Re: [Paparazzi-devel] Quadrotor with Lia 1.1 - a sudden loss control aut

[Loic Drumettaz]

Hello,

From the plots I think what could have happened: 

-In Hover mode the quadrotor is climbing up to almost 30 meter (which is wrong!)

- Then the pilot switches to RC_DIRECT mode and the quadrotor starts to descend

- But the vertical speed becomes too high: almost 10 m/s according to the plots

- It is then very hard to decelerate in such conditions, even with throttle at 100%...

Again, it is sometime difficult for the pilot to switch from hover to direct mode, specially when hover mode is not tuned.

So, it seems that the original problem comes from what happened in hover mode: the quad was climbing. 

To solve this problem, i would recommend:

1- Reduce vibrations with mounting the lisa on damping foam. This may help the adaptatve control to find the right nominal throttle. 

2- As an alternative you could also use the NOMINAL_HOVER_THROTTLE define. See: http://paparazzi.enac.fr/wiki/Control_Loops

3- Check baro sensivity with altitude and vertical speed measurements on ground (well, on stairs)

4- BMP085 is very sensitive to sunlight, it must be covered.

Regards