Dev Call Aug 27, 2018 2300 UTC

Issues & Pull Requests


  • Update





Rover Boat

  • Sailboat
1 Like

Attendee count (max): 22

UTC2305 -

  • Behaviour change for loiter mode for tradheli
  • When disarmed or inflight or autorotation using motor interlock disabled
  • Allow pilot to get manual control from loiter mode
  • Navigation controller active as aircraft spooling down
  • Behaviour changes coming in 3.7 to do with spool states
  • These are smaller changes for 3.6
  • Interlock disabled and aircraft disarmed sets position to current position of aircraft
  • Wasn’t this supposed to be handed by the don’t-navigate-near-ground?
    • Might have been broken
    • Soften-for-landing?
    • Land-complete-maybe is a slightly easier thing to achieve
  • Backportable to 3.5?
    • Probably
    • 3.6 is close
  • merged!

UTC2314 -

  • Random intermittent faults with safety switch
  • Brd_safety_mask parameter when it should be allowing channels it is not
  • Tlog required
  • When you reboot it sets safety on when you reboot
    • Communication with iomcu may not be reliable after it reboots
  • MdB: Nates I think is because we don’t have a mixer…
  • Safety mask is a mixer feature and we don’t do mixers on ChibiOS?
    • No, it’s not a mixer feature
    • Controlled by the FMU, not by rc on IO
    • Shouldn’t need mixer
    • [9:18 AM] To Weekly devcall: Also, if it’s an intermittent failure it’s not a good theory?
  • Tridge has to concentrate on OBC stuff
  • Two incidents
    • Reboot-in-flight issue meant he didn’t have control
    • On reboot no surface control
    • Works on NuttX
      • No px4io mixer support in ChibiOS
    • MdB: also kinda curious how it rebooted IE electrical things probably take out IO processor anyways
    • [9:21 AM] (Channel) MdB: So you still wouldnt have servos…
    • [9:21 AM] (Channel) MdB: (until we got far enough in the cycle)
  • Rebooting in flight isn’t really supported
  • Doing the mixers is on the todo list
  • Need to find a mechanism for reproducing the intermittent one
    • Or at least a pattern

UTC2314 -

  • Meant to be NFC
  • Grew a bit of FC
  • Reduce code duplication and calls to _start_cmd
  • Fixes outstanding bug where we’d get into an infinite loop
    • Set-current-cmd and lots of do-cmds
  • Fixes mission restart related to jump-to-landing
    • You will get naks if you mavlink set-cmd to do-cmds that don’t have a nav after them
  • Removes code duplication
    • Removed code didn’t have the protections in it
  • Lots of SITL testing done
  • Cleanup prefacing more work to come
  • Should we be using 0 as a flag value here?
  • [9:29 AM] To Weekly devcall: I considered suggesting adding a parameter to indicate whether you should use the index - but decided what MdB had was as good.
  • Tridge is going to do some testing
    • [later] merged!

UTC2331 -

  • 3 questions.
  • Use godbolt to test the std::swap
  • Not being pure-virtual is OK (already the case in HAL)
  • MdB: we do have a bswap set of helpers already as well for swapping (all defer to builtin though)
    • libraries/AP_Common/missing/byteswap.h
  • No screams of pain, we can move on!

UTC2334 - stick-mixing and rudder

  • Adds extra version of plane flight options to only do mixing in manual stabilize or acro
  • Removes caching of input
  • Doesn’t change default behaviour
  • Makes no-stick-mixing mean exactly that
  • Diff makes things a little non-obvious
  • merged!

UTC2340 -

  • Relative to what?
    • I.e. cumulative or not?
  • Randy thinks it should be relative to baseline
  • Tridge thinks it should be relative to current speed
  • Would be based on target airspeed
  • Add an enumeration value?
  • Lost-packet problem
  • Peter to do a PR for discussion based on it being a non-cumulative value

UTC2354 - math float-equals

  • No objections
  • Merged

UTC2355 - uavcan refactor

UTC2358 -

  • Frsky telem
    • IMU Temperature
  • Was too complicated last week
  • Brought out a variable
  • Still too complicated
  • Could fix it ourselves
  • Too many ternary operators
  • Has a merge commit in it
  • Needs a squash and rebase

UTC2358 -

  • Tridge couldn’t fault this
  • Nate’s not on call
  • Needs a rebase

UTC0002 - DroneCore update, Peter

  • Similar to DroneKit
  • Tested. Required some changes, present in Peter’s branch
  • Required two PRs:

UTC0010 - redundancy

UTC0012 - streaming updates

UTC0016 - Plane refactoring

  • Want to get Tom’s stuff in
  • Both Randy and tridge are keen

UTC0016 - Plane update

  • Log download will be faster on USB in latest beta
  • 3.9.1 will go out shortly
  • Safety fix in in last week
    • Reboot causing vehicle movement
    • Trim settings
  • More test flights required
  • 3.9.2 in next few weeks
    • More boards
      • Hwdef.dat gives some surety
  • Three new boards in last week
    • Drotek p3 pro, speedy, ?
  • S0up active ATM
  • Rescue mode in betaflight?
    • Rtl mode in Copter without a compass
    • Yaw determination without compass
    • Creates fly-forward assumption
      • Kind of like drift mode
  • Outstanding bug
    • Only boot on battery power and not on USB?
    • Tridge spent some time chasing it yesterday
    • Bootloader issue?

UTC0023 - Copter - new rc out soon

  • Reasonably close to 3.6 official release
    • Few weeks
  • Tower issue
    • 3dr usb id problem
    • Forcing usb strings in hwdef.dat makes it work
  • Need to check if rc calibration is working on USB
    • Mavlink2 vs mavlink1
    • Probably MP is looking for the wrong one
  • PR for fixing twitch related to new loiter code

UTC0028 - Rover

  • Sailing code fits OK so far
    • No tacking so far
    • Coming along well
    • Interesting discussion and papers going on
    • Iampete
    • Need to incorporate roll
    • As much sheet on while remaining level
    • Airflow over sail
    • Windvane support coming
    • Randy to test in next week
  • Efforts continue on getting non-gps navigation using lidar and ROS
  • Hectorslam does a terrible job on a bumpy surface
    • Replace with google’s cartographer?
  • Obstacle-distance message support coming

UTC0043 - CAN testing with Nate

  • Will try to look at it tonight

UTC0043 - quick PR to go in to fix builds

UTC0044 - CI and caching issues

  • Both systems have problems
  • OK ATM

UTC0046 - funding update

  • Partner-sign-up for August is looking good!
  • Three others
    • Thanks to Hex!
  • Over 40 partners at the moment?!
  • Tridge’s Believer arrived
  • Tom’s arrived
  • Paul’s was damaged
    • Will try to get a replacement wing
  • Randy’s volunteered to write the Wiki page

UTC0050 - contribution of the month

  • Not merged yet
  • Tridge is considering nominating s0up for board ports

Thanks to Conor and others for help on the notes this week - Peter