Well, this time it was pretty clear, with the RC system, I increased Kd until it starts to oscillate and then lowered it a bit.
My next modification will be to make the average of the reading at least during the 10ms between calculations of the PID, I think it should improve the differential term. The gyro itself includes a low pass filter also.
To make the tuning you mention I need to finish my engineering degree before
at least learning some matlab I suppose. At the time I could take the data without much difficult, but I would not know what to do after that.
However it already works well enough