Re: [Paparazzi-devel] Plane drops after take off on energyadaptative loop

[Felix Ruess]

FYI, there was a bug/regression in energy_ctrl that has been fixed
now: https://github.com/paparazzi/paparazzi/pull/1597

On Fri, Mar 11, 2016 at 5:16 AM, Joseph Muhlhausen <address@hidden> wrote:
> Just to report, I re-trimmed the plane and re-tuned the pitch loop, I also
> increase the initial nominal cruise pitch to 1 degree and played with the
> energy total and energy diff gains, it seems to have solved the issue.
>
> Thanks again!
>
>
>
> On Wed, Feb 24, 2016 at 10:51 AM, Joseph Muhlhausen <address@hidden>
> wrote:
>>
>> Thanks Guys!
>>
>> It looks like I am flying too slowly, I honestly thought the AoA was
>> around 5 degree and didn't think of checking the log. I will test the plane
>> again this weekend with your recommendations and report.
>>
>> Joseph
>>
>> On Wed, Feb 24, 2016 at 2:29 AM, Balazs GATI <address@hidden> wrote:
>>>
>>> Dear Joseph!
>>>
>>> Flying close to the stall speed is not the safest way of operation,
>>> because the aerodynamic behavior of the airplane changes very non-linear.
>>> Thus the controller won't be able to keep the demanded altitude, speed, AoA,
>>> as you would expect.
>>>
>>> Of course UAV is the best device to research stall but I suppose you
>>> don't like it during take-off. :)
>>>
>>> Balazs
>>>
>>>
>>>
>>>
>>> 2016.02.23. 21:57 keltezéssel, Christophe De Wagter írta:
>>>>
>>>> If the high pitch angle is fine for you, just type the desired-pitch
>>>> value you see during normal cruise as nominal cruise pitch in your
>>>> airframe file. Otherwise your controller will start with this value (in
>>>> your case about 20 deg) too little pitch and it takes a few seconds for
>>>> the controller to correct such a big angle, especially when you activate
>>>> it in (slight) underspeed, in which the initial reaction is to slightly
>>>> nose down a few degrees extra before it finally notices the dramatic
>>>> dive and starts to correct for this using the diff i  gain.
>>>>
>>>> Nominal-cruise-pitch is  used as initial value for energy control. Put a
>>>> reasonable value in there ( in your case RadOfDeg(22) linking at your
>>>> graphs )
>>>>
>>>> On Feb 24, 2016 02:17, "Joseph Muhlhausen" <address@hidden
>>>> <mailto:address@hidden>> wrote:
>>>>
>>>>     Tobias, thanks! my plane is tuned pitch down to make landing easier,
>>>>     I will modify that. The standby waypoint in 50 m above the deroute.
>>>>
>>>>     Christophe, thanks! Yes my nominal pitch is high because I fly
>>>>     really close to stall speed and thus need a high angle of attack.
>>>>     Could I instead increase the nominal throttle and fly with a pitch
>>>>     near 0?
>>>>
>>>>     Best,
>>>>     Joseph
>>>>
>>>>     On Tue, Feb 23, 2016 at 12:57 PM, Christophe De Wagter
>>>>     <address@hidden <mailto:address@hidden>> wrote:
>>>>
>>>>         It takes a few moments for the integrator to find out you need a
>>>>         20 degree pitch up to stay level, which explains the sudden dive
>>>>         followed by normal flight. As default value for the pitch
>>>>         integrator the cruise pitch from the airframe file is used.
>>>>
>>>>         On Feb 24, 2016 01:53, "Christophe De Wagter"
>>>>         <address@hidden <mailto:address@hidden>> wrote:
>>>>
>>>>             Fill the nominal cruise pitch (about 20 to 25 degrees) into
>>>>             the nominal cruise pitch: RadOfDeg(25)
>>>>
>>>>             PS: your cruise pitch looks quite large. Might want to look
>>>>             at the tuning of your inner loop, the placement of your IMU
>>>>             etc...
>>>>
>>>>             On Feb 23, 2016 20:56, "Tobias" <address@hidden
>>>>             <mailto:address@hidden>> wrote:
>>>>
>>>>                 Hi Joseph,
>>>>
>>>>                 energy-loop is the way to go, just to answer your last
>>>>                 question. :-)
>>>>
>>>>                 regarding the unwonted behavior - first thing to check
>>>>                 is: is your airframe well tuned? I mean is it holding
>>>>                 altitude and speed well. after a wile?
>>>>                 If so you should check probably the values for ( )
>>>>                 against the flight log for straight flight and climb.
>>>>                 <!-- auto throttle inner loop -->
>>>>                         <define
>>>> name="AUTO_THROTTLE_NOMINAL_CRUISE_THROTTLE"
>>>>                 value="0.3"/>
>>>>                         <define
>>>> name="AUTO_THROTTLE_NOMINAL_CRUISE_PITCH"
>>>>                 value="0."/> <!-- default 0 -->
>>>>                         <define
>>>> name="AUTO_THROTTLE_CLIMB_THROTTLE_INCREMENT"
>>>>                 value="0.1" unit="%/(m/s)"/>
>>>>
>>>>
>>>>
>>>>                 in the scenario you described you are changing from
>>>>                 fixed values to the energy loop - its always a good idea
>>>>                 to have the plane in a rather similar attitude/speed to
>>>>                 do this. So what is your alt for stby? just thing about
>>>>                 it - if it is the same alt as your deroute from the
>>>>                 previous block try to increase the stby alt or decrease
>>>>                 the deroute alt
>>>>                        <exception cond="GetPosAlt() > ground_alt+50"
>>>>                 deroute="Standby"/>
>>>>
>>>>                 this allows the plane to catch up.
>>>>
>>>>                 cheers Tobi
>>>>
>>>>                 _______________________________________________
>>>>                 Paparazzi-devel mailing list
>>>>                 address@hidden
>>>>                 <mailto:address@hidden>
>>>>
>>>> https://lists.nongnu.org/mailman/listinfo/paparazzi-devel
>>>>
>>>>
>>>>         _______________________________________________
>>>>         Paparazzi-devel mailing list
>>>>         address@hidden <mailto:address@hidden>
>>>>         https://lists.nongnu.org/mailman/listinfo/paparazzi-devel
>>>>
>>>>
>>>>
>>>>     _______________________________________________
>>>>     Paparazzi-devel mailing list
>>>>     address@hidden <mailto:address@hidden>
>>>>     https://lists.nongnu.org/mailman/listinfo/paparazzi-devel
>>>>
>>>>
>>>>
>>>> _______________________________________________
>>>> Paparazzi-devel mailing list
>>>> address@hidden
>>>> https://lists.nongnu.org/mailman/listinfo/paparazzi-devel
>>>>
>>>
>>> --
>>> Balazs GATI, PhD
>>> associate professor
>>>      Department of Aeronautics, Naval Architecture and Railway Vehicles
>>>      Budapest University of Technology and Economics
>>>
>>> Address:   Budapest
>>>           Stoczek u 6. J. ép. 423
>>>           1111
>>> Tel:       +(36)-1-463-1960
>>> Fax:       +(36)-1-463-3080
>>> Homepage: http://vrht.bme.hu/
>>>
>>>
>>> _______________________________________________
>>> Paparazzi-devel mailing list
>>> address@hidden
>>> https://lists.nongnu.org/mailman/listinfo/paparazzi-devel
>>
>>
>
>
> _______________________________________________
> Paparazzi-devel mailing list
> address@hidden
> https://lists.nongnu.org/mailman/listinfo/paparazzi-devel
>