Hello,
I have a Skywalker X-8 with a Pixhawk running ArduPlane V3.2.3, and I'm trying to get the automatic takeoff working smoothly.
Currently, the aircraft is setup with a Turnigy SK3 3548 700kV motor, a Master Airscrew S-2 series 11x8 prop, and a 5s 5000mAh battery. It is launched using a 5mm bungee.
The problem is that once the aircraft releases from the bungee, it pitches nose up to between 50 and 80 degrees. It does not stall, but it shoots through the desired altitude. The pitch limit and TECS max pitch are both set to 30 degrees. The max takeoff throttle is set to 100, and throttle slew 100.
In reviewing the logs, the desired pitch is 30 during launch, so the autopilot is considering that parameter before disregarding it. The airspeed peaks about midway through the launch, then falls off very close to stall. The lowest point corresponds to the max altitude. Also, the pitch I-gain is very high at the beginning of the launch before decreasing.
To me, this looks like a combination of TECS using pitch to maintain airspeed and the bungee holding the nose down which is causing the the I-gain to ramp up.
Why does ArduPlane not modulate the throttle to maintain airspeed, but instead uses pitch?
Short of playing around with desired pitch angles, altitudes, and max throttle settings, it there a way to mellow out this tendency?
I've attached log, GPX, and .mat files if that helps.
Thank you very much for taking a look at this, and any insights you might have are appreciated.
Joe
Replies
FWIW, we currently use a takeoff pitch angle of only 12° with just 1 fail in over 400 reliable takeoffs, speed being controlled by takeoff throttle but with the new firmware looking to fix that.
I do not doubt you managed to get it to work.
I have two different planes this was tried out on, (I always set pitch to between 8 and 11 degrees)
I can't remember when it worked last, it's a long time ago. And I see no reason for it to fail (please prove me wrong by looking into this.)
Will's logs contain the same problem when it comes to controlling pitch.
Grant and Andre,
Haha! I promise that I read that page before loading up a flight plan, and had everything setup as best I understood from that page you linked to, as well as my other experiences with APM/Pixhawk and Mission Planner.
I'm not sure why it isn't showing the switch to auto, but I definitely had a takeoff command and the mode set to auto.
Here's a new log file. It has three launches, and they all show auto before launch, and have the same pitch behavior.
When I write a flight plan, I select all of the waypoints, DO- commands, etc..., then add a TAKEOFF command at the top. I set the desired pitch and the altitude, then write the flight plan to the autopilot. Is this not the correct way to implement the automatic takeoff?
Thank you for your help,
Will
2015-05-14 12-25-43.zip
Hi Will. The Lat/Long of your takeoff waypoint is 0,0 so during takeoff the plane tries to fly to the west coast of Africa. Fix that, try again and the let us know how you go.
Thanks, Grant
Grant,
I didn't think that Takeoff needed a lat/long. I thought that it locked the heading until the altitude was reached, then turned on course, but I will definitely give it a try.
The weather doesn't look good here for a test until later in the week, so I'll report once it's flown.
Thanks for your help,
Will
@Will,
you are right, the lat/lon doesn't matter in the takeoff command. It just holds the heading from when it started the takeoff.
For the upcoming release (probably today or tomorrow) APM:Plane does now reduce throttle if we go over speed on takeoff. Previously it always used max throttle on takeoff, which could cause TECS to raise the pitch a lot.
Cheers, Tridge
Tridge,
Thanks for confirming that. I had suspected it, and SITL doesn't need it, but I'd never seen it set up, so I wasn't sure.
Hurray for the devs! New ArduPlane firmware makes my week!
Thanks,
Will
That's better, I won't go into all details, but from what I see in 15 seconds is.:
near line 194k you start, mode is auto, command is 22 (MAV_CMD_NAV_TAKEOFF) , min Pitch is 30.
CTUN.NavPitch=30 at line ~195715 target pitch is exceeded, but at line 196687 the RCOU.Ch2(elevator servo) is still increasing wildly even past 71degree - and past a point where CTUN.NavPitch starts to sink.., then you take over in manual.
Quick conclusion; (which is in my bug report) - Takeoff fail to control pitch properly.
Unless Grant sees something else, I would urge you to add a comment, your log, to my bugreport.
Relevant information, we both are using the airspeed , I am using EKF, you not. your goes wildly up, mine fly like a trimmed plane, but with no meaningful pitch control.
Andre,
Thank you! I'll wait to see what Grant says before making a comment on GitHub, but your insights are very valuable.
Also interesting is the pitch I-gain. It starts increasing when the mode is switched to Manual before launch, and doesn't stop until the aircraft starts moving on takeoff. This looks curiously like a bug in the controller, since I would expect the I-gain to sit at zero while the mode is Manual.
Thanks,
Will
ooops, sorry . I assumed you did it before and knew the procedure, now that Grant wrote what he did, I see the log shows you are in manual mode, with no takoff command.
-
1
-
2
of 2 Next