2pc CW 2 pc CCW Rotation 12" 4.5 pitch
Compact 4000mAh 3S 25C Lipo Pack
Afro ESC 20Amp Multi-rotor Motor Speed Controller

PPM encoder for PWM signal from the 5ch RC controller

From the threads i've gone through, i have:

  • Made sure that the proper motors are attached to pixhawk in the proper manner (shown in the picture).
  • Made sure that the propellers were spinning in the right direction.
  • Made sure that the clockwise and counter clockwise props are on the right motors (seen in the picture).
  • calibrated the accelerometer in mission planner. Values given when on flat ground is

Pitch: (±0.01 - 0.2). Roll: (±0.01-0.4).

  • Ran motor test to see if vibrations/battery affect accelerometer. They do not.
  • tested over grass and flat concrete.
  • Started with Arducopter 3.01, upgraded to 3.2 to see if the problem might be fixed.
  • Radio gives proper outputs and was calibrated.
  • ESC's are calibrated.
  • Made sure it was in + configuration.
  • Ran motor tests.

A couple of comments about the quad:

  • The pixhawk is not centered on the base, its a little off (towards the back towards motor #4).
  • The weight is pretty well distributed but not perfect.
  • starts flipping at about 50% throttle.
  • always flips towards the corner (between 2 arms, could this mean that there is extra pitch and roll?)
  • using 5 channel controller and a ppm encoder for rc. All channels work (1-5) properly.
  • I have logs, no idea how they work but i just downloaded them from the pixhawk using mavlink.

I've tried fixing this problem for 2 weeks but i cannot find anything wrong with the quad. I've looked at multiple forums and all the solutions were not applicable to my quad (usually configuration problem with props/motors). I'd love some help or any suggestions. 

Thank you.

Views: 6755

Reply to This

Replies to This Discussion

Liban,

From my own experience, two suggestions:

1)  I believe strong, consistent flipping during take-off is most likely to be due to mis-wiring the Power Distribution Board (PDB). Your controller is sending 4 signals to the PDB, and the PDB is distributing those signals to the ESCs.

If you get the match-up wrong, the controller re-acts to changes in orientation (wobbling during take-off - inevitble) by sending a corrective signal to the wrong motor/ESC, which amplifies the instability instead of off-setting it, which causes the controller to send a stronger signal, which makes the problem worse...resulting in a sudden, violent flip-over every time you start getting airborne. The controller thinks it's correcting the problem but just making it worse. 'Positive feedback loop'.

(The story of my first real-world flight, by the way. I think I'm not alone).

Check those PDB connections very carefully. Check them again.

You say 'proper motors connected to controller in proper manner- see picture', but it's not possible to see if you've soldered the Controller-to-PDB wires correctly.

2) Somewhere around the 3.0 firmware update (3.01?), Randy advised not to touch the pitch or roll controls on the transmitter during take-off, or you 'may experience flipping on take-off'. Just centre them, use only the throttle control until you're well airborne (10 feet?), and trust the controller (only if it's properly wired!) to keep the craft stable until then.

This also works (also from hard experience).

Caveat - I have no Pixhawk experience, but I don't think it changes the 'PDB mis-wiring' danger.

George

Add-on to my previous post:

If the PDB is wired correctly but facing 180 degrees in the wrong direction, you will get the same flip-over problem. (If 90 degrees off, severe instability and crash on take-off).

George

For the quad, we are using a simple breakout cable. I don't think this could cause the problem.

As for controls, the flips occur with only throttle. Even after assuring that the sticks were centered with calibrating the rc inputs and calculating their median value and making sure not to change them.

Ok (no PDB);

Though it wasn't actually the battery-to-ESC (power) connection I was worried about (your second picture).

It's the Controller-to-ESC signal connections I'd still suspect first.

I can't see those connections in your picture. How are you connecting the ESCs to the flight controller?

To be sure everything's wired correctly, some people suggest holding the quad (very carefully) above your head and, with mid-power to motors, apply pitch and roll on the transmitter and see if the quad wants to tilt in the expected directions. If not (if opposite to expected directions), the output connections from the controller are going to the wrong ESCs/motors (or your transmitter controls are 'reversed', but unlikely since you're not touching pitch/roll during the flips).

More safely (my choice), with quad on the ground as usual, don't give enough throttle to lift off - just 25% or so - and in short, quick bursts, apply extra pitch in one direction then the other, then the same for extra roll. If the quad 'tilts' in all four directions the way it should, it's not a wiring problem.

My advice is don't assume it can't be a wiring problem - check.

George

 

I checked the Motor configuration but I did not do a hand test. I will check the motor configuration again and  perform the test on monday (Quad is not with me). Thank you.

Liban_O,

After you check all of the above a couple of more thoughts

1. During the radio calibrations the pitch control when moved away from you or forward must cause the bar in Mission Planner calibration display to go down not UP.

2. It often takes a bold jump to get the quad off the ground say 5 feet up so that the FC can start to control the motors. If you are too cautious it starts to tilt 6" off the ground and the fc can't do anything about it.

3. Balance is very important. Make sure the aircraft is not heavy on one side or another. Tie four strings to the arms at the plate and lift it up and see if it balances.

It is frustrating but hang in there and you will figure it out.

Need the logs for this problem.  It will tell us how the flight controller is responding to the takeoff.  It will tell us which motor is being powered and what the flight controller thinks is happening.

 

I don't know if these are the right files.

Attachments:

Both these files show motor 3 is running hot and that your RCIN are all at zero.  They should be at 1000-2000.

For example at mid stick or Pitch and Roll at takeoff should be at 1500 for center.

Don't know how your getting the unit to takeoff.

 

Mike

 

I tried looking at the logs and the RCIN's are all 0 (don't know why). RCOUT has values that look like what my quad is receiving.

The min for my rc controller is 975 and the max is 1980. I made sure to make sure all sticks were at their mid (except throttle which was at 975). I don't know why the RCIN values are all 0 as the quad was obviously responding to throttle.

For motor 3 running hot, I did notice that one of the motors, when spun with the hand, had a little more resistance in its spin. When i start up all the engines, it looks like it spins the same speed as all the others.

Here is a more recent log of this morning. RCIN values are looking normal now, i do not know why they weren't there before.

Attachments:

Here is a video demonstration of the motors.

  • Armed and spinning at about 10% throttle.

https://www.youtube.com/watch?v=s05ziNOQxPI

  • Mission planner motor test order A-B-C-D

https://www.youtube.com/watch?v=XmSB1P1V2WI

Reply to Discussion

RSS

© 2017   Created by Chris Anderson.   Powered by

Badges  |  Report an Issue  |  Terms of Service