Re: [Paparazzi-devel] Is the multirotor software of paparazzi not save?

[Christophe De Wagter]

1. Many motor controllers can spin up much faster than they spin down. If your gains are very high and oscillations occur, than because of this the average real throttle becomes higher than the average commanded throtlle. This is not specific to paparazzi.

Reduce your gains a little, buy better motor controllers or add a software filter that slows down the spin up just as much as the slow down.

2. This would confirm that your d gains are very high.

On Jun 22, 2014 5:33 PM, "Prof. Dr.-Ing. Heinrich Warmers" <address@hidden> wrote:

Hi,

i have to report dangerous  behaviour of the paparazzi multirotor software:

1. by fitting the values of the basic attitude control system i observed in 2013 the following: The multirotor shows vibration of the frame and climb also if the throttle value where set to zero  by the rc stick! The only way to command to ground was to arm all motors by the kill switch. I think there is a vibration generated motor control value.  This is a really safety problem or a wrong control idea. For RC flights the RC-Sticks have to control the motors overall.
2. By hart landing the motors get full throttle values and the multirotor jumps same times. 
3. The GPS fix is same times wrong. It is possible to get 0 0 coordinates (middle in the Atlantic see) wiht a gps fix.  This is also shown for fixed wing. I see this by the use of lea4 from üblox.  It seems that same messages are wrong.
4. The lisa/M v2.0   in combination with the aspirin2.2 does not run out of the box in combination with the  current stabile version (5.04).

You have to add in  <subsystem name="imu"           type="aspirin_v2.2"/>

 <configure name="LISA_M_BARO" value="BARO_MS5611_SPI"/>

If you forgot the second line you mutirotor will go up and away and you will don’t see to multirotor again if you switch the  automatic mode.  The reason is that there is no baro altitude information and the multirotor will always climb.

You can monitor baron altitude by changing the altitude by hand (with armed motors) and observing the altitude information of the control station.

 

The point 3 an 4 can easily solved by some lines of code. I hope that someone will do this soon. Second I want to set up a safety page for the startup of paparazzi multirotors.

Regards

Heinrich

 

 

 

_______________________________________________
Paparazzi-devel mailing list
address@hidden
https://lists.nongnu.org/mailman/listinfo/paparazzi-devel