Servers by jDrones

Dev Call Aug 30, 2021

Issues & Pull requests





Attendee count (max): 20

UTC2302 -

  • Good first issue was done
  • Logs internal errors very close to when the happen
  • Can be merged when

UTC2304 -

  • New board
  • Bootloader definition missing?
  • wants a README

UTC2306 -

  • Correct field descriptions on PSC
  • Merged

UTC2309 -

  • Check all GPSs for fix and health
  • Could make blended GPS an actual backend

UTC2310 -

  • Align yaw to gps yaw
  • Not really gps yaw this is gsf yaw
  • Use yawsource enumeration?
    • Every backend would need to know about that
  • Tested

UTC2320 -

  • Make terrain margin a float…
  • Changes recent stuff we added
  • Not in new stables
  • Max needed to use old incorrect terrain data across entire globe is larger than what fits in an int16
  • Change to use metres instead
  • Changing units of a parameter isn’t something we like to do
  • But it has only been in very briefly
  • Also made default values a different value and a wider allowed range

UTC2325 -

  • Rst emit for latex pdf

UTC2328 -

  • Now have a 1-d model
  • C++ library for JSON example
  • Peter to check this over and then merge

UTC2329 -

  • Pitch_cd_flight option to for GCS and OSD to include or exclude Plane’s pitch-trim from the reported numbers

UTC2331 -

  • Fix compiler warning
  • Iteration to initialise seems extreme
  • Cast _v to void* instead?
    • We’ve done that elsewhere
  • Which compiler?

UTC2334 -

  • Fix corner case in new auto-rtl mode
  • Merged!
  • New issue to be created on having an “ok_to_enter” callback

UTC2340 -

  • Move taking of semaphore
  • Should make OSD and crsf have more time
  • Can be merged after it has been flight-tested and log looked at

UTC2344 -

  • Eliminate get_location
  • We’ll rename get_position to get_location later
  • merged!

UTC2346 -

  • Couldn’t get past the “neutral values” RC check
  • Plane: remove use of recompute_pwm_no_deadzone in training mode

UTC2348 -

UTC2356 -

  • Need to set the init_ms
  • Breaks initial transition into qloiter and rampdown after being in assisted flight
  • Significant issue - backport to 4.1
  • Transitions to qloiter and/or assist second one might be dodgy
  • Probably won’t crash you but will look untidy

UTC2359 -

  • Remove safe-pwm from Rover and then from everywhere else it had its fingers
  • Need to redo the iofirmware
    • Probably won’t commit new builds just for this
  • IOMC message

UTC2359 -

  • Stop having to modify things other than hwdef when adding boards

UTC2308 -

  • Zeroing accumulators without holding locks!
  • merged!

UTC2309 -

  • Speed improvements for USB on F7
  • Merged!

UTC2312 -

  • Save vast amounts of flash by using data structures
  • Mechanically done
  • Some weird ones
  • Several little things to fix up but can then be merged

UTC2323 -

  • Simplify get_wp_destination_loc
  • Needs more testing

UTC2327 -

  • Fix comments
  • Merged!

UTC2331 -

  • Polarity on the IOMCU pin?
  • Make sure heater is left off if disabled

UTC2332 -

  • Fixes wobbly horizon on ahrs cal
  • One loop without assisting as well…
  • merged!

UTC2332 -

  • Airmode in forward flight for surface-less tailsitters
  • Qassist is always active except at zero throttle
  • This adds in airmode which is active at zero-throttle
  • Need to adjust the defaults file for the Swan-K1
  • No need to set the throttle-min any more
  • Will allow for “gliding” in these vehicles
  • Needs a flight test on a surface-ful tailsitter
  • May allow for transition fixes
  • Z-controller init issues?
    • Stop using z-controller initially
    • Run q-stab code with low throttle until we hit some threshold
    • Then init the z controller
    • Another state in the back-transition?
      • PH is trying to avoid adding them….
  • Tailsitters are weird

UTC2342 -

  • Plane update
  • Another week, another beta
  • In-sync with Copter
  • Pretty happy with the way things are going in beta series
  • Big change is 60ms scheduling delay fix
    • More important than expected
    • Bug can actually happen at any time
      • Not just on timer wrap
      • 32-bit would be a watchdog reset

UTC2350 -

  • Copter update
  • Generally a good story
  • Slight uptick in issues over last week
  • Dijkstras is not working in guided in position control mode
    • Guided mode doesn’t use wpnav any more so can’t use dijkstras
      • Can use simple avoidance
      • Working with Leonard on this one
  • Crsf and gps not working well together?
    • Extraordinarily bad crystal
    • 6600ppm crystal error
    • Capacitor error?
    • GPS-time vs system-time advance showed it quite clearly
  • Channel 6 tuning of wp speed is busted
    • Constantly changes target speed, s-curves don’t like that
    • Change things so we don’t poke too often?
    • Lots and lots of calculations to redo the s-curve
      • Currently do it at 3Hz?
      • Slow that down and require it to be stable for 1 second before making the change?
  • Pitch control in loiter seems significantly changed
    • We have logs for the problem?
    • Need PID data!
  • Surface tracking with downward-facing lidar is laggy
    • Blocker for release?
      • Do we need to get 4.1 out the door?
        • So make this a non-blocker?
        • There are bad bugs that are fixed in 4.1 and we’re not back-porting them

UTC0004 -

  • Everyone seems to be happy with this release candidate
  • Bendy-ruler is working less-well?

UTC0007 - GSoC update

  • All done!
  • Three successful GSoC students!
  • Rishabh’s blog post series is coming
    • Next week or so

UTC0008 - off the floor

  • Lincoln
    • Flight control software from scratch? In four months?
    • Bit of history of ArduPilot
    • Expectations have changes for autopilot software
      • Older videos are laughable now-a days
      • People now complain if you’re off by more than cm
      • Could probably put a sketch on an Arduino in 4 months which got off the ground, but why would you?
    • Freertos?
      • We use that on the esp32
    • Research channel!
    • [11:15 AM] Josh Henderson: also trying to do both large hardware & large software development is way too much… for a small team even with baseline software that is stable.

UTC0017 - close

Servers by jDrones