Today I launched my AR Wing Pro as I have done dozens of time before. However, this time I had changed the video system from a 5.8Ghz type to a 1.2Ghz system and nothing else changes other than upgraded to 4.6.0 arduplane. Normally I use auto Take off and send the plane as shown in the video. It flies 200m out and start circling at elevation 60m. This time it does not keep the elevation but loose hight until it hitted the ground. For transmitter I use TBS Tracer which is a 2.4Ghz system on a Radiomaster TX16. This might not be compatible with the 1.2Ghz video system (200mW)? I have attached the *.BIN file. I dont have the skills to analyse the file so can any please look if it is possible to locate the reason for not keeping the 60m elevation. Please let me know if further info is required. The video is just from a GoPro camera. I did not have the time to get the googles on. Thank you. /Brian
Video of crash
BIN file for analysis
It looks like your plane was unable to maintain roll control and TECS couldn’t handle that correctly.
It looks like yaw rate was coupled into pitch rate controller at large roll and prevented it from increasing elevator deflection.
TBH I would increase loiter radius, so it isn’t loitering at 40° roll.
Just to echo on what Lupus said. What seems to have happened is that your altitude and speed controller (TECS) was not able to keep the set altitude in loiter mode due to the fact that the loiter radius set was too low WP_LOITER_RAD = 45. Since the Roll_Limit_Deg = 45, the aircraft has rolled up 45 degrees trying to maintain the 45m radius, this high of a roll caused the pitch to be coupled with roll and the controll was unable to pitch up to maintain the correct altitude and started losing it up to the point of the crash. The VTX change had nothing to do with it, throttle was kept at max and the battery voltage/current was healthy. This most likely means that you changed something in the parameters between the flights or have not used the TAKEOFF mode before. If you insist you didin’t please send a log of a healthy flight with auto TAKEOFF from before so that I can compare.
I would also point out that there was significant roll control deviation. At 45° it would probably have been fine but at 60 the controller couldn’t keep up.
Hi, I have now checked my random backups of the *.param files and the loiter radius was earlier 80. I have no idea of why it changed to 45. I assume that updating Ardupilot are not changing any values. Sorry for not checking but thanks a lot that you spotted it. I have made dozens of take offs with 80 without issues even in windy conditions.
I will give it another try when the it stops raining 
Earlier succesfull takeoff flight where it loiters a lot before I got to the transmitter: https://www.dropbox.com/scl/fi/yllwwly72szqtv4ner590/00000038.BIN?rlkey=x1xv59v2masyysn7fdp0vjpty&dl=0
Param file: https://www.dropbox.com/scl/fi/y13m2vjj9t8ji0lloq94x/FullParameterList_22-07-2025.param?rlkey=l11cvb0kug7xnepfakw91n9bj&dl=0
i guess this is something to do with the weird channeling you used
you are definitely not using standard order
this is the answer why i see the RC2 are not having output at all
and when i saw the rest of your servos options it made me give up
because the rest can only be seen from your cabling
I wouldn’t use any automatic or assisted mode prior to AUTOTUNE completion.
I use shake to wake for delaying the motor start, works like a charm.
Thank you all for your input to my first post.
Since the intend to autolaunch with v4.6.2 I have changed the loiter radius to 90 meter without luck. The AR Wing Pro still crashes when going into loiter after autotakeoff. I have now downgraded to Arduplane v4.3.4 as I had a .param backup for that version. However, I just downgraded to v4.3.4 and all setup was still there and I did nok make use of the .param backupfile. Yesterday I tried again and now with success :-). The plane goes nicely into loiter after autotakeoff is completed and I can put on my googles and start flying. My humble conclusion is that something has changed in v4.6.2 that does not agree with my setup.
The .BIN can be found here: BIN file
A video of the successful autotakeoff is found here: Video of launch, flight and landing
Let me know if any questions or feedback to my setup or equal, thanks.
Hello, I’m sorry to hear about the crash again. The *.bin file from the crash with 4.6.2 and 90 meters of radius would be interesting in comparison to 4.3.4.
Rolf
I can find no such stale PRs. Since you can point to exactly where in the code you’d like a change, why don’t you craft a PR to fix the issue?
Ok, so there was a recent merge about which you complained before it was merged AND after.
You can point to the exact behavior you want. A PR is welcome.
Hi Rolf, please find the file in the below and the associated video.
.BIN file here: BIN file
Videofile here: Video link
In the video I disarm at 30sec to avoid a crash at full throttle.
You flew Fw 4.3.4 (..00048.bin) with significantly more efficient roll PIDs than Fw 4.6.2.(..00046.bin).
Poor roll compliance ultimately leads to a crash.
I would like to be convinced otherwise, but I still don’t see how the firmware version could be relevant.
Rolf
Hi, during the latest flight I made an autotune mostly on elevator but was interupted by poor video. Perhaps thats why it has changed?
Yes.
But to have an overview, let’s summarize:
In the log browser, the top line shows your file number and the Arduplane version.
With radii that are far too tight, I can well imagine that flying wings sometimes spiral downwards in the loiter circle and sometimes don’t.
However, I still don’t understand why the flight failed with a 90-meter loiter circle and the previous roll parameters.
Rolf
Same here. I dont know why it failed. It takes off well and when it starts to loiter in right hand direction it goes straight to the ground and I disarm at 30sec. It is possible to hear the beep in the video. From here the plane straights out and crashes “gently”.
Hi, I use the standard 8X5" prop and spins clockwise looking from the back. When launching the plane overhead I dont have issues with torque roll. I have never had a motor rip out of the fuselage but I see your point and will look into setting the motorbrake. Actually it helps to take off speed with the prop turning.
Please don’t discuss offtopic political opinion or we will moderate post again.
There is nothing worth mentioning on this, there was no incidents in drone sense.
Thanks