Copter-4.6.0-beta3 has been released and can be installed using MP or QGC’s “Beta firmwares” link. Alternatively it can be downloaded directly from firmware.ardupilot.org
The changes vs -beta2 are in the ReleaseNotes and copied below
Board specfic changes
AEROFOX-H7 support
AET-H743-Basic support
ESP32 memory initialisation fix
MicoAir743AIO and MicoAir743v2 support
OrqaF405Pro supports ICM42688 IMU
TBS Lucid Pro support
Plane related enhancements and bug fixes
RC aux channel option for C++ QuikTune
TECs uses external HAGL (if available) for landing
DDS/ROS2 pre-arm check and copter takeoff service
Bug Fixes and minor enhancements
BATTx_OPTIONS param desc fixed
BLHeli telemetry ESC index fixed on boards with IOMCU
CAN manager logging critical race condition fixed (only occurred if CAN_LOGLEVEL was 1 or higher)
RunCam/OSD menu movements obey RC channel reversal
Topotek camera gimbal pitch rate control direction fixed
TradHeli AutoTune rate and accel limiting fixed
VTOL-quiktune script gets QUIK_ANGLE_MAX to prevent bad oscillation if tuning algorithm fails
In general 4.6.0 seems quite stable but there is one stubborn issue to do with QuadPlane attitude control that needs to be cleared up before we can promote 4.6.0 to be the stable release.
As always we very much appreciate any help we can get with beta testing so all feedback is very welcome!
woops deleted by mistake.
Loiter is working ok and stabilize is a little up and down overall ok. default parameters with a small kick to PID’s currently.
Out of level landing seems to distract the loitor landing from shutting down see video, I also don’t think i could disarm until switching to Stabilize then the bird leveled and i was able to stick disarm. I guess you can call this a landing runaway. Update. I just ran a s6 battery on a test flight and the test feels more concise with the KV and prop size I am using.
Thanks. I think the landing issues are more to do with Loiter mode possibly jamming into the soft earth not knowing what to do and thinks its still airborne. Notice loitor did not shut down and ramped up the power, only until i switched to Stabilize could i shut it down. I do see this error from time to time but not always.
To recap. Loitor landed fine but did not shut down, more power continued even that the stick was down lifting the rear of the craft up like a runaway. After waiting I switched to Stabilized the motors powered down and i manually disarmed. It’s like the old jump bug at landing.
Always room to improve, but I run some very good hardware mostly plug and play and was a tester during production. My biggest issue is trusting the battery settings. I getting a bad battery on this version of firmware on a s4. My charger seems to think is ok. Moving to s6 has improved the flight and does not give an error.
Set up a switch for Land mode - you can even use a 3 position switch for Land and RTL (I have to remember how I did mine… )
Land mode descends straight down and disarms quickly.
I wonder of the battery voltage issue is just the sensor calibration, since nothing has changed for quite some time in this regard in the code.
“sensor calibration” I had to turn the battery monitor off this time because of the bad battery error using 4s. but I normally just set it to 4 in the parms and measure the voltage. My new s6 does not give a error tho. What are your steps when setting up the battery?
I connect a power supply set to the lowest voltage that would ever occur, or a low voltage battery, like 3S for a 4S system, and measure the actual voltage applied.
Enter it in the Battery monitor calibration “measured voltage” field.
This should give an accurate reading at low voltages where it matters most.
And I use the failsafe values from the MissionPlanner - Initial Parameters widget, there’s various Lipo and LiIon options in there too.
BTW, we’ve added a new LDET log message to help with support calls around failure to detect landing (PR is here). I’m pretty confident that there haven’t been any changes in this area in 4.6 vs 4.5.
Update. I just ran a s6 battery on a test flight and the test feels more concise with the KV and prop size I am using.