Dev Call Jan 11, 2021

Issues & Pull Requests




1 Like

Attendee count (max): 22

UTC1100 -

  • General agreement that moving things into libraries is a good idea
  • Can be merged when Peter’s tested the parameter metadata output

UTC1107 -

  • Moving messages definitions into AP_AHRS
  • Want CAN GUI tools to pull documentation stuff

UTC1115 -

  • What frame should be used for AVD_F_ALT_MIN parameter?
  • Is ADSB avoidance safe for use for Copters?
  • Should clarify the parameters is AMSL not relative
  • Can’t change the meaning of a parameter like this as existing users may be using existing behaviour
  • Different flight levels mean different standards
  • QNH for moderate flight levels
  • If people want the flexibility of using relative then it can be added later
  • We have a similar parameter for RTL_ALT_FRAME

UTC1120 -

  • Merged
  • Fix units

UTC1122 -

  • Merged
  • “DCM” vs “EKF0” text fix

UTC1123 -

  • In RC failsafe have mount go to neutral position (or retract)
  • Naming?
    • Randy’s not a huge fan of the name
  • Change parameter name
    • Parameter name and internal variable name
  • Can be merged once Peter’s happy
    • Make it pure virtual?
      • Try it and see what breaks

UTC1127 -

  • PRIu64 etc etc causing issues
  • People find them hard to read
  • KDEcan isn’t working right now?
    • Tom’s shipped a device to Sid to sort it…
  • Merged!

UTC1139 -

  • Tidies rudder arming
  • Randy to review and merge when he’s happy

UTC1142 -

UTC1144 -

  • Sid’s checked this was good with dirdiff across the output to the script
    • Configure-all and then dirdiff
  • merged

UTC1145 -

  • Merged
  • Ftp upload of parameters?
    • Transactional?
  • Still no FTP download of parameters on FTP?
  • Wider problem with crossfire, this will help
    • James is working withthe crossfire people on the problems

UTC1149 -

  • Solo roll control
  • Needs a quick code review now?
    • Need to get Buzz or Matt or someone to test it instead
  • Simulator may not simulate roll

UTC1151 -

  • Correcting pin initialisation for some ESCs
  • ESCs getting confused?
    • Not initialising
    • Do we need to do this elsewhere?
    • Bootloader not rebuilt as part of this PR?
    • Commit message needs help
    • Tridge will work on this a bit
  • Others may have seen this…
    • Tom’s doing inheritance for board and bootloader to avoid duplication
  • Reset state of microntroller might still stuff things
    • PhilipK is adding an enable pin with a hold-down to avoid motor spin on boot
  • Tridge will merge after testing

UTC0001 -

  • Don’t have to init the submodules on checkout
  • Peter to test
  • We have a configure option to not play with submodules
    • –no-submodule-update

UTC0004 -

  • Subclasses for heli autotune
  • Statistical test on Callisto
    • Consistent results except D term in pitch axis
  • Tried to keep this working as it was when it was one library
  • Randy will look at this

UTC0009 -

  • More parameter reset issues
  • Post-bootloader-update
  • PX4 project is also getting parameter resets on CubeOrange
  • H7-based rather than specific to Orange, probably
  • We now have a set of backup parameter things!
    • Thanks Giacomo!
    • Tricky to analyze
    • Copter 4.0 has the logging of the last 100
      • Much better chance of getting data around it

UTC0016 -

  • Fixed compilation error due to macro conflicts
  • Merged!

UTC0017 -

  • Home vs origin for offsets?
    • Relative to what?
  • Rotation?
  • Reinterpreting a bit in everyone’s storage?
    • Tridge is worried we might consider already-stored missions as local
    • We need to test missions which have been stored with older versions of the code
    • Especially if crosstrack is set
  • Needs to be tested in both STIL and on stm32 and on linux boards
  • Throw away current method of storing items and put them into a flexible structure would be nice
    • Need a partner to sponsor that one
    • Maybe just on very-large-memory boards
  • A valid PR we would like to support, but this needs much more testing

UTC0035 - Plane updates

  • beta is out 4.0.8 beta 1
    • Thanks to Elizabeth for flying the beta on the weekend
    • Tiny, tiny change, will go out soonish
    • Any problems on the Copter beta would also apply
      • Which is likely to go out soon

UTC0038 - Copter updates

  • EK3 still initialising fix still required
    • Still waiting on Paul for that one
  • S-curve stuff actively being worked on
    • Spline curves working in a separate branch
      • Flying like 4.0
    • Calculator can come up with pos/vel/acc at each point on the path
    • Old calculator only did position
      • Needs to provide other data
    • Maybe 30%-50% of the way there
  • Tridge found a bug in the low-pass filter code which might be relevant
    • Filter was initialising from zero
  • Still poking away at 3D object avoidance
    • May still make its way into 4.1

UTC0045 - Rover

UTC0047 - need Wiki for VectorNAV

  • Warning it will only work on Plane and Rover because @50Hz

UTC0048 - autotune busted on a Hexa?

  • Turns out position-hold feature in autotune lead to a bad tune
  • Twitch twice in same direction enhancement to try to speed it up…
    • Can gain a chunk of speed then bounces against the circle and rotates
    • [11:49 AM] Peter Hall: You have to also turn off the rate target filters on 4.0
  • Need to contain the speeds we reach with the double-twitch
    • Very bad on quadplanes!

UTC0052 - close

1 Like