Dev Call Dec 20, 2021

Issues & Pull Requests




1 Like

Attendee count (max): 22

UTC1102 - Report CRSF link quality in place of RSSI by julesgilson · Pull Request #19042 · ArduPilot/ardupilot · GitHub

  • use link quality in place of RSSI in CRSF
  • Thanks to julesgilson for first contribution

UTC1104 - AP_OSD: fix code causing errors while compiling out OSD by shiv-tyagi · Pull Request #19146 · ArduPilot/ardupilot · GitHub

  • Q_option to force fw transition on fail instead of qland
  • Test flown today
  • CI broken right now
    • Enforcing holiday break on developers
  • Still does qland if option not set
  • MergeOnCIPass

UTC1108 - AP_Scripting: add support for getting fw version by IamPete1 · Pull Request #19397 · ArduPilot/ardupilot · GitHub

  • Add support for ArduPilot version numbers in scripts
  • MdB requested individual fields which has been done, but review still outstanding
  • Can be merged if MdB is happy

UTC1111 - Plane: tailsitter: add gains to scale control surface vs motors by IamPete1 · Pull Request #19463 · ArduPilot/ardupilot · GitHub

  • Scaling fix for copter tailsitters
  • Pitch with motors or elevators? This allows you to set gains separately to surfaces vs motors is adjustable
  • Current default of 1 means no change
  • Wikineeded
  • Conflicting PIDs may cause oscillation
    • This is the sensible approach
    • SLT quadplanes have the two-pid problem
      • Desired rates are linked to help avoid it
  • Merged

UTC1118 - Gyroflow logging by IamPete1 · Pull Request #19531 · ArduPilot/ardupilot · GitHub

  • Adds dedicated message for gyroflow video stabilization
  • Our IMU data is filtered….
    • Get_gyro and get_accell
    • But not bias-corrected
  • 300 bytes of flash
  • Off by default and a custom build server thing?
  • Scripting?
  • Needs to be full-rate logging
  • Quaternion isn’t used ATM but PH threw it in anyway
  • Need RATE and IMU and accel biases
    • RATE message isn’t relevant in plane fixed wing (most of the fields)
  • Does the thing have legs?
    • Randy worked on the DPTH message
      • Worked quite well
  • There’s a PR against gyroflow
  • This needs to correct for center-of-focus on the camera
  • VSTB rather than GFLO?
  • Randy likes domain-specific methods
    • Makes it easy for users to use the feature
  • Feedback from gyroflow labs would be nice

UTC1136 - Pr fix multiple includes by Jaaaky · Pull Request #19538 · ArduPilot/ardupilot · GitHub

  • Fix multiple includes
  • Merged!

UTC1138 - AP_BattMonitor: Remove double cast in mah calculation by arBalasquide · Pull Request #19540 · ArduPilot/ardupilot · GitHub

  • Needs copyright fixed
  • Can be merged out of call once copyright issue is fixed

UTC1140 - Reduce number of log IDs allocated to vehicles, rearrange by peterbarker · Pull Request #19562 · ArduPilot/ardupilot · GitHub

  • Adjusts log ids to be more sensibly allocated
  • Merged

UTC1142 - RC_Channel: added RC_SWITCH_TYPE parameter by tridge · Pull Request #19568 · ArduPilot/ardupilot · GitHub

  • Adds rc switch type to allow for radiomaster tx16s push-buttons work nicely
    • Surprisingly tricky right now
  • Use rc channel options instead
  • Massive GCS changes required
  • Even this one would need the GCS to be fixed
  • GCS needs to know the thresholds to get interfaces correct
  • Two-channel-mixing in ArduPilot would be quite nice

UTC1155 - RC: fixed serial protocol detection bug by tridge · Pull Request #19576 · ArduPilot/ardupilot · GitHub

  • Fixes regression
  • Crsf change for no serial port
  • Find_serial changes options
  • New have_serial
  • AP_RCProtocol bug too
    • If redetection set and you detected via bytes (on uart) it would never re detect correctly by not re-searching
  • Serialn_option is a relatively new thing and we’re still bedding it in
    • Find_serial, get uart, call serial manager to set options sounded a bit clunky
    • So find_serial set options for you
  • Doesn’t just affect pixhawk4 mini
  • How did this bug get out there?
    • Beta testing didn’t catch it?
    • Most people don’t run the betas
    • 80-odd flight controllers in many configurations!
      • Multiply probabilities you end up with small numbers
    • Timer-capture rc input doesn’t have this
    • We can do all protocols on both ports
    • Rcin port is serial4 uart on pixhawk4 minit
    • Most other things are timer-capture
  • 4.1 there are three boards which are affected
  • Merged, will be in next Plane stable too

UTC0008 - AP_SmartAudio: fix a bug that would starve the lower priority thread by MingchenZhang · Pull Request #19579 · ArduPilot/ardupilot · GitHub

  • Smartaudio fix for too much CPU
    • May starve e.g. scripts
  • Delay 100 is ineffective
  • Should be backported
  • merged

UTC0010 - AP_BoardConfig: disable STLink debug pins by default by tridge · Pull Request #19559 · ArduPilot/ardupilot · GitHub

  • Backport for st-link debug pins
  • ESD (“stray electron”) could cause flight controller to reboot
  • Pressing debug connector on board would cause reboot!
  • Disable by default
  • BRD_OPTIONS to enable the debug pins
  • Why not on AP_Periph?
    • Consequences are less severe on Periph
    • It probably should be but too much bother right now
  • Pull-ups PR also coming soon
    • Our config floats these pins
    • Sid makes them asserted one way or the other
  • We do have an option to assert lines to pull things down
    • Some boards are using that
    • Modalai pixracer pro pixracer periph and one more board
    • ST electronics say we should pull low by default
      • … but current draw….
      • Defaults here is one-up, one-down….
  • mRo don’t even make the board in question but found the problem for us
  • Will need to offer partners running 4.0 a patch
  • This will go into the next Copter stable rc
    • Rc2
    • Ph4-mini as well
    • And crsf patch

UTC0025 - Plane update

  • New 4.1.6 beta coming soon with fixes discussed on this call in it
  • Synthetic airspeed bug?
    • Important but hasn’t been looked at yet

UTC0026 - Copter: 4.1.0 issues list · Issue #16478 · ArduPilot/ardupilot · GitHub

  • Copter update
  • Rc2 today probably
  • Some more issues have come up this week
  • Set-position-target-local-ned ignore yaw/ignore-yaw-rate might be interpreted incorrectly?
  • Yaapu not working on omnibusf4?
  • Copter-4.1 not working if using a fresh repository
    • Tridge has fixed this one!
  • Rc2 and plane 4.1.6
    • Double-rebase showed up some stuff
      • Invensesense sensors
      • Covariance row reset
      • Autotune rate limits
      • They’re all good!

UTC0029 - Rover-4.1.3-rc1 released for beta testing - #2

  • balance -bot broken in acro mode?
    • Regression from 4.0 to 4.1?
  • D-shot still to be looked at
  • 4.1.3 beta test is out there

UTC0032 - Heli update

  • In-flight motor shutdown
    • Met criteria for landing
    • H-collective-mid land-complete-maybe
    • Did stuff as-designed…
    • H-collective-mid at zero is recommended
      • Shouldn’t be descending at zero
      • Max-was 140pwm above that
      • Min was 140pwm below that
      • Typical for helicopters
      • Symmetric, 3d setup
      • Collective was all the way down?
        • H-collective-mid was probably not set up correctly
    • Tridge doesn’t like the detection being payload dependent
    • Heli might want a much more extended time before disarm is done?
    • Rising air can fool vehicles
      • Slope lift on side of hill
      • Extending the time doesn’t stop it from happening
        • Transients mean you won’t “land”
    • Mid isn’t supposed to be half-way between the two endpoints
    • Maybe have a regime people go through to determine things?
      • E.g. do a 1m/s descent rate, check the hcol-mid
  • Big clear warnings in Wiki that this is a dangerous parameter and one you should pay attention to
  • Option to disable the detection regime and have an option for force-landed
    • Use a switch to disarm?
  • Two-position switch to disable detector
  • Auto modes won’t detect landing if in auto and not in a landing sequence
    • Disarm happened at 50m

UTC0052 - –out for waf not working?

  • Tridge tested it as working
    • –out requires an absolute path

UTC0059 - happy holidays to all

  • Thanks to Craig for running these calls

UTC0059 - close

1 Like