Dev Call Sept 20, 2021

Issues & Pull Requests




Attendee count (max): 24

UTC2302 - How to advise user when DMA issue arises in CRSF · Issue #3674 · ArduPilot/ardupilot_wiki · GitHub

  • DMA warning is confusing users / no solution specified
  • Document it on the Wiki?
  • an arming warning?
  • Have we quantified the effect? Do we know how much of an issue it is?
    • Can we just not care?
  • 400kbaud and 100 kbaud
  • Detect symptom in prearm (high packet loss) and give an arming error?
    • 1 in 10 packets doesn’t matter
    • Past a certain level it starts to matter
    • Even at 400kbaud some boards won’t see a lot of loss even when not using DMA
  • Bumped to DevCallEU
  • Second problem
    • Reboot means baud rate negotiation is lost so we may not connect

UTC2306 - Add Linux shortcuts for MAVProxy and MAVExplorer by Hwurzburg · Pull Request #924 · ArduPilot/MAVProxy · GitHub

  • Shortcuts in Linux for MAVProxy and MAVExplorer
  • Merged!

UTC2306 -

  • Support multiple clients with a timeout
  • Permit UDP clients
  • Some dissension on timeouts and default behaviour for a single connection

UTC2317 - OSD Sidebars by squilter · Pull Request #16441 · ArduPilot/ardupilot · GitHub

UTC2322 -

  • Making weathervaning accessible to Copter
  • Properly tested?
  • So long as this has been tested it can be merged
  • Randy needs to approve this
  • Will be back on DevCallEU

UTC2325 - AP_Mission: fix infinite loop with nav commands that won't start by peterbarker · Pull Request #18726 · ArduPilot/ardupilot · GitHub

  • Infinite loop in Mission
  • We won’t allow more than 255 do commands in a row now
  • Mergeoncipass

UTC2329 - MS5525 driver fix conversion when temperature <20C or pressure negative by hendjoshsr71 · Pull Request #18363 · ArduPilot/ardupilot · GitHub

  • 32-bit SITL?
    • Board type added in later PR
  • Ms5525 driver ix for <20C

UTC2330 - Copter: do not enter airmode when armed with switch by IamPete1 · Pull Request #18478 · ArduPilot/ardupilot · GitHub

  • Don’t enter airmode when armed with switch
  • Param conversion from 41 to new parameter values
    • With airmode for one vehicle, without airmode for the other
  • All this so we don’t change behaviour
  • Deletes the q-option for airmode on arming
  • Merged!

UTC2336 - Generate board list by peterbarker · Pull Request #18502 · ArduPilot/ardupilot · GitHub

  • Generate board list
  • Peter and tridge will test after the call

UTC2338 -

  • Fix units in reporting via mavlink for quadplane
  • Merged!

UTC2340 -

  • New airspeed sensor for AP_Periph
  • Objection
    • Add_backend macro
    • Lack of discoverability
    • No call to function in C++
    • Generated by hwdef compiler
    • Same pattern we use for baro/mag/imu

UTC2342 -

  • Peter to look at this after the call and merge if happy

UTC2343 -

UTC2346 -

  • Add build sitl 32 option
  • Should we have a separate directory for 32 and 64 bit?
  • Tridge thinks not
  • Avoid combination of options
  • Will change to use same dir for both

UTC2353 -

  • Change loiter altitude control like fbwb
  • Currently we stick-mix but we will return to current altitude
  • This means we’ll change at new altitude
  • This needs to be predicated on stick-mixing
  • Inputs outside trim might be issues with this
  • New flight option bit to get old behaviour back
  • Why wouldn’t this be relevant in other flight modes like circle?

UTC0001 -

  • Check imu within calibrated range
  • Now with bonus prearm check
  • Arming temperature check as absolute but INS as a margin?
  • Interaction with INS temperature calibration?
    • Ability to turn off with hwdef
  • Heater temperature / ins temperature are different things?
    • Both looking at IMU temperature?
    • May not in the future?
  • Have to turn off arming check if you’ve calibrated your temperature sensor?
    • No, just set the margin to 0 to disable
  • We want temperatures to come calibrated on autopilots
  • If your IMU is too hot out in the field then you can’t really cool it?
  • Go back to just a heater temperature check
  • Heater + temperature calibrate?
    • Heater might get you into the calibration range
    • … but heaters also create interference with others sensors
    • So set min temperature to bottom of calibration range
  • We’ll remove the arming check for ins

UTC0018 -

  • Removes some dead code in yaw handling via GCS mavlink
  • Shouldn’t be checking for local frames
  • Randy will need to review and approve

UTC0019 -

  • A could happen but scripting can trigger this
  • Currently Can set a timeout which will never trigger if scripting is involved
  • Merged!

UTC0024 -

  • Gets rid of manual mode trim auto
  • Used to Trim plane in manual mode then trigger this to copy trims so fbwa works
  • For people used to flying a plane and trimming it using transmitter
  • Trim_auto triggered things automatically so once Plane was trimmed in manual
    • Locks in current rc input as trims
  • Paul usually does manual flight first
    • When transitioning traditional RC Plane he uses the feature we’re getting rid here
  • People are supposed to be using SERVO_AUTO_TRIM
  • Some nasty traps in this one
    • In case of fence breach we’d fiddle the parameter
      • But not for battery failsafe!
      • So if you had stick input when your vehicle failsafe we’d stick you with significant RC offsets
  • Remove TRIM_AUTO but add new RC_OPTION to trigger this function
    • Add roll/pitch rate check

UTC0039 -

  • Don’t add parameter mode when armed
  • Causes telem loss
  • Crsf-only change
  • Merged!

UTC0042 - Plane update

  • Plane: stable 4.1.0 issues list · Issue #15941 · ArduPilot/ardupilot · GitHub

  • Some EKF issues have come up

  • High vibration check in EKF would continuously reset height and velocity state every second

    • So EKF wouldn’t get too far away from measurements
    • But doing it every 100ms instead of every 1s
  • Velocity reset isn’t happening correctly so current vibration failsafe test is failing

    • Bad vel estimate causes climb
  • So if you don’t know your vv should you return the GPS velocity?

  • Copter might end up with more robust vibration handling once this is in

  • Need to check EK2?

    • Doesn’t have clip indicators but do have height timeouts
  • Changing accel clip limit on rc switch useful for testing this

    • 16G to 2G
    • Clipping on very vehicle
  • High vibration means lower gains through controller too

  • Then there’s also the fly-away code which removes accel pid code from loop (just used velocity control)

    • 4* reduction when using DCM too (quadplane)
      • Randy thinks this is too big
    • Parameter for backoff?
  • 200ms delay when falling back to GPS data

  • Wind state init fixes


  • Still considering when to do stable release

UTC0101 -

  • Copter update
  • Yaw change on guided mode takeoff
  • Previously would not change yaw
  • Now might turn in direction depending on perceived speed
  • Probably only one more RC before push out stable
  • At least a week out
  • No unexplained crashes in forever
  • Reject commands which specify a force

UTC0105 - tradheli update

  • Autotune!
  • Some good alpha-level testing results from Steve
    • Success!
  • Ready for review now

UTC0108 - Rover update

  • 4.1 is out!
  • Risks are lower compared to other vehicles
  • Point releases for sure but nothing yet
  • Scurves next!

UTC0109 -

  • Log groundspeed undershoot
  • Merged!

UTC0112 -

  • Makes unhappy message repeat
  • Merged!

UTC0114 -

  • Enum cleanup
  • Merged

UTC0117 -

  • ZealotF427 RC input issues
  • Still under discussion

UTC0119 -

  • What impact will this have?
  • Floats on servo inputs!
  • Fixes 1% throttle stuff
  • Fixes 1% slew rate
  • Needs flight testing
  • Might affect tri frame
  • Nice intermediate thing
  • Doesn’t change last_pwm array
    • So won’t fix all throttle problems
  • Scaled to narrow range issue
  • Might help on rovers?
  • SITL testing should be valid as HAL layer is still integer
  • Nice patch!

UTC0125 -

  • Move into MOT if still required
  • RC_SPEED thing

UTC0128 - ArduPilot foundation now exists

  • Bank account coming

UTC0129 - close


I am one of those “grumpy old men” who use TRIM_AUTO.
For conventional fixed wing planes I still believe in a full manual successful first flight and then add the auto pilot functionality later on.
I am also aware of the dangers of TRIM_AUTO as implemented today. Changing the flight mode at the same time with the sticks off-center will give a “bumpy ride” … Don’t ask me how I know.

I know it’s different from the recommended procedure for the first flight in the wiki. It works for me.
You may ask why I don’t use the recommended SERVO_AUTO_TRIM. From my point of view it hides any tendency of a mechanical bad setup until it runs out of available servo trim range.
From my point of view a well made mechanical setup and properly trimmed plane is the base of all reliable working planes with or without auto pilot.

PS: I would be happy with a similar functionality done by a new RC_OPTION too.

1 Like

Thanks for the feedback, the change has now been merged, the TRIM_AUTO param is gone but the functionality remains in the form of a new RCx_OPTION, “155: set roll pitch and yaw trim to current servo and RC”. The only effects at the moment, but it will work its way into a stable release eventually.


Hi @iampete,
I have to say Thank you for the new RCx_OPTION.