Hi kwaiweng

I'm not 100% sure about #1 problem, but #2 problem is simple:
You must take a reference 'mid' reading of all gyros when motors are stopped - just before takeoff.
That way you have the most up to date 'mid' value readings.
I also take a reading of the pitch, roll and yaw transmitter stick readings at same time, because these are also part of the 'mid' equation.
In my software, if throttle is very close to zero, then I continually read and update all 'mid' values untill throttle is increased again.
Danvel's explanation of his algorithm mentions this too.

Because of temperature drift etc, you will end up with a different 'mid' value for each gyro/stick channel. The idea is to have 3 (pitch, roll & yaw) completely separate algorithms running.

PS: Do you have a website where you can post a video clip of your DF flying ?


Cheers ADI