I have a quadcopter with TBS discovery frame and 10inch props running on 810Kv motors on 3S. The quadcopter has an asymmetrical mass distribution since the heavy battery pack at the back causes the weight to be heavily skewed towards the back. The quad is really unstable and a pain to even hover properly, even in angle/horizon.

I was going through another thread on this stackexchange, and was trying with low PIDs to no avail. Turns out I need high PID gain for a craft with low power to weight ratio (around 8 : 3). My tune still seems to have some issues, since my quad was still having oscillations and it suddenly lost altitude and slammed hard into the ground when faced with a sudden gust of wind while hovering at 20m. Doing fast moves like rolls and flips with this quad doesn't really appear to be a good idea before setting up the PIDs properly since it is a bit to large and risky. Most of the tutorials I have followed seem to be stuck on gauging response to sharp turns and pitches and rolls, so I am not sure how to go about it. Increasing I or decreasing P and D makes the quad even not wanting to lift up from the ground, oscillating on the ground like a pendulum along the midline.

This drone is just a hobby project with no special use case. I will be using it just for leisure flying.


1 Answer 1


The discovery frame uses the old DJI plastic arms and the arms are not very stiff. So you get a lot of flex in the arms. This leads to two major problems: vibrations and the quad has a relative long delay until force of one motor really starts to rotate the quad. This problem gets amplified if you are using out of balance motors or very felxible propellers. This would explain the unstable flight you are describing.

I'm afraid that this is a "hardware problem" and not a "software problem" that you can fix with some PID settings. Modern flight controller firmware is built with certain assumptions about the used frames/escs/motors. A 10 inch prop on a 810KV motor with 3S and a plastic frame isn't a target platform for betaflight or any other race/freestyle oriented FC firmware.

What are your options? Make the frame as rigid as possible and try to dampen the FC so you get rid of the vibrations (e.g. put double sided tape under it). I don't know what FC you are using but try to get some blackbox logs to see the frequency range of the vibrations and try to add more filtering to clean up the signal for the FC. After that you can start to tune the PIDs and maybe get a flyable quad. You can tune the quad without any flips and rolls because with a low power/weight ratio the quad will probably crash in the process. So just tune the quad for regular forward flight.

  • $\begingroup$ I want to make a few clarifications here: a> The flex on the quad is not a lot, I have flown this quad with Ardupilot on a Pixhawk before, but decided to switch over to Betaflight due to troubles with some hardware and some really bad crashes b> The motors and props are well balanced. c> I have managed to tune the quad so that it is more than just "flyable", but it is still far from what I'd like. d> I am using a CLRacing F7 dual gyro v2.1 e> I am clueless about filtering, but still reading. Stuck a bit since the flash storage is just insufficient for a full flight at 2K logging rate. $\endgroup$ Commented Mar 5, 2021 at 17:42

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.