Custom Quadcopter PID Stabilization Help..

I am building a custom quadcopter based off an Arduino flight controller. I attempted to fly it today. It went about 8 feet in the air, hovered steadily for almost a second, then began to oscillate slowly, then rapidly, then finally flipped over in the air and crashed. I attached this video (https://youtu.be/RdITL_6V534) to give a fairly accurate description of what happened. It broke one propeller.

I started my P gain for Pitch/Roll PID at 0.8. I and D gains were at 0 for Pitch/Roll Loops. P, I, D gains was set at 0 for Yaw loop. I verified the motors responded in the right directions to my PID loops with the gyro and controller inputs.

Is it my fault I didn't immediately move the quadcopter around? Should I not let it attempt to hover untuned and immediately move it around and start tuning? Would it be better in my case to tie the quadcopter down and try to tune it that way because of what happened? Did I set the P gain for Pitch/Roll too high to begin with? Any advice would be greatly appreciated.

I've been getting suggestions that this is because the D gain is set to 0 for the Pitch/Roll loops...

You need to be a member of diydrones to add comments!

Join diydrones

Email me when people reply –

Replies

  • So I decided to give it two more tries before I built a stationary stand and things didn't go so well. I recorded the crash (https://youtu.be/T3bjvrg7_-I) and you can see what happens. In summary, uncontrollable oscillations when I tried to take off. It crashed a few days ago and other forums told me that was because my D gain for Pitch/Roll was set to 0. I started it at 5 and this issue persisted. I'm not sure what to do. Is it my gain values again? My PID loop?

    My PID Gain Values:

    Pitch/Roll

    P: 0.5

    I: 0

    D: 5

    Yaw

    P: 0.5

    I: 0

    D: 0

  • Okay, so I start with a small value for D and work up from there. 

    What is your opinion on this tuning technique? (start at 12:00)

    https://www.youtube.com/watch?v=JBvnB0279-Q

    Chris Anderson said:

    If your D gain is set to 0, you don't really have a PID -- you just have a P, which is unstable in an undamped system like a quadcopter. You need to add damping, which is what the D term is for. Set it for something above zero and experiment. 

  • 3D Robotics

    If your D gain is set to 0, you don't really have a PID -- you just have a P, which is unstable in an undamped system like a quadcopter. You need to add damping, which is what the D term is for. Set it for something above zero and experiment. 

This reply was deleted.

Activity