Dev Call June 21, 2021

**Issues & Pull Requests"





1 Like

Attendee count (max): 20

UTC2300 -

  • Smart battery info message
  • To be broken into separate PRs

UTC2315 -

  • Log in more reasonable units, and correct some units
  • merged

UTC2320 -

CAN from scripts

  • Problem doing can_iface_select in scripting if something else is reading from the same CAN port
  • Hard-lockup issues still
    • exclusive access required
  • Currently needs a delay to make it work
  • Need to get rid of the delay
  • Will test in SITL

UTC2327 -

  • Reject mission if wait numbers negative
  • Should we squash to zero?
  • Easier on the user?
  • Check for NAN - and infinities
  • Have the GCS validate the number

UTC2344 -

  • Fix build of vrubrain
  • Too big
  • Merged
  • Variation between boards is mainly features enabled

UTC2348 -

  • Added for experimental board
  • Orientation was wrong
  • Merged

UTC2353 -

  • Update dshot settings at 1Hz
  • merged

UTC2354 -

  • Check if EKF origin exists before setting origin in non-GPS modes
  • Able to set scripting over-and-over again from scripting which is bad
  • A few comment fixes but can then be merged

UTC0000 -

  • Can be merged after rebase is done
  • Disable airspeed on a switch

UTC0002 -

  • Save a few bytes
  • But not on all platforms
  • Saves bytes here, loses bytes there
  • Merged
  • Amilcar used the same strategies in here and saved 140 bytes in new code

UTC0009 -

  • Add water depth
  • Sonar surveys usually add a delta for GPS antenna height
  • Quaternions?
    • Load into a spreadsheet and you’re stuffed?
    • Stay with roll/pitch/yaw
  • Distance is absolute raw - doesn’t account for temperature

UTC0022 -

  • Rover bits for above
  • Always sens for every measurement

UTC0028 -

  • Add sample values for some bitmasks
  • Tridge would prefer bit selections
  • Need to click in just the right way to get the bitmask options up
    • Maybe clicking in the range field should bring up the bitmask?
    • Clicking in the edit box brings up the interface but whether things are selected whether the bitmask thing shows
  • Randy is fine with removing Values

UTC0046 -

  • Force-descend only in last phase of descent in LAND
  • Force descend is about land detection not actually about descend its not about making the vehicle go down
    • Limit_vector.z
  • Needs equivalent quadplane change
  • Force_descend should not be set for a long time
  • Callisto testing landing @15m/s

UTC0058 -

  • Glitch when RTL engaged in auto mission
  • Stopping point calculated where vehicle currently is
  • Need to better handle kinematic inconsistency with z velocities
  • Vibration failsafe not working is not good

UTC0108 -

  • Smbus still in on a lot of boards
  • Leave it in, don’t want to break stuff

UTC0110 -

  • Working far-from-home is a problem
    • More than what we thought
    • 0.2m/s quantization at 12km sort of thing
  • Double-precision maths is probably the fix
  • Not new for 4.1
    • Do we want the fixes into 4.1?
  • This is for awareness, not for merge today
  • Also affects SITL
  • Speed costs on F4?
    • This brings it in for everyone
  • This PR doesn’t bring EKF changes in
  • CubeBlack is the big question
    • Highest CPU load of any board we have
    • F4
    • More likely to be flown long distances
  • Drone racers might suffer from CPU
  • Having just this in?
    • Pushes the problem out from 12km to 100km
  • Tridge is planning on doing a bunch of bench testing to see what the CPU cost is
  • 17823 cleans up APIs a bit
    • Some evil casting down from Vector3 to Vector2
      • Saves a bunch of data copying
    • Saves CPU
    • Saves flash
    • Saves RAM

UTC0131 - Plane update

  • Quadplane airbrake fix went in
  • New beta imminent
  • Autotune isn’t working as well as it should

UTC0132 - Copter update

  • Flight mode change / rc option conflict detection
  • Inconsistent alts fix coming
  • Two requests that we allow aligning vis odom with gps-for-yaw
    • Maybe
  • Loiterbucketing we just talked about
  • Reports requiring investigation
    • Ground effect compensation not working as well as 4.0 when no GPS
  • Core-x-unhealthy if using GPS for yaw as yaw source
  • DJI-MSP-OSD requires serial flow control to be turned off
  • When switching between yaw sources compass/gps using scripts doesn’t flop after being flipped
  • Bouncy terrain following
  • Rocking loiter when on ground
  • Closed off many issues this week
    • Thanks to all helping, inc Andy and PH

UTC0134 - Rover update

  • Lost accel cal parameters?
  • Rover’s avoidance of fence questions
  • PRX lib has disappeared on 1MB boards

UTC0137 - GSoC update

  • Pending PR from Will for custom firmware server
    • Next day or so merge

UTC0138 -

  • Raising barriers to entry for non-native English speakers
  • Nobody had strong feelings on the call

UTC0140 - close