Dev Call Nov 11 2019 2300 UTC

Issues & Pull Requests

Notch Filter



1 Like

Attendee count (max): 21


  • Welcome Gus!
  • Welcome David Ardis

UTC1102 -

  • Old PR from Chris
  • Touched up a bit to do accel and decel
  • Race condition between test code and the ArduCopter code
    • The parameter is being dynamically updated by the code
    • The test is setting the parameter but the code is changing it underneath
  • Using parameters as temporary storage space
    • E.g. WPNAV_SPEED
  • Bill will change this to use a different variable
    • Some sort of an accessor?
    • Won’t be updating the parameter
  • Would be nice to compact logic a bit

UTC1111 -

  • No objections on the call.
  • merged!

UTC1114 -

  • Too strict on our error checking
  • Need a better story around “recovery mode”
  • Randy will merge after testing

UTC1118 -

  • Maybe we should stream out “recovery mode” instead of “config error”

UTC1120 -

  • Numerical or alphabetical?
  • On the initialisation enumeration case statement
    • Shows that whoever added the RC options couldn’t have tested them in SITL

UTC1121 -

  • Tridge needs to look at this one
  • Add a comment about the call to longitude_scale
  • Bumped to next week

UTC1125 -

  • Peter still working on it
  • Devcall left on it and we’ll look at it next week

UTC1128 -

  • Just documentation fix
  • Question about lidarlitev1 - probably same driver
  • Merged!

UTC1131 -

  • Needs tridge
  • Left devcall and moved on

UTC1133 -

  • Merged!

UTC1134 -

  • Needs tridge
  • Left devcall and moved on
  • Sanity check was added

UTC1135 -

  • Can be merged once the the zigzag stuff is removed
  • Zigzag does its own stuff and doesn’t need the higher layer’s protection

UTC1140 -

  • Ready for merge?
  • SITL-only
  • Needs RealFlight
  • Discuss topic to be linked in
  • Randy to review

UTC1143 -

  • Randy agrees new messages are OK

  • We can refine the algorithm once this is in

  • New messages coming

  • Randy will merge

UTC1149 - notch-filter tuning

  • Amazing results?!
  • Pushed P up a vast amount 60%
  • Went back and forth
    • Including attempting to run new tune with filter disabled
      • Which didn’t go well
  • Solo + this will be awesome
  • The Wiki instructions were sufficient to do all this
  • FFT analysis to determine where to put the filter
  • Mat acknowledges he has a tendency to see and hear what he wants to see
    • so he was somewhat rigorous with his testing here
    • Even on stock Solo there’s a significant change
    • This code doesn’t go into EKF
      • But Leonard said the control is more stable that there’s less garbage on the accelerometers which are going into the EKF
        • Less jitter
  • PR in place to update the frame parameter to enable the notch filter
  • New PID values for green cube
  • Frame resonance with/without gimbal?
    • Matt will experiment
  • The worse your Copter is the better this feature is
    • Even perfect Copters will
  • Andy is working on some more magic filter
    • To eliminate the manual steps
      • Dynamic FFT to work out where the issues are
  • If you have a Copter and you have vibe issues
    • You need to play with this

UTC0004 -

  • Beta testing continues on both Rover and Copter
    • RC option issue
    • Otherwise in pretty good shape
    • Oneshot not working?
    • Blheli telemetry issue resolved
    • Dual-gps for yaw needs help
    • Copter
    • Lots of users
    • Lots of reports to investigate each day
    • Navio2 leds not working
    • D-shot?
    • iBus not working for some users?
    • Cal not working?
    • GPS blending not working?
    • Parameter downloads slow when CAN enabled?!
      • User gives a brisk scenario on how to reproduce
    • Getting stuck on takeoff mission item when switching from loiter
      • Probably in 3.6
    • Probably some way off
    • Help appreciated!

UTC0011 - Close

1 Like