Dev Call March 13, 2023

Development Team
1

Issues & Pull Requests

DroneCAN UAVCAN Cyphal

Plane

Copter

2

Attendee count (max): 21

UTC1101 - Syncing APMPlanner with ArduPilot Firmware sever · Issue #1280 · ArduPilot/apm_planner · GitHub

  • Releases done from github now for APMplanner?
  • Traditionally from firmware.ardupilot.org
  • Tridge is working with Bill to get stuff up on firmware
  • But tridge wasn’t aware releases were being done from github
  • The App Image is Linux-only?
    • Arne says Linux is the only system he can build
    • It builds on Windows
      • Could use CI to do it…
      • No demand for Windows?
  • Arne would like help releasing MacOSX once or twice a year
  • We’ll archive the old versions
  • Tridge didn’t have any luck starting it

UC1115 - CAN: support Himark CAN servos by tridge · Pull Request #22331 · ArduPilot/ardupilot · GitHub

  • DroneCAN sub-protocols
  • Peter and tridge need to discuss
  • Will discuss on DevCalEU

UC1117 - AP_Motors: move Thrust Linearization functions and variables to own class by IamPete1 · Pull Request #22787 · ArduPilot/ardupilot · GitHub

  • Move thrust linearization functions to their own class and variables
  • Need to be careful about what we merge as anything we put in will go into 4.4
  • We don’t want to do a rebase
  • Tridge is a bit wary of large changes ATM
  • Maybe this is 4.5?
  • Randy wants to make sure that different pressures and voltages do impact the thrust linearisation
  • Tridge is hoping to get next beta out this week
    • Doesn’t want big changes in right now

UTC1130 - Add and use a "bool read(c)" method to AP_HAL by peterbarker · Pull Request #22990 · ArduPilot/ardupilot · GitHub

  • Add and use a bool read(c) to HAL
  • Bump back to 4.5

UTC1132 - Add defines for AP_Generator backends by peterbarker · Pull Request #23049 · ArduPilot/ardupilot · GitHub

  • Generator defines
  • Merged!

UTC1134 - https://github.com/ArduPilot/ardupilot/pull/23062

  • Make AC_PRECLAND optional and put it in build_options.py
  • Peter will review

UTC1136 - Fix Failing Tests by lthall · Pull Request #23109 · ArduPilot/ardupilot · GitHub

  • Fixes Python checks in 4.3
  • Merged!

UTC1138 - Document some SITL parameters by peterbarker · Pull Request #23126 · ArduPilot/ardupilot · GitHub

  • Document some SITL parameters
  • Tridge is thinking about cleaning up things in SITL parameters
    • A 4.5 thing?
    • IMU parameter class, GPS parameter class, mag parameter class, airspeed parameter class
    • Parameter conversion?
      • Will possibly break MisisonPlanner integration
      • And SITL-on-hardware
  • Huge number of SIM parameters
  • A better structure is needed
  • merged!

UTC1143 - Abort PAYLOAD_PLACE landing aux function by Hwurzburg · Pull Request #23129 · ArduPilot/ardupilot · GitHub

  • Abort payload place aux function
  • Payload place in plane really wants scripting
  • Henry wants an aux switch to abort the landing
  • Should this be a “landing abort” instead?
    • If you’re in QRTL?
      • Require a do-land-start?
    • There’s a mavlink command to do the landing abort
  • When we have options for qrtl and qland we can re-use this switch option
  • This is good for now
  • Merged!
  • Fixed-wing glide-slope landings have an “Abort Landing” button - we should use that for vtol landing

UTC1153 - Yaw D Autotune by andyp1per · Pull Request #23131 · ArduPilot/ardupilot · GitHub

  • Yaw-d-autotune
  • Will automatically tune it
  • May be beneficial on thrust-vectored vehicles
  • Average user / wiki? Premature to recommend use of this in the general case
  • So no updates to wiki or advice at this point
  • Also very relevant for fixed-cant multirotors
    • E.g. callisto has a 3-degree cant
    • Some d-yaw does work well here

UTC0000 - Add second battery sensor to KakuteH7v2 by andyp1per · Pull Request #23136 · ArduPilot/ardupilot · GitHub

  • Adds second battery sensor on KakuteH7v2
  • Might have a bounds problem on the new default on MatekL431
  • This might change scaling on some boards?
    • Need to ensure defaults.parm and this volt-pin change make sense

UTC0006 - AP_Scripting: bump SCRIPTING_HEAP_SIZE for large RAM (>=1MB) devices by magicrub · Pull Request #23142 · ArduPilot/ardupilot · GitHub

  • Bump up heaping script size for large-RAM devices
  • including H7 boards
    • From 100kB to 200kB
  • Changes in memory layout in default Pixhawk6x bootloader meant we need to add the multi-heap feature to make large allocations work as it fragmented memory
  • Worst-case in scripting is where you’re on the boundary
    • Could we emit a high-watermark warning?
    • Bit hard to tell if you really are on the edge as garbage collection might be run more often
      • Move to manual garbage collection?
      • We do one form of gc at the end of each loop
  • Object-oriented style reduces memory usage dramatically
    • Have to use the right form for this
    • Check the structure of the aerobatics LUA script to see where it works well
  • MergeOnCIPass

UTC0017 - AP_HAL_ChibiOS: remove mavlink camera from minimized boards by peterbarker · Pull Request #23160 · ArduPilot/ardupilot · GitHub

  • Remove mavlink backend from minimized builds
  • There’s a thermal camera from FLIR which might trigger?
  • Maybe phase1?
  • We previously always broadcast to components
  • Now you have to specify a backend
  • Randy’s good with this going in
  • Merged!

UTC0024 - AP_Generator: rename fuel_remain_pct to fuel_remain_scale by peterbarker · Pull Request #23168 · ArduPilot/ardupilot · GitHub

  • Proportion?
  • Ratio?
  • Scale?
  • Airspeed is ratio
    • That’s a multiplier on the raw air pressure
  • Peter will come up with something and merge it

UTC0030 - Add and use option to compile SLCAN support out of code by peterbarker · Pull Request #23174 · ArduPilot/ardupilot · GitHub

  • Remove SLCAN from minimized builds
  • MergeOnCIPass
  • Can change CI to be Pixhawk1-1M-bdshot

UTC0033 - AP_Terrain: terrain offset max default to 30m by rmackay9 · Pull Request #23176 · ArduPilot/ardupilot · GitHub

  • Adjust maximum altitude deviation to 30m
  • There are arguments both way for this deviation number
  • 1 in 10 or 20 flights this was a problem
  • Cliff problem
    • Flying on beach with a cliff you can have trouble if your vehicle things it’s on top of the cliff rather than at the bottom
  • This is all about resolving zero-terrain-from-database and actual-zero-terrain-because-you-are-on-the-ground
  • Common case is user taking off from the ground
    • But 90m grid means you’re only accurate at intersections of thegrid lines
      • 1m precision vertically
  • What we have works well except where the clamping cuts in
    • So Randy wants to move the clamping
  • [11:46 AM]Peter Hall: Basicaly its not absolute terrain following, its local topology following.
  • 11:46 AM]Peter Hall: We follow the ground slope not the abs alt.
  • Grabbing GPS altitude at a single point in time is perhaps the problem?
  • Could also be flaws in the database
  • SRTM database is getting long in the tooth
  • Merged!

UTC0059 - Plane update

  • Report from a partner on 4.3 where the ailerons and elevator were winding up on the ground
    • Flight went just fine
  • Tridge wants to branch for 4.4 now
  • CubePilot want the 5-IMU-support stuff in
  • IMU parameters will be a little odd in 4.4 on some rare boards
  • 4.5 will be be IMU parameter rename

UTC0102 - Copter update

  • 4.3.5 is out!
  • Couple of new issues
    • NMEA output problems
    • Long-standing bug
    • PR is ready
  • yaapu telemetry not showing second battery
  • USD rangefinder on CAN was a config error
  • Thanks to Yuri Rage for tracing that one
  • 4.4 issue
    • Terrain alt clamping
    • New takeoff code in 4.4 has some issues
    • Takeoff command using relative is still susceptible to terrain database issues

UTC0107 - Fix CubeOrange build when HAL_MINIMIZE_FEATURES is set by peterbarker · Pull Request #23177 · ArduPilot/ardupilot · GitHub

  • Fix CubeOrange build when HAL_MINIMIZE_FEATURES
  • Merged!

UTC0109 - Rudder arming allowed in ALL modes · Issue #23181 · ArduPilot/ardupilot · GitHub

  • udder arming allowed in all modes
  • Bug?
  • Maybe disable stick-mixing instead?
  • Would be nice to get this in Copter
  • This will be disruptive to the users
  • Immediately after arming, squash yaw input channel to trim value until we see it within dz?
    • With a timeout?
  • Throttle-zero on quadplane is magic
    • There’s a bit which changes that
  • No yaw until you release it?
  • If your rc4 trim is off you might never get rudder back
    • Timeout with slew rate to fix that
  • Auto-rolling-takeoff state machine is stopped until rudder is seen as neutral
  • Wait_for_neutral needs to be reset at least
  • Could end up with full-throttle stick mixing accidentally with current formulation

UTC0132 - Plane: don't trigger RC failsafe until RC has been received for the first time by IamPete1 · Pull Request #23189 · ArduPilot/ardupilot · GitHub

  • Don’t RC-failsafe until we’ve seen an RC signal
  • This is a 4.5 thing
  • Forced-arming could be problematic
  • Arming could trigger an RTL which would be a problem?

UTC0144 - Tools:reserve KakuteH7-Wing board type by Hwurzburg · Pull Request #23196 · ArduPilot/ardupilot · GitHub

  • Reserve a board type for KakuteH7-Wing
  • Dibs!

UTC0146 - gitmodules: Bump up benchmark to v1.7.1 release by Ryanf55 · Pull Request #23197 · ArduPilot/ardupilot · GitHub

  • Benchmark
  • Maybe move up to 22.04 on CI?

UTC0146 - conference

  • Book your flights? Too late?
  • Please update attendance sheet (ping tridge!)
    • Need it to for food-planning
  • Make sure you’re on the conference chat on Discord

UTC0149 - close