Hello,

I ran across this thread over at google code drones discussion and would like to bring attention to it, hoping 3DR chimes in and can clarify if indeed the rear blades need to be swapped (and documentation updated to reflect this). You will notice on your Y6 that if you yaw, it induces a pitch change.  Here is a portion of the discussion where a solution is found: 

Problem definitely solved : it was caused by the rear arm propellers arrangement (reversed compared to the front arms).


Here is how i did solve it :

- put a pusher prop on the rear arm upper position (instead of the normal one)

- put a normal prop on the rear arm lower position

- exchange ESC outputs for those two motors and reverse spinning directions


So now i have an Y6 with the 3 pusher props in upper location on the three arms.


No more Yaw / Pitch coupling at all.


This is confirming what i thought, a thrust asymmetry is induced if a propeller tandem is reversed on a arm.


The probable reason :


On a coaxial setup the two propellers of a same arm do not have the same efficiency. During yawing, we are accelerating three propellers, and slowing down the three other ones.

The problem is that the vertical thrust result is not the same if one arm has a different prop arrangement : rising the speed of the lower prop is more efficient than rising the speed of the upper prop.

So if two arms of the Y6 frame have pusher props in the up location, and the other arm have the pusher prop in the lower location, then when yawing one side, we produce a different vertical thrust on the rear arm compared to the front arms, because we are not speeding up the same arm propeller. This is causing a pitch derivative coupling.

The coupling is only derivative because the induced pitch unleveling is compensated by the APM accelerometers as soon as it is detected but with a small delay.



So the definitive solution is to put the three pusher props in the same location on the arms. That's it.


I think that reintroducing a top bottom ratio is not the solution, at least not without a more complex motor mixing to avoid axis coupling.

 

Olivier.

 

If a simple change can increase stability, I think it's really worth looking at and addressing.