Just been out to maiden my new fx-61 Phantom wing build. First plane I have ever used Ardupilot on (I am experienced with Copter but new to Plane). I have 3.5.2 running on a Pixhawk Lite FC, M8N GPS unit, no airspeed sensor fitted. I hand launched in Manual mode, it flew really well - nicely balanced and responsive. I later switched to Stabilize, and the plane instantly changed attitude to climb vertically, then stalled, then on me trying to recover, it climbs vertically again and on it goes in this cycle. I quickly switched back to manual and recovered from this the first time. A few minutes later I decided to gain some altitude in manual and this time try FBWA - I got the same results as before :-( I tried to recover, thinking I had switched back to manual flight mode (was actually in Stabilize - not used to the modes mixes on my radio for this model yet!), but of course that cycle of climb/stall started again and I ended up after several loops of that cycle stalling into a nose dive into the ground! Fortunately a spongey muddy field, so no real damage done.
If anyone is willing to check my flight log to see if they can suggest any possible cause, I would be most grateful.
Dataflash log is here (too big to attach): https://dl.dropboxusercontent.com/u/12131631/forum-links%28do%20not%20delete%29/1.BIN
Replies
Thank you very much for your reply dear Paul,
I cannot control/move by sticks the surfaces of the plane (except one surface only)
Also, when i tilt or roll the plane by hand it didnt react to inverse the direction except that one surface. all of this is in the FBWA mode... furthermore in FBWB it is worse nothing at all will try to react or move using an input sticks when applied.
I really lost in this i tried many many things but no luck, i dont know what i missing whether missing information or setting.
I am noob in ArduPlane, but the ArduCopter i did all of that easily.
Please if you have any hint, tip, idea or suggestion please tell me. I am trying to solve this before many days...
If you can sir see my thread it shows all my setup:
http://diydrones.com/forum/topics/newbie-in-fbwa-mode-surfaces-not-...
Thanks, i appreciate your respond very much, it gives me hope :D :D
Nobel,
Sorry, I'm not 100% sure I understand the meaning of your question.
In FBWA mode the surfaces still move with the RC inputs, but they are also assisted - so if you hold the plane level then the control surfaces should be level - if you tilt the nose down then the elevator will tilt up to try and correct back to level. The same applies to the roll with ailerons.
In RTL mode the plane will return to a position above the launch position and just circle, but you can override with Rc controls although it reduces the input somewhat.
Dear please how i can move the surfaces while i am in FBWA or any auto mode ? << of course i am in a 3D GPS lock.
I have 5 servos to control surfaces, 2 elev. and 2 Ailerons. and 1 rud. and it is functions as it should be in Manual mode.
???!!!
First proper flight today since fixing the issue and it flies like its on rails. Very happy with it!
I think I have found the solution! Basic rookie mistake in elevon setup! The elevon roll response in FBWA was working correctly, but the elevon pitch response was reversed, so when the plane pitched up, it was being sent elevator surface up commands instead of down as it should be to get it back level - this would totally explain why one second it was climbing vertical and the other diving straight down! I had to set the CH2_REV = -1 to resolve this. So glad I did a dig into those parameters now! Thanks James for the response, gave me the nudge to go figure it out ;-)
That's exactly what I did James. I got parameter settings from the parameters file attached to the first reply in this thread, which many fx61 flyers seem to use successfully: http://diydrones.com/group/fx-61-phantom/forum/topics/apm-phantom-w... (Pieter) and used those. One difference in my build is that I have no airspeed sensor yet, so that is disabled. Otherwise (aside from calibration values) the gains I am using are as per the ones advised in that params file, with the exceptions of:
MIXING_GAIN - I have as 0.7, the param file has 0.9
NAVL1_PERIOD - I have 20, the params file has 24.
ARSPD_ENABLE - I have 0, the params file has 1.
ARSPD_OFFSET - I have 2.41221881, the params file has 2139.531 (this is a huge difference)
ARSPD_PIN - I have 15, the params file has 0. (not sure which airspeed sensor they would be using with this set to zero - for Pixhawk it says that for a digital 12c sensor, this should be set to 65 and for an analog ADC 6.6v connection, this should be 15)
ARSPD_RATIO - I have 1.99, the params file has 2.2
ARSPD_USE - I have 0, the params file has 1.
My pitch gains are these:
PTCH2SRV_D,0.02000000
PTCH2SRV_FF,0.00000000
PTCH2SRV_I,0.20000000
PTCH2SRV_IMAX,1500
PTCH2SRV_P,0.69999999
PTCH2SRV_RLL,1.10000002
PTCH2SRV_RMAX_DN,0
PTCH2SRV_RMAX_UP,0
PTCH2SRV_TCONST,0.50000000
and roll gains:
RLL2SRV_D,0.05
RLL2SRV_I,0.2
RLL2SRV_IMAX,1500
RLL2SRV_P,1
RLL2SRV_RMAX,0
RLL2SRV_TCONST,0.5
and a couple of other important ones:
MIXING_GAIN,0.7
NAVL1_PERIOD,20
The article with the param file was written in march 2014 though, so Arduplane would have likely changed quite a bit since then - do you recall any refactoring of PID values in more recent versions?