Dev Call Apr 14, 2024

CraigElder · April 15, 2024, 10:38pm

Methodic Configurator

Issues & Pull requests

github.com/ArduPilot/ardupilot

solved the logical error in grouping of escs into 4

ArduPilot:master ← adityaomar3:ap-esc

opened 07:34PM - 30 Sep 23 UTC

adityaomar3

+4 -1

ESC_Telemetry will not send MAVLink MSG_ESC_TELEMETRY_1_TO_4 if ESC_TELEM_MAX_ES…CS (PWM/SERVO) count is <=3 as logic was to keep idx = ESC_TELEM_MAX_ESCS/4 It gives rise to 3 big issues that are :- 1.It will leave some values if ESC_TELEM_MAX_ESCS which are not completely divisible by 4 e.g. 1,2,3,5,6,7,9.... 2. (next_idx + idx) % num_idx it will give error if escs are less than 4 3. for loop will never start if escs count is less than 4 i.e the message inside loop will be skipped even if escs are present. Also if escs are not in multiple of 4 then it will skip 1 iteration Solution in commit: This is rectified by logic to add 1 to num_idx if ESC_TELEM_MAX_ESCS is not multiple of 4.

github.com/ArduPilot/ardupilot

GCS_MAVLink: respect txbuf flow control for FTP messages

ArduPilot:master ← brad112358:ftp_txbuf

opened 03:53PM - 30 Jan 24 UTC

brad112358

+32 -17

This gives slow radio links a fighting chance of getting FTP bulk download worki…ng efficiently, even when they use a baud rate which is much higher than their current bandwidth. The strategy is similar to what Ardupilot already does for PARAM_VALUE messages. This should eliminate the need to disable FTP for parameter download on slow to moderate speed radio links like mLRS. It allows removal of a hack in mLRS which results in a decrease in parameter download time for 19 Hz mLRS from 45-60 seconds to 11-17 seconds. Other mLRS message rates show similar improvement. This should also be good news for the ELRS rc-mavlink branch. Also, improve get_last_txbuf() to allow message flow in the unlikely event that we stop receiving radio_status messages. This has also been tested with TBS Crossfire and SiK radios and found to have no negative impact on parameter download.

github.com/ArduPilot/ardupilot

AP_HAL_ChibiOS: Pixhawk6X peripheral power on at startup

ArduPilot:master ← rmackay9:chibios-pixhawk6x-periph-power

opened 10:21AM - 10 Apr 24 UTC

rmackay9

+6 -8

This addresses the Hereflow startup issue recorded in the [4.5.0 issues list](ht…tps://github.com/ArduPilot/ardupilot/issues/26103) and discussed [here on the forums](https://discuss.ardupilot.org/t/copter-4-5-0-released/115614/47). The issue is that the HereFlow sensor connected to a Pixhawk6X does not start successfully with 4.5.0 but does with 4.4.4. The temporary work around is for users to set BRD_BOOT_DELAY to 3000 which slows AP's startup until after the flow sensor has initialised. This PR includes a better fix by providing power to the Hereflow sensor as soon as the autopilot is powered on. Note that this change affects more than just the CAN port (but I'm not personally sure what other ports are affected). This has been tested on real hardware including: - Confirmed the user's report is correct. E.g. Pixhawk6X with Hereflow with 4.4.4 works, 4.5.0 does not - Confirmed that setting BRD_BOOT_DELAY to 3000 allowed 4.5.0 to work - Confirmed that changing the nVDD_5V_PERIPH_EN to LOW allowed 4.5.0 to work with BRD_BOOT_DELAY = 0 This PR changes both the "HIPOWER" and "PERIPH" definitions. In fact, only the 2nd definition must be changed to resolve the issue but I suspect they should be consistent. From a quick search for "nVDD_5V_PERIPH_EN OUTPUT HIGH" these other boards may have similar issues: - ARKV6X - DrotekP3Pro - Durandal - fmuv5 - JFB100 - PixC4-Jetson - Pixhawk5X - RadiolinkPIX6 I'm unsure if other boards might have similar problems.

github.com/ArduPilot/ardupilot

AP_NavEKF3: don't look at compass learning with GPS yaw

ArduPilot:master ← tridge:pr-ekf3-compass-fallback

opened 05:48AM - 11 Apr 24 UTC

tridge

+6 -2

if using GPS yaw we don't want to do a yaw reset on completion of GPS yaw, so ju…st don't detect compass learning this avoids an issue where COMPASS_LEARN=3 is set in flight with EK3_SRC1_YAW=3 for GPS with compass fallback. That leads to a yaw alignment reset on completion of compass learning which leads the EKK to diverge we really need to better understand the divergence as well Some clues: - when the bug happens we end up with huge compass offsets. The compass_learn=3 system uses GSF, which should be unaffected, so this doesn't make sense

github.com/ArduPilot/ardupilot

Plane: Fix FBWB/CRUISE missing zero crossing of elevator input

ArduPilot:master ← pieniacy:kp/fix-fbwb-cruise-alt-windup

opened 02:15PM - 14 Apr 24 UTC

pieniacy

+7 -6

# Summary In FBWB/CRUISE Plane can drastically overshoot altitude and continue …to climb/descent long after pilot gives opposite elevator stick input, resulting in crashing into the ground or drastically overshooting the ceiling. This fixes the issue. # Scenarios ``` CRUISE_ALT_FLOOR 0 FBWB_CLIMB_RATE 5 TECS_SINK_MIN 2.5 TECS_SINK_MAX 5 ``` ### Descent Imagine you are at 120 meters in CRUISE and want to do your max descent with lowest airspeed. You lower your throttle to zero and command full nose down. You requested slowest airspeed, so your plane will not be able to sink 5 m/s as directed, more like 2.5 m/s. You descent to 20 m and want to pull up, so you move your elevator stick for full nose up, you expect your airplane to pull up and climb, but it does not, it continues down and will hit the ground. ### Climb Imagine you want to climb from 20 to 120 meters in CRUISE but you want a fairly high airspeed. You raise your throttle and command a full nose up. You requested faster airspeed, so your plane will not be able to climb 5 m/s as directed. At 120 meters you realize you reached your maximum legal altitude, so you move your elevator stick for full nose down and expect your airplane to level off and descent, but it does not, it continues up and will probably reach an altitude of over 150 meters. ## Reproducing Both scenarios can be easily reproduced in SITL. I can confirm it is also a case with a physical airplane on ArduPlane 4.5.1 with ELRS (at 50 Hz). # Reason https://github.com/ArduPilot/ardupilot/blob/233f344a37e177343f372bec31cb2cc5890a26ee/ArduPlane/navigation.cpp#L400-L404 In `Plane::update_fbwb_speed_height(void)` a condition to reset target altitude only checks if elevator input became zero, but does not check for zero crossing (pilot moving the stick through zero). So the behaviour is different if pilot moves the stick slowly, in which case a zero will be registered and altitude will be reset and moves the stick fast in which case it will not be registered and the height setpoint has been wound up by as much as 100 meters. What makes the issue even worse is that the loop to update height target is running at only 10 Hz: https://github.com/ArduPilot/ardupilot/blob/233f344a37e177343f372bec31cb2cc5890a26ee/ArduPlane/navigation.cpp#L383 This makes it even easier to jump over the deadzone with a realistic stick movement. # Solution The only change in this PR is to replace the condition to check for the input becoming zero to also check zero crossings. I believe the most efficient way to do this is to check if the input is no longer positive or no longer negative. It preserves the previous behaviour and also triggers if the input jumped over zero. In terms of implementation one could alternatively multiply previous and current input and check for negative (indicating different signs), but this way it should be more efficient as it does not require floating point multiplication, which can take a few more mpu cycles. # Resources Here you can see before and after the fix from SITL. Notice that on the left image the aircraft crashes into the ground despite full pull up being commanded for 10 long and terrifying seconds. ![image](https://github.com/ArduPilot/ardupilot/assets/7950377/3b18dab6-fb42-471b-b8e6-92a0195f10cc) Here are the logs to produce above figures: [FBWB_BUG_BEFORE_FIX.BIN](https://drive.google.com/file/d/1-iWT6s_3F7-b-tqa43ZhoZBfrCK26RsO/view) [FBWB_BUG_AFTER_FIX.BIN](https://drive.google.com/file/d/1I4wbHaIzGliulEPAZmbHEae-xpLOdBO-/view) # Remarks [Documentation of FBWB](https://ardupilot.org/plane/docs/fbwb-mode.html) is not very clear on the somewhat hidden altitude target, which is not always intuitive in FBWB/CRUISE. Most users will not be aware that such target altitude windup can occur. The fact of the target winding up in some configurations of commanded airspeed/vspeeds is disputable, one could argue that the stick should indeed set the hidden altitude target strictly according to the `FBWB_CLIMB_RATE` without taking into account real conditions like slow airspeed or randomly encountered thermal lift. On the other hand flying in FBWB/CRUISE tends to be rather interactive (often FPV) and the pilot probably expects that elevator input will control current vertical speed, not an altitude target that is not known when flying VLOS or FPV. This can lead to crashes or (subjectively) uncommanded climbs. # Severity In my opinion the bug is severe and should be fixed asap. The fact that an airplane will behave differently whether you rise your elevator stick slow or fast is in my opinion not acceptable. What's worse, slow stick movement will register zero and reset the altitude target. However, faster stick movement, which might occur suddenly in response to approaching ground, obstacles, or other aircraft, will not reset the altitude target. Instead, it will continue the climb or descent until the setpoint unwinds, which for perfectly realistic scenarios can easily take 10-15 seconds, resulting in crashes.

github.com/ArduPilot/ardupilot

AP_NavEKF3: removed velocity corrections for IMU position

ArduPilot:master ← tridge:pr-ek3-ins-pos-no-vel

opened 08:42PM - 15 Apr 24 UTC

tridge

+4 -12

this is a possible fix for #20710 the velocity corrections introduce a lot o…f noise in the reported velocity to external controllers. With the velocity correction enabled a VTOL aircraft may oscillate badly if the IMU position is significant. Using replay I have experimented with using dal.ins().get_gyro() instead to add some filtering, but a fair bit of noise still gets through. The best results in replay have been with only applying the position corrections and not the velocity corrections This is an example on a large quadplane: ![image](https://github.com/ArduPilot/ardupilot/assets/831867/d5df991c-8ed2-4191-b19e-d00d1550af3d) the yellow line shows the much better velocity east output with this change versus the original in blue this aircraft has: - INS_POSn_X = 0.5 - GPS_POS1_X = -0.1 - GPS_POS1_Y = -1.2 the issue was noticed as the aircraft oscillated in QLOITER mode with the correct INS and GPS position parameters. Setting INS_POSn_X==0 stopped the oscillation Some other possibilities: - use the gyro from AP_InertialSensor which has been through notch filtering. The IC engine in this aircraft produces a lot of noise which is reduced with notch filtering using RPM

Plane

Copter & Rover

github.com/ArduPilot/ardupilot

Copter/Rover-4.5.0 issues list

opened 03:30AM - 31 Jan 24 UTC

rmackay9

Copter Rover DevCallTopic ReleaseAdmin

This is a list of issues discovered during Copter and Rover 4.5 testing. The [is…sues list for Copter/Rover-4.4 is here](https://github.com/ArduPilot/ardupilot/issues/23192). [Backport PRs](https://github.com/orgs/ArduPilot/projects/8/views/1). Reports requiring investigation: - [ ] Lua rangefinder broken ([discussion](https://discuss.ardupilot.org/t/send-sensor-distance-by-using-serial-in-script/116238)) - [ ] Gremsy can't be used as 2nd mount ([discussion](https://discuss.ardupilot.org/t/gremsy-zio-does-not-work-if-set-to-mount-2/116127)) - [ ] Stabilize mode Internal Error in Land Detector ([discussion](https://github.com/ArduPilot/ardupilot/blob/master/ArduCopter/land_detector.cpp#L59), [code](https://github.com/ArduPilot/ardupilot/blob/master/ArduCopter/land_detector.cpp#L59)) - [ ] Guide mode using thrust causes internal error ([PR](https://github.com/ArduPilot/ardupilot/pull/26784)) - [ ] 3 of 12 Pixhawk6X display IOMCU error ([discussion](https://discuss.ardupilot.org/t/copter-4-5-1-released/115973/3)) - [ ] Hereflow requires BRD_BOOTDELAY ([discussion](https://discuss.ardupilot.org/t/copter-4-5-0-released/115614/47), [PR](https://github.com/ArduPilot/ardupilot/pull/26753)) - [ ] SmartRTL flow of control error ([pre-existing issue](https://github.com/ArduPilot/ardupilot/issues/17416)) - [ ] DShot 600 jerky movement without being armed ([discussion](https://discuss.ardupilot.org/t/copter-4-5-0-released/115614/21)) -- configuration error (MOT_PWM_TYPE = 0 but SERVO_BLH_MASK non-zero) - [ ] Pixhawk 6C SBUS out not working with Tarot gimbal ([discussion](https://discuss.ardupilot.org/t/copter-4-5-0-released/115614/12)) - [ ] Lua mount:set-angle-target binding doesn't work on Matek H743 mini V3 ([discussion](https://discuss.ardupilot.org/t/copter-4-5-0-released/115614/3)) - [ ] RPLidarS1 erratic distances on Rover ([discussion](https://discuss.ardupilot.org/t/rplidar-s1-erratic-distances/115561), [possible fix](https://github.com/rmackay9/rmackay9-ardupilot/commits/rover-rplidar-update-rate/)) - [ ] CAM messages with Img field "0" appearing in logs without pictures being taken (reported by Randy) - [ ] Throw mode reliability ([discussion](https://discuss.ardupilot.org/t/diagnose-unpredictable-throw-mode/113314)) Confirmed issues: - [ ] Jump message missing from telementry? ([discussion](https://discuss.ardupilot.org/t/copter-4-5-1-released/115973/25), [issue](https://github.com/ArduPilot/ardupilot/issues/26802)) - [ ] EKF loses attitude estimate if COMPASS_LEARN set to 3 (inflight) and GPS-for-yaw used ([PR](https://github.com/ArduPilot/ardupilot/pull/26762)) - [ ] EKF Wind estimates not available when using GPS-for-yaw ([issue](https://github.com/ArduPilot/ardupilot/issues/26650)) - [ ] APM_Config can't be used to include/exclude features ([issue](https://github.com/ArduPilot/ardupilot/issues/26654)) - [ ] Neopixels broken ([discussion](https://discuss.ardupilot.org/t/multiple-neopixels-dont-flash-in-sync/113620), [issue](https://github.com/ArduPilot/ardupilot/issues/26295), [PR-to-blame](https://github.com/ArduPilot/ardupilot/pull/23974)) Enhancements or Issues that may not be resolved before the stable release: - [ ] GPS consistency checks Alt when GPS can't provide alt ([discussion](https://discuss.ardupilot.org/t/incorrect-calculation-of-gps-positions-diffs-during-arming-checks/116033), [issue](https://github.com/ArduPilot/ardupilot/issues/26749)) - [ ] Custom Build server support for CUAV Pixhack v3x with 3 IMUs ([discussion](https://discuss.ardupilot.org/t/ardupilot-custom-firmware-builder-fmuv3-missing-necessary-for-cuav-pixhack-v3x/115823), [product page](https://www.cuav.net/en/pixhack-v3-autopilot/)) - [ ] AP_OSD: Add CRSF link stats display fields ([PR](https://github.com/ArduPilot/ardupilot/pull/25844)) - [ ] Network params not visible for Pixhawk5x ([discussion](https://discuss.ardupilot.org/t/copter-4-5-0-beta2-available-for-beta-testing/113156/13), [issue](https://github.com/ArduPilot/ardupilot/issues/26397)) - [ ] ViewPro does not track GPS location well ([discussion](https://discuss.ardupilot.org/t/viewpro-gimbal-point-camera-here-function/98279/33), known limitation of ViewPro API) Wiki issues/enhancement requests: - [ ] Dynamic Notch filter instructions need consolidation ([issue](https://github.com/ArduPilot/ardupilot_wiki/issues/5878)) - [x] Auto mode LOITER_TURNS does not support CCW? ([discussion](https://discuss.ardupilot.org/t/ccw-loiter-turns/114282)) - [ ] Add vertical position controller tuning instructions ([discussion](https://discuss.ardupilot.org/t/z-oscillation-in-position-hold/113149/4)) - [x] Add warning for AutoTune ESC desync ([issue](https://github.com/ArduPilot/ardupilot_wiki/issues/5760)) Mission planner issues/enhancement requests: - [ ] fixup camera gimbal page ([discussion](https://discuss.ardupilot.org/t/question-about-camera-gimbal/101063/7)) - [ ] Loiter turns radius incorrect ([issue](https://github.com/ArduPilot/MissionPlanner/issues/3305)) QGC issues/enhancement requests: - [x] QGC sending invalid roll, pitch, yaw angles ([discussion](https://discuss.ardupilot.org/t/mount-invalid-angle-targets-problem/113885), [QGC issue](https://github.com/mavlink/qgroundcontrol/issues/11215)) Resolved issues: - [x] RTL engaged below Home alt and using Terrain alt returning too high? (reported by Tridge and PeterB) -- RTL seems to be acting as expected. It was engaged at 121m above terrain and so that's the altitude it maintained until it returned home. - [x] AP_Mount's default mode being overwritten by RC input ([discussion](https://discuss.ardupilot.org/t/copter-4-5-0-released/115614/69), [PR](https://github.com/ArduPilot/ardupilot/pull/26738)) -- resolved for 4.5.2-beta1 - [x] T265 not working ([discussion](https://discuss.ardupilot.org/t/vision-positioning-using-t265-does-not-work/115679), [related PR](https://github.com/ArduPilot/ardupilot/pull/23240)) -- user was using same sysid/compid as autopilot - [x] Developer issues integrating CAN RangeFinder ([discussion1](https://discuss.ardupilot.org/t/issues-with-interegrating-can-sensor-using-lua/115585/5), [discussion2](https://discuss.ardupilot.org/t/help-related-to-dronecan/115874), [issue](https://github.com/ArduPilot/ardupilot/issues/26704)) -- worked at 1Mbps - [x] GPS_TYPE parameter descriptions in MP using 4.4 ([issue](https://github.com/ArduPilot/ardupilot/issues/26681#issuecomment-2034416514), [PR](https://github.com/ArduPilot/ardupilot/pull/26740)) -- resolved for 4.5.2-beta1 - [x] ROS2 / DDS install issues ([discussion1](https://discuss.ardupilot.org/t/compiling-arducopter-4-5-0-with-enable-dds-failed/115632), [discussion2](https://discuss.ardupilot.org/t/colcon-build-packages-up-to-ardupilot-dds-tests-failed/115720)) -- caused by user's installed tools falling out of date. - [x] Near Fly-away with 4.4.4 (EKF failsafe was disabled, [discussion](https://discuss.ardupilot.org/t/internal-errors-0x400-l-856-cnstring-nan-uncontrollable-behavior-in-rtl-and-loiter/115598)) -- no further communication, likely bad compass cal - [x] KakuteH7 fails to initialise ([discussion](https://discuss.ardupilot.org/t/kakute-h7-gps-and-radio-problem/115931)) -- also failed on 4.4.4 so Randy has concluded that it is a hardware issue - [x] Siyi mount stalls occasionally and bouncing near pitch and yaw limits ([discussion](https://discuss.ardupilot.org/t/copter-4-5-0-beta4-available-for-beta-testing/115149/11)) -- Siyi firmware / hardware mismatch - [x] Pixhawk6C compass issue ([4.4 issue](https://github.com/ArduPilot/ardupilot/issues/25946), [4.5 discussion](https://discuss.ardupilot.org/t/copter-4-5-0-beta2-available-for-beta-testing/113156/3)) -- no feedback - [x] do-condition-yaw bug (also in 4.4, [PR](https://github.com/ArduPilot/ardupilot/pull/24719)) -- resolved for beta5 (assuming there is a beta5) - [x] Copters turns North in Guided_NoGPS ([issue](https://github.com/ArduPilot/ardupilot/issues/26357), [discussion](https://discuss.ardupilot.org/t/set-attitude-target-with-guided-no-gps-lack-of-documentation/111845/24), [PR](https://github.com/ArduPilot/ardupilot/pull/26407)) -- fixed for beta5 (assuming there is a -beta5) - [x] ViewPro not holding level during hard maneuvers ([PR](https://github.com/ArduPilot/ardupilot/pull/26556)) -- resolved in beta4 (although no confirmation from user) -- fixed for beta5 (assuming there is a -beta5) - [x] altitude jump when GPS/Non-GPS transitions switching altitude source of ExtNav to Baro (reported by Randy) -- should jump when moving from camera->baro, cannot reproduce otherwise. - [x] MP's MAVFtp can't download SYS files ([MP issue](https://github.com/ArduPilot/MissionPlanner/issues/3309)) -- SITL only issue - [x] Logging removed on some boards? ([issue](https://github.com/ArduPilot/CustomBuild/issues/70), [PR](https://github.com/ArduPilot/ardupilot/pull/26421)) -- needs backport? - [x] Sending unknown message ([issue](https://github.com/ArduPilot/ardupilot/issues/26358), [PR](https://github.com/ArduPilot/ardupilot/pull/26361)) -- resolved for -beta3 - [x] increasing k_format_version does not reset parameters (dev question, [issue](https://github.com/ArduPilot/ardupilot/issues/26336)) -- cannot reproduce, responses stopped - [x] MAV_CMD_DO_SEND_SCRIPT_MESSAGE not working? ([discussion](https://discuss.ardupilot.org/t/is-mav-cmd-do-send-script-message-unsupported-on-copter-4-5/113787)) -- only supported as Mission command - [x] TKOFF_RPM_MIN working? ([4.3.x discussion](https://discuss.ardupilot.org/t/takeoff-blocked-waiting-for-esc-rpm/113256)) -- fixed in 4.4 - [x] Record video while armed broken ([PR](https://github.com/ArduPilot/ardupilot/pull/26217)) -- resolved for -beta3 - [x] Parachute conversion issue ([issue](https://github.com/ArduPilot/ardupilot/issues/26132), [PR](https://github.com/ArduPilot/ardupilot/pull/26151)) -- resolved for beta2 - [x] Siyi running old camera fw message ([discussion](https://discuss.ardupilot.org/t/siyi-a8-camera-fw-version-reporting-issue/112870), [original discussion](https://discuss.ardupilot.org/t/siyi-a8-mini-4k-ai-mini-zoom-gimbal-camera-ai-identify-tracking-4k-1-1-7-inch-8mp-sony-sensor-6x-zoom/92133/673), [PR](https://github.com/ArduPilot/ardupilot/pull/26168)) -- resolved for beta2 - [x] Matek H743 watchdog ([discussion](https://discuss.ardupilot.org/t/locally-compiled-arducopter-4-5-0-beta1-fails-on-matekh743-due-to-watchdog/113029)) -- development environment issue - [x] Stuck thread ([discussion](https://discuss.ardupilot.org/t/stuck-thread-and-failed-data-logging-error/112711/2)) -- exists in 4.4.4 as well, two devs say it's fine so concluded it is a hardware failure on the user's autopilot. - [x] AutoTune crash -- ESC de-sync ([discussion](https://discuss.ardupilot.org/t/copter-4-5-0-beta1-available-for-beta-testing/112504/11)) - [x] Here4 does not support GPS-for-yaw (aka Moving baseline) ([discussion](https://discuss.ardupilot.org/t/gps-yaw-moving-baseline-with-dual-dronecan-here4/112357))

GSOC

Conference

CraigElder · April 16, 2024, 4:50am

Attendee count (max): 14

UTC1100 - Add test for copter CompassMot by peterbarker · Pull Request #19655 · ArduPilot/ardupilot · GitHub

UTC1100 - solved the logical error in grouping of escs into 4 by adityaomar3 · Pull Request #25156 · ArduPilot/ardupilot · GitHub

Correct ESC telemetry batch sending
Peter’s suggestion looks good
Can be merged once it is simpligied

UTC1110 - Add test for copter CompassMot by peterbarker · Pull Request #19655 · ArduPilot/ardupilot · GitHub

Add test for Copter compass-bot
Can be merged once conflict fixed

UTC1112 - hwdef: bdshot version of the MatekF765-SE by FOG-Yamato · Pull Request #25273 · ArduPilot/ardupilot · GitHub

Bdshot MatekF765-SE
Merged!

UTC1113 - Bounds correctness for AnalogIn_IIO by landswellsong · Pull Request #25747 · ArduPilot/ardupilot · GitHub

Bounds correctness on GPIO
Merged!

UTC1115 - GCS_MAVLink: respect txbuf flow control for FTP messages by brad112358 · Pull Request #26098 · ArduPilot/ardupilot · GitHub

Respect txbuf flow control for FTP messages
There’s a better radio_status message coming!
Merged!

UTC1117 - 3DR Control Zero H7 OEM Revision G support by TOTON95 · Pull Request #26463 · ArduPilot/ardupilot · GitHub

3DR Control Zero H7 OEM RevG support
Checklist for new boards
- Some clocks are getting very bad clock drift
  - 7 seconds across a flight….
- Simple test to see if that’s right
- Flow control pins actually work
- RC input actually works
- Crystal works properly
- Peter will fix up commit list if we’re otherwise happy

UTC1121 - Support for MicoAir405Mini by Minderring · Pull Request #26637 · ArduPilot/ardupilot · GitHub

PicoAir405Mini
Merged!

UTC1121 - Tools: added DroneCAN dependencies by tridge · Pull Request #26745 · ArduPilot/ardupilot · GitHub

DroneCAN in CustoimBuild server
Merged!

UTC1123 - AP_HAL_ChibiOS: Pixhawk6X peripheral power on at startup by rmackay9 · Pull Request #26753 · ArduPilot/ardupilot · GitHub

Peripheral power at startup
A few changes e.g. 6C 5X and Durandal supported
B like CubeOrange in terms of maintaining level

UTC1129 - SITL: correct diagnostic message obout what AirSim-in addr we bound by peterbarker · Pull Request #26759 · ArduPilot/ardupilot · GitHub

Correct SITL diagnostic error message
Merged!

UTC1131 - AP_NavEKF3: don't look at compass learning with GPS yaw by tridge · Pull Request #26762 · ArduPilot/ardupilot · GitHub

Don’t look at compass learning with GPS yaw
Paul’s got a better fix
4.5.2 for sure

UTC1133 - Copter/Blimp: remove redundant home is set check by peterbarker · Pull Request #26766 · ArduPilot/ardupilot · GitHub

Remove redundant check
Randy’s raised some questions, will be back with this one

UTC1137 - Copter/Blimp: remove redundant home is set check by peterbarker · Pull Request #26766 · ArduPilot/ardupilot · GitHub

Stop reaching into location object to get alt
Peter to change to get_alt_m and then merge

UTC1142 - AP_AHRS: remove call to get_home() by peterbarker · Pull Request #26769 · ArduPilot/ardupilot · GitHub

Remove call to get_home in AHRS
Merged!

UTC1147 - AC_AutoTune_Heli: factor out roll/pitch/yaw values by peterbarker · Pull Request #26780 · ArduPilot/ardupilot · GitHub

Clean up use of centidegrees accessors
Conflicts with Bill’s rework of autotune
Can wait
Peter to look at the autotest

UTC1152 - Plane: Fix FBWB/CRUISE missing zero crossing of elevator input by pieniacy · Pull Request #26796 · ArduPilot/ardupilot · GitHub

Plane: Fix FBWB/CRUISE missing zero crossing of elevator input
TECS time constant?
- Reset vertical filter
- Faster pull-up
- Time constant is very slow
FBWB or cruise
Elevator demands climb/descent rate
Looking for elevator at zero to set-altitude-current
Usually integrates stick to get a target height
Only resets when you neutralize the stick
Should happen on cross-over from positive to negative
But if you really move the stick rapidly then you can miss that cross-over
Max difference between target and actual?
Fbw climb rate too big warning?
Parameters based on FBWA measurements bad?
How do you test MAX_SINK_RATE?
Patch because we’re using TECS a bit wrong

UTC0001 - AP_GPS: remove dedundant crc32 routine from Nova by peterbarker · Pull Request #26798 · ArduPilot/ardupilot · GitHub

Remove redundant crc32 check
Merged!

UTC0002 - AP_NavEKF3: removed velocity corrections for IMU position by tridge · Pull Request #26808 · ArduPilot/ardupilot · GitHub

Remove velocity corrections for IMU position
Paul wants something slightly different
Bad on Copters where IMU is significantly different from CoG
Slight different change in here which is now Paul-approved
Avoid EKF being affected by user-configurable filters
- Would need to replay-log filtered IMU data
- This change with just doing the filtering gets 90% of the way there
Gyro noise vs velocity noise measurement ratio is a complex story
- The current PR doesn’t require that
- INS pos to zero and work to zero to avoid problems with it

UTC0013 - AP_VisualOdom.h：a bug in get_delay_ms(), return type error · Issue #26800 · ArduPilot/ardupilot · GitHub

Bug in get_delay_ms return type error
4-year old bug

UTC0018 - Plane update

Compass-learning-with-yaw
Velocity-correction-imu-position
- Can’t get good tune on big aircraft
- Only affects people with ins offsets
Fbwb in cruise
Beta next week

UTC0021 - Copter update

Weeks of bugfixes to do
Closed 5
Opened 7
Randy appreciates all the help from other DevTeam members
RTL coming back too high
- Tridge will go back and have a look at the log
If you terrain-follow on way home
- If you RTL when 100m above home
- It will come up 100m above terrain
- Never descends
Gully… terran following… terrain margin…. Wpnav_terr_margin
- 10m
- As it approaches the margin it will slow down
- 20m higher than RTL height
- Sitting at 400ft when RTL was triggered
- Climbed when it shouldn’t
- Needed to fly lower or set the terr_margin lower
- If you go over a cliff you’re going to breach a terrain limit….
- Will share log with Leonard
RC targeting bug for mount
- Default mode problem
T265 not working….
- Don’t consume our own message caught people out using 1:1
Repeat messages in waypoint navigation
- Doesn’t print message if there’s no limit on the jump
LUA rangefinder not working
- No autotest for it
- Tofsense-m1
Two mounts
- Not working now?
- Some do work according to Tim
Internal error from land detector
- One for stab, one for guided
Iomcu firmware load problems on PX4-IOMCU firmware
- Holding the safety down is problematic if you don’t have one connected to the IOMCU
GPSs which don’t give altitude are problematic?
Suspicious we’re not getting terrain offset flying correct
- WP_TERR_MRGN
- A lot of sharp pitching the 10m margin is threshold for slowing down
  - Sudden change to slow down
  - Need a proportional response?
    - It is?
    - Pitching over course of second
    - 60-degrees-per-second-desired?
      - Camera can only do 30….
    - Seems a bit much?
    - Maybe just look ahead for the stopping distance?
      - Less pitching?
    - Use slope from terrain database?
      - Stop the terrain being a surprise to it
- Two different margins, an upper and lower margin?
- Battery consumption vs speed
  - A lot less power at speed vs hovering
- Need to get Andy’s fence stuff in before we play too much
- Callisto at 17m/s uses same current as at hover
- 17% less battery on Angel at 11m/s than when hovering
  - 1200s on this Copter
  - Callisto is 20%
  - It’s airspeed, not groundspeed which matters!
  - Ideal is actually a target pitch

Dev Call Apr 14, 2024

Attendee count (max): 14

UTC1100 - Add test for copter CompassMot by peterbarker · Pull Request #19655 · ArduPilot/ardupilot · GitHub

UTC1100 - solved the logical error in grouping of escs into 4 by adityaomar3 · Pull Request #25156 · ArduPilot/ardupilot · GitHub

UTC1110 - Add test for copter CompassMot by peterbarker · Pull Request #19655 · ArduPilot/ardupilot · GitHub

UTC1112 - hwdef: bdshot version of the MatekF765-SE by FOG-Yamato · Pull Request #25273 · ArduPilot/ardupilot · GitHub

UTC1113 - Bounds correctness for AnalogIn_IIO by landswellsong · Pull Request #25747 · ArduPilot/ardupilot · GitHub

UTC1115 - GCS_MAVLink: respect txbuf flow control for FTP messages by brad112358 · Pull Request #26098 · ArduPilot/ardupilot · GitHub

UTC1117 - 3DR Control Zero H7 OEM Revision G support by TOTON95 · Pull Request #26463 · ArduPilot/ardupilot · GitHub

UTC1121 - Support for MicoAir405Mini by Minderring · Pull Request #26637 · ArduPilot/ardupilot · GitHub

UTC1121 - Tools: added DroneCAN dependencies by tridge · Pull Request #26745 · ArduPilot/ardupilot · GitHub

UTC1123 - AP_HAL_ChibiOS: Pixhawk6X peripheral power on at startup by rmackay9 · Pull Request #26753 · ArduPilot/ardupilot · GitHub

UTC1129 - SITL: correct diagnostic message obout what AirSim-in addr we bound by peterbarker · Pull Request #26759 · ArduPilot/ardupilot · GitHub

UTC1131 - AP_NavEKF3: don't look at compass learning with GPS yaw by tridge · Pull Request #26762 · ArduPilot/ardupilot · GitHub

UTC1133 - Copter/Blimp: remove redundant home is set check by peterbarker · Pull Request #26766 · ArduPilot/ardupilot · GitHub

UTC1137 - Copter/Blimp: remove redundant home is set check by peterbarker · Pull Request #26766 · ArduPilot/ardupilot · GitHub

UTC1142 - AP_AHRS: remove call to get_home() by peterbarker · Pull Request #26769 · ArduPilot/ardupilot · GitHub

UTC1147 - AC_AutoTune_Heli: factor out roll/pitch/yaw values by peterbarker · Pull Request #26780 · ArduPilot/ardupilot · GitHub

UTC1152 - Plane: Fix FBWB/CRUISE missing zero crossing of elevator input by pieniacy · Pull Request #26796 · ArduPilot/ardupilot · GitHub

UTC0001 - AP_GPS: remove dedundant crc32 routine from Nova by peterbarker · Pull Request #26798 · ArduPilot/ardupilot · GitHub

UTC0002 - AP_NavEKF3: removed velocity corrections for IMU position by tridge · Pull Request #26808 · ArduPilot/ardupilot · GitHub

UTC0013 - AP_VisualOdom.h：a bug in get_delay_ms(), return type error · Issue #26800 · ArduPilot/ardupilot · GitHub

UTC0018 - Plane update

UTC0021 - Copter update

UTC0108 - conference update

UTC0115 - GSoC

UTC0115 - close