Re: [Paparazzi-devel] Throttle problem with AIRSPEED sensor (Gautier)
From:
Christophe De Wagter
Subject:
Re: [Paparazzi-devel] Throttle problem with AIRSPEED sensor (Gautier)
Date:
Thu, 14 Jun 2012 21:58:53 +0200
Gains should always precede the integrators. Then the value in the integrator is the trim value. For airspeed the trim value would be the cruise throttle. There are several integrators that can integrate beyond the limits of the control, especially with higher gains.
-Christophe
On Thu, Jun 14, 2012 at 8:13 AM, Eduardo lavratti <address@hidden> wrote:
Gautier ... i am very curious to know how you discovered the problem. how you debuged ?
are you used log file ploter ? if yes, what parameter you look ?
about your replay, what are the other control loop available ?
flix sugest the NEW_Control is this that you like to tell me ?
sorry for my poor english regards.
Date: Sun, 10 Jun 2012 20:52:42 +0200 From: address@hidden
To: address@hidden Subject: Re: [Paparazzi-devel] Throttle problem with AIRSPEED sensor (Gautier)
Hi Edouardo,
Here is what happen during your flight:
According to your airframe file, you are flying with the standard
control loop but with the airspeed (actually ground speed) control.
Is it what you wanted ? In this mode the pitch is controlled by the
vertical speed error (and thus altitude error), while the throttle
is controlled only by the air and ground speed error.
When you first engaged AUTO2, the ground/air speed error is close to
zero (or even slightly negative). The weakness of the PI controller
used in this loop is that you can't set a "cruise throttle" value as
initial throttle setpoint. In this situation, the output of the
control is 0 throttle, resulting of speed loss which would have
increase the integral action if the time spend in AUTO2 had been
long enough.
As it was not the case, the second switch to AUTO2 leads to a 0
throttle, but this time increasing. This would have been enough to
continue the flight if your plane had not stall due to a combination
of pitch up setpoint and high bank angle.
This stall produced an important drop of your air/ground speed,
which in the end saves your day as it allows the integrator to fully
"charge" itself. For your last attempt to enter AUTO2, the throttle
setpoint was close to 100%, resulting of a correct vertical control
of your plane.
To summarize: integrators are real shit, but no, you can't just get
rid of them like that. The correction I can advise to you is to add
a cruise throttle setpoint corresponding to your throttle for cruise
flight (somewhere close to
sw/airborne/firmwares/fixedwing/guidance/guidance_v.c:351).
Last point: in your radio file, I think it is better to have min and
neutral equals for THROTTLE channel, so you don't have -100% display
in the GCS.
Good luck for your flights (in a very nice place by the way ;) )
Gautier
Le 10/06/2012 01:17, Eduardo lavratti a écrit :
Gautier, this is the logs that show a problem when i engage
AUTO2.
The flight start after 330 seconds
i tried to engage auto2 for 3 times.
on the 3th time the throttle work fine.
this is my last configuration.
i hae another logs with same problema but the airframe file is
different.