Plane 4.4 release

kunlin · November 12, 2023, 12:17am

Hi, this is my log file.

tridge · November 15, 2023, 4:34am

I have just released plane 4.4.3. Changes from 4.4.2 are:

fixed setup of ICM45486 IMU on CubeOrangePlus-BG edition
disable AFSR on IxM42xxx IMUs to prevent gyro bias for “stuck” gyros
fixed AK09916 compass being non-responsive
implement GPS_DRV_OPTION for using ellipsoid height in more GPS drivers
fixed SIYI AP_Mount parsing bug
configuration fixes for BETAFTP-F405 boards
fixed origin versus home relative bug in quadplane landing and guided takeoff
correct mavlink response for no airspeed sensor on preflight calibration
protect against notch filtering with uninitialised RPM source in ESC telemetry
allow lua scripts to populate full ESC telemetry data
added YJUAV_A6SE_H743 support
fixed uBlox M10 GPS support on boards with 1M flash

Happy flying!

kunlin · November 20, 2023, 12:59am

@tridge Hello. May I ask if there are any abnormalities found in the log?

tridge · November 21, 2023, 8:43am

I have had a look and see the log ends in flight. Some questions:

I note that this is not a flight controller we support in ArduPilot. I see you have an open PR to add this board here: Support for ElfinH7 Wing by suiaiyun · Pull Request #24543 · ArduPilot/ardupilot · GitHub - we should try and get that merged in
can you please find the tlog (telemetry log) for the flight on the GCS - that may offer us some clues as to what went wrong
can you tell me what telemetry radio you are using? It is logging RAD messages which would imply SiK or RFD900, but the RemNoise field is always 255 which is very strange
the firmware is 4.4.0, but we did not support the H750 with CAN in 4.4.0. I see you have backported the changes, which should work, but means you are running untested code
it isn’t clear to me how you are doing parameter storage on this board? The hwdef.dat does not seem to define any mechanism

Hopefully the tlog will give us some clue as to what went wrong.

kunlin · November 21, 2023, 9:32am

During the flight, I used the mobile version of QGC, and by default, I did not save the tlog log. I also did not find it on my phone’s memory card.
The remote control I am using is MLRS, which supports Mavlink data transmission, similar to the remote control of Siyi.
The firmware version I am using is 4.4.0. I tried to compile 4.4.2 and it was successful, but it did not run successfully on my flight control. It stopped at the bootloader jump. （I set the parameter “FLASH_SIZE_KB” to 2048 to include all the functions.）
The parameters are saved on the tf card, in a file called arduplane.cfg.

There is a random issue: there is a probability of getting stuck in the bootloader jump during the startup process and not being able to enter the flight control system normally. I don’t know if this accident is related to this.

trag123123 · November 22, 2023, 9:10pm

@tridge Hello, I was using the firmware 4.0.7 for my quad plane and it was flying perfectly during quad mode. I updated the firmware to 4.4.0. The plane started to shake during take-off and also during landing, which led to it falling several times. Is there a major change in the parameters? Was the code causing the problem, or what?It was all the crashes that happen to the plane during landing, Unfortunately, I stored the log in the memory that crashed together with the plane crash. The remote measuring device I am using is not compatible with the mission, but I have the parameters that were on the plane. My problem is with the code parameters only. I downloaded the parameters that were in version 4.0.7 to The parameters in version 4.4.0 and I did not change any of them, but I saw that with the new version 4.4.0 during landing my plane was unable to balance and then failed to land. These are the parameters that I used.As for the rear throttle, I turn it off using the command DO_REPAT_SERVO just before the landing point.

trag123123 · November 22, 2023, 9:10pm

2023-11-04 08-27-29.param (16.4 KB)

Cheers, Tridge

makeflyeasy · November 25, 2023, 3:12am

Looking forward to the 4.4.4 update.

tridge · December 5, 2023, 7:42pm

I’ve just released plane 4.4.4-beta1.
Changes from 4.4.3:

CubeOrange Sim-on-hardware compilation fix
RADIX2HD supports external I2C compasses
SpeedyBeeF405v4 support
DroneCAN battery monitor with cell monitor SoC reporting fix
ProfiLED output fixed in both Notify and Scripting
NTF_LED_TYPES parameter description fixed (was missing IS31FL3195)
Scripting bug that could cause crash if parameters were added in flight
STAT_BOOTCNT param fix (was not updating in some cases)
don’t query hobbywing DroneCAN ESC IDs while armed
clamp empy version to prevent build errors

Please test!

tridge · December 7, 2023, 10:02pm

@trag123123 that parameter file is not complete. Do you have the bin log from a flight?

trag123123 · December 11, 2023, 2:57pm

@ tridge

There is a problem that I am facing with Firmware 4.4.0 due to the value of max_t (main loop rate). I connected to the Cuav v5+ system via the com port (USP), and I ran the system with version 4.4.0 and set the parameters to default after I activated q_enable, also the value SCHED_LOOP_RATE is 300, which is the default value. I made the system run for a while, switching modes, then I analyzed the .bin record attached below, and I saw that the value of max_t = 5 and 4.5, and the correct value for it is 3.3ms. This problem did not appear to me when I was using version 4.0. 7. What is the solution please?: 00000001.BIN - Google Drive

and This is another log,Together form 4.4.3 It’s the same problem:

and This is another log with the 4.0.7 Firmware and its default parameters, also with the same steps q_enable = 1 and SCHED_LOOP_RATE = 300, it was connected to the system via com (USP), and with the same system used for 4.4.0 Cuavv5+, it is noted that the value of max_t here is at 4.0 .7 equals 3.3ms.

tridge · December 20, 2023, 11:55pm

@trag123123 I just tested a CUAVv5 with 4.4.4 and I do not see any scheduling errors when I load the parameters from your log. Please re-test with 4.4.4.
Note that you can see scheduling errors without downloading the log file by setting SCHED_DEBUG=1 and looking in the messages tab. This is what I see:

AP: PERF: 0/1500 [3446:3209] F=300Hz sd=11 Ex=0
AP: PERF: 0/1500 [3420:3245] F=300Hz sd=10 Ex=0
AP: PERF: 0/1500 [3463:3203] F=300Hz sd=10 Ex=0
AP: PERF: 0/1500 [3419:3247] F=300Hz sd=9 Ex=0
AP: PERF: 0/1500 [3495:3169] F=300Hz sd=14 Ex=0
AP: PERF: 0/1500 [3410:3258] F=300Hz sd=9 Ex=0
AP: PERF: 0/1500 [3432:3234] F=300Hz sd=9 Ex=0
AP: PERF: 0/1500 [3424:3235] F=300Hz sd=9 Ex=0
AP: PERF: 0/1500 [3439:3227] F=300Hz sd=10 Ex=0
AP: PERF: 0/1500 [3434:3232] F=300Hz sd=10 Ex=0
AP: PERF: 0/1500 [3458:3210] F=300Hz sd=9 Ex=0
AP: PERF: 0/1500 [3454:3216] F=300Hz sd=12 Ex=0
AP: PERF: 0/1500 [3454:3214] F=300Hz sd=9 Ex=0
AP: PERF: 0/1500 [3422:3244] F=300Hz sd=10 Ex=0
AP: PERF: 0/1500 [3416:3250] F=300Hz sd=9 Ex=0
AP: PERF: 0/1500 [3413:3257] F=300Hz sd=11 Ex=0
AP: PERF: 0/1500 [3426:3240] F=300Hz sd=11 Ex=0
AP: PERF: 0/1500 [3469:3200] F=300Hz sd=11 Ex=0
AP: PERF: 0/1500 [3413:3250] F=300Hz sd=9 Ex=0
AP: PERF: 0/1500 [3465:3198] F=300Hz sd=11 Ex=0

this shows no long loops (0/1500) and a small loop standard deviation

tridge · December 20, 2023, 11:56pm

I’ve just released ArduPilot plane 4.4.4. Changes from 4.4.3:

CubeOrange Sim-on-hardware compilation fix
RADIX2HD supports external I2C compasses
SpeedyBeeF405v4 support
DroneCAN battery monitor with cell monitor SoC reporting fix
ProfiLED output fixed in both Notify and Scripting
NTF_LED_TYPES parameter description fixed (was missing IS31FL3195)
Scripting bug that could cause crash if parameters were added in flight
STAT_BOOTCNT param fix (was not updating in some cases)
don’t query hobbywing DroneCAN ESC IDs while armed
clamp empy version to prevent build errors

Happy flying!

trag123123 · December 25, 2023, 1:10pm

Fine, thank you
Cheers, Tridge

YupsUAV · January 12, 2024, 12:35pm

Had some very weird EKF3 behaviour because of GYRO data corruption. Please have a look at my thread:

jason_tonight · January 13, 2024, 1:56pm

Hello, I used plane 4.4 to fly the quadplane On the first flight, the route moved at a low frequency in roll. After the flight landed, I modified the ff and p of roll, and modified the ff and p of pitch. When I took off again, the plane hit a tree. Yes, the elevator is invalid. I speculate that the ff of the pitch should be increased to be correct. I would like to ask if the new pid has a greater weight on ff. P only affects the speed of reaching the angle.

makeflyeasy · January 20, 2024, 1:52am

@tridge Looking forward to the next version update.

trast911 · February 9, 2024, 9:59am

Good day everyone. I have had issues since I updated to 4.4.4 quadplane firmware on my MFE striver equipped with Pixhawk 6c and f9p holybro. Before i’ve been updated from 4.3.7, everything was working fine, but now when I’m trying to takeoff in QHOVER manually vehicle starts to spinning by YAW axis and performs unstable hovering causing a crash. Everything was recalibrated and still have no result.

Osman · March 10, 2024, 9:02am

Hello Tridge,
Wondering if you’d have some time to review the dataflash log from my recent quadplane flight that ended up in a crash. I have a video of the crash as well that can further help analyzing what went wrong. Let me know and I will upload both the log file and video for your kind advice.

Thanks
Regards
Osman

tridge · March 12, 2024, 8:08pm

This was the first post by your user ID on this forum, so I don’t know how to find the log or crash report you are referring to. Where did you post it?