Dev Call Feb 14, 2022

Issues & Pull Requests




Attendee count (max): 21

UTC1100 -

  • Allow conversion of parameters into scripts
  • [10:05 AM] Peter Barker: I really don’t think removing working and valid C++ code is appropriate. At the very least it will make the feature inaccessible on 1MB boards. The code mentioned can already be compiled out via hwdef.
  • [10:07 AM] Josh Henderson: Agree.
  • [10:07 AM] Peter Hall: very small set of people using the custom build server to get those features on 1M boards.
  • [10:07 AM] Peter Hall: There going to lose them by default at some point anyway if they have not already.
  • merged!

UTC1103 -

  • pause/continue concept with SCurve navigation
  • How should we pause/unpause as we exit submodes
    • If you’re in position amode ang get paused, and then get told to do something with the attitude control, should you unpause?
  • Multiple-command case is a corner case already
  • Leonard thinks we should just stay paused regardless of sub-mode change
    • Also fixes race conditions with queued commands
  • So all submodes it will stay paused and if the flight mode changes we unpause

UTC1111 -

  • This was old C code which got moved to .cpp
  • Full cleanup wasn’t done
  • So this is a reasonable cleanup
  • Merged
  • This is about type safety
    • Nullptr is a pointer, NULL is not

UTC1113 -

  • Stops using guided_WP_loc
  • Need to look at the avoidance thing
  • Peter will test a lot and work with Tom to get this in, will bring back

UTC1124 -

  • Tridge doesn’t really like the attribute forcing people to check results
  • We’ll merge this
  • But we need a follow-up to get rid of IGNORE_RETURN

UTC0135 -

  • Navigator CI tests
  • merged!

UTC0138 -

  • @SYS/timers.txt support
    • Timer information via @SYS
  • Need to have Andy around for this one
  • Tridge Played around on cuav LED and it was good
  • @SYS files are awesome in log files
  • Flash cost for having the diagnostics…
    • Having all the diagnostic information all the time is useful…
  • Probably put these diagnostics onto periph nodes
    • Add a SYS_DEBUG parameter enum 0 don’t 1 dma 2 uarts 3 whatever
    • As soon as it sees the parameter change it dumps once to CAN GUI
    • Use FTP to get data?
    • A pair of CAN messages a CAN node can send, one to define a dataflash log message, one to actually log against it?
  • Getting rid of -bdshot boards?
    • Prereq is dynamic DMA allocation stuff
    • It is possible, just difficult
      • Alt-configs?
    • [10:49 AM] Peter Hall: Last time I looked you can do it now on some boards.
    • [10:49 AM] Peter Hall: The timers happen to come out the same.
    • Comes down to who can share DMA channels
      • SPI can share
      • ADC can’t
      • Third category which wants dedicated access but it gets the dedicated access at run-time rather than compile-time
    • Extensions to BRD_ALT_CONFIG to be able to switch between tables?
      • We’d still need partial DMA assignment
      • We could also say we just won’t do DMA for some things

UTC0152 -

  • Having the locals is useful for debug
  • We’d accept either way but we really don’t want to move between the two
  • closed

UTC0156 -

  • Merged
  • Documentation fixes around gsf

UTC0157 -

  • Get_gryo_latest should include pitch trim
  • Correct use of pitch trim from ahrs view
    • Affects quadplanes if you’ve fiddled some parameters
  • Rot_view includes pitch trim
    • Rotation is one of the enums
  • Rotate was cheap, the “*” is expensive
  • once/loop call
  • In Copter we run the rate controllers before doing the AHRS update
  • Get_gyro_latest call is actually not useful?
    • Just nuke get_gyro_latest?!
    • Remove it completely in a future PR?
  • Peter Hall will put a test in to make sure that the numbers are just the same
  • EKF3 already adds the drift (most of the time)
  • We’re not logging what we’re using right now
  • View needs and has cached gyro value including pitch trim

UTC0008 -

  • frontend/backend split for temperature sensors
  • Can be compiled out
  • TEMP1_ TEMP2_ or TEMP_ TEMP2_
  • Former not latter
  • Temperature sensor params class should make that easy enough
  • How do we connect this to other things?
    • You’ve got 5 temperatures - now what?
    • Tom is overriding ESC temperatures with other temperature sensors
  • H7 we can get microcontroller temperature
    • Logged in POWR
    • Should be a temperature backend

UTC0019 -

  • VSTB to use latest gyro
  • The entire call should go away, so this is DelayMerge pending Closed not needed
    • See earlier discussion in this DevCall

UTC0020 -

  • Range-check the PWM input
  • Should use pre-existing values
  • 900 sometimes come in as a legitimate value
    • Tridge has seen it
  • 800 might be a better idea
  • Can use the same value in motor stuff
  • Consistency is good!

UTC0023 -

  • Spracing documentation and add to CI
  • Should we knock out one of the builds to reduce CI times?
  • Merged!

UTC0033 -

  • GSoC students are doing a great job this year!
  • Fixes our build when using AVSUAS.xml
  • MergeOnCIPass

UTC0036 - Plane update

  • Mostly same as Copter
  • Aerobatics update
  • Need to do build system fixes for running multiple betas

UTC0037 -

  • Copter update
  • Beta is going well
    • I-term buildup fix worked well
    • Ibase frame gap fix in there as well
  • 4.2 afterwards!
  • Leonard’s PRs to come in before we beta

UTC0042 - CAN-over-mavlink!

  • Filtering is coming!
  • Along with a mavlink fix
  • Slcan can go away
  • Firmware updates can be updated on a CAN GPS over a slcan radio
    • 5 or 10 minutes to update it….
    • Slow link….
    • Updating one ESC at a time will be fun on an octocopter
  • Boat at sea on a 2 month journey
  • Mavlink-over-dronecan should also be revisited

UTC0046 - yaw reset bugfix on transitions for stable?

  • H-wing flipping upside down as it transitioned
    • Cause was found ages ago
    • Yaw reset using euler rates but because tailsitters earth-frame yaw rate becomes body-frame roll this doesn’t go well
    • Just stops resetting rates on tailsitters
  • Custom controllers for transition on fancier quadplanes are coming to make things cleaner

UTC0050 - ship landing for Copter?

  • C++ code or LUA?
    • Two hooks needed to use LUA
      • Set_velocity_match
        • Gives a base ground speed reference frame
          • Ship’s velocity
      • A way of updating the target position during the landing
        • Both in RTL land approach and descent
        • Switch to guided and you can do it in the lua script already…
    • Twitching on multirotors as they’re more responsive
    • Could probably use updated guided stuff on Plane instead…
    • Takeoff and landing and failsafes in lua scripts would need to be done for aborts and whatnot
      • LUA can kind-of-sort-of handle failsafes by noticing they’ve happened and re-taking control
  • This is something Leonard would love to get done

UTC0058 - anything else?

  • WSL2?
  • Guided-mode radius?
  • Canfd support?
    • Not in master yet
    • Probably a 4.3 thing
  • Gps-in for periphs
    • Why gps?
      • Why not EKF estimate?
  • If things don’t get in then we really shouldn’t worry as we’re aiming for bi-annual releases now!
  • Volz servo feedback thing
    • Dsdl and ardupilot PRs in place

UTC0105 - Rover update

  • Still working on trying to get SCurves in
    • Testers wanted!
    • Four or 5 people have popped up to help, but not sure SCurves will be in in time
    • Several little issues popped up in terms of limits in Rover
      • Spate of little fixes have gone into Rover for limit handling
      • Edge cases fixed up
      • Copter handles limits well
        • Rover not quite so much

UTC0108 - close