Auto Tuning

am working with a Tarot 650 quadcopter using Pixhawk 6x, 16×5.5 inch propellers, 350 KV motors, 6S 16000 mAh LiPo, and a front-mounted payload (~800 g). The payload is required for the application (gripper / robotic arm), and the center of gravity has been mechanically adjusted to be near neutral.

Observed Issues

Aggressive Pitch Response with Front Payload

When commanding forward motion, the drone initially pitches forward as expected.

Due to the front payload inertia, the nose drops faster than desired.

The controller then overcorrects, causing the nose to pitch back up aggressively.

This results in a nose-down → nose up oscillatory behavior, especially noticeable during small pitch commands.

Movements feel over aggressive rather than smooth, despite CG being reasonably balanced.

Yaw Stress and Diagonal Motor Heating

During hover (even without forward motion), cross-diagonal motors heat up more than the others.

CG is verified to be acceptable, so heating does not appear to be caused by front rear imbalance.

Yaw corrections appear to be working continuously, suggesting yaw loop over-activity or torque imbalance.

Current yaw gains were initially high (ATC_RAT_YAW_P and ATC_RAT_YAW_I both ~0.22).

AutoTune Behavior

AutoTune was attempted after Stabilize → AltHold at ~18 m altitude and also on a PID rig.

On free flight, AutoTune aborted mid air and the vehicle descended abruptly (no damage).

Log analysis shows autotune_failed, with no battery voltage sag.

On the PID rig, AutoTune consistently fails after ~1 minute with “AutoTune failed, return to manual mode.”

This suggests AutoTune is unable to converge for this heavy, high-inertia configuration.

Low-Altitude and Initial Motion Behavior

At takeoff and low altitude, the drone tends to move in random directions.

In Stabilize mode, it does not hold position (expected), but initial corrections feel exaggerated.

Transmitter trims are centered; accelerometer calibration has been performed.

How to tackle this

Can you try lowering MC_PITCH_P ?


Consider that this could also be caused by the mounting of the motors.


Autotune can fail, if your initial guess is way off. Do you have a simulation for that vehicle where you could try to tune it?
Also, consider setting MC_AT_APPLY to 0 and then analyze the data later on the ground from the log.


A log would be helpful to debug this. Could be caused by the bad tuning.

First time it wasn’t working later i had to fine tune it to come to those approx parameters.

Now the thing is the diagonal motors were heating i haven’t noticed them earlier but without the payload also the same thing even the centre of gravity was good. I already checked the motor mounts were not tilted. Also esc are good and working l, i tested them seperately