Servers by jDrones

Dev Call Aug 2, 2021

Issues & Pull Requests





Attendee count (max): 23

UTC2302 -

  • Crsf needs DMA for best performance
    • Tells you you do need it
    • Doesn’t tell you how to fix it
    • We don’t need DMA on H7 now we’re using the fifo
    • F405 wing passthrough issue?
  • 400kbaud
    • Autobauding
    • Negotiating
  • Put some effort into interrupt latency might be able to not lose bytes at 400kbaud
    • Above 230 you should have dma on receive
    • For xmit you’ll just use more CPU
    • Receive dma that matters
  • is Andy’s PR

UTC2305 -

  • Closed as not enough detail

UTC2308 -

  • Need to get the mavftp stuff upstream
  • Closing this as we’ll never implement it

UTC2309 -

  • We would accept a patch for this

UTC2312 -

  • Closed as must have been fixed already

UTC2312 -

  • We can’t 100% know we’re flying….
  • Add is_flying ?
    • Might not start to fly
  • Won’t fix the heuristic based on any report we’ve received so far
  • Use disarm_delay in Plane?
    • If throttle is suppressed for any period of time disarm?
    • Leonard has had a disarm midair with extreme upward draft and a quadcopter

UTC2316 -

  • Assigned to Randy

UTC2320 -

  • SITL ride-along instances via JSON
  • Merged!
  • Rename?
    • Ride-along?
    • Secondary?
    • Redundant?

UTC2323 -

  • Param for i2c address, move analogue and i2c params
  • Lots of testing done including valgrinding
  • Can now set address of battery device
    • Parameters….
  • Changing addresses on device?
    • Done in lots of different ways
  • Another 1/2k of flash on the 1MB boards
    • Few 1MB boards would have smbus batteries
      • We want to get an option around this
        • There is one here
    • Disabled on F405?
      • Do we know why are we using more flash?
    • Shouldn’t be
    • Nm needed!

UTC2336 -

  • Atmospheric tables for high-altitude flight?
  • Josh did that maths…
    • Use equation rather than table
    • 80 pascal error using equation
    • Wonky interpolation errors?
  • Still some work to be done

UtC2339 -

  • Qmode tidying!
  • Tested in SITL
  • Saves a few bytes

UTC2342 -

  • Alt frames for fences
  • Can’t change user-set parameters….
  • Should be a relatively small patch
  • Tridge will help out with this

UTC2347 -

  • handle command-long and int within scripting
  • Ringbuffer
  • Lots and lots of flash!
  • Can stop the main code handling code and do it themselves
  • Register for command long of 31010 for example
  • Register_command_int and register_command_long
  • Can send ACK
  • Receive is non-blocking…
  • Why so much flash space?
    • New Buffer class
    • User data type for command long and command int
      • Generated bindings again!
    • 6kB
    • Templates?
    • Nullable with uint bitmask…
    • Part of the idea of scripting is to pay for interpreter once we have functionality
    • Compile scripting bindings with -Os
    • Durandal is down to 269kB
    • Starting to look a bit iffy
    • This is 4 or 5% of our 2MB board space
    • Really need to find space used
  • RAM cost for new bindings?
  • There’s a PR for libraries in SITL

UTC0003 -

  • Remove force set position typemask from messages
  • Nice cleanup
  • Can be merged once clean

UTC0005 -

  • Still want to look at show view vs vtol mode
  • OSD doesn’t work in VTOL modes
    • This now fixes things in vtol modes
  • OSD is corrected for fixwing modes pitch-trim-cd and it now works in vtol modes
  • Semaphore?
    • Probably doesn’t matter so much here
  • Merged!

UTC0009 -

  • Remove gcs pitch offset
  • So GCS gets non-pitch-corrected attitude
  • Having people hold the stick in fbwa because of confusing GCS data is bad
  • Flight option bit

UTC0012 -

  • No change in default behaviour
  • Can set maximum log rate on a per-backend basis
    • So everything still gets logged but each message a a maximum rate
  • Might be removed in future for a per-message-rate thing
  • Can be updated in real-time
  • More comments needed on the API?
  • When do you use this?
  • DSF.Dp going up means you could use this
    • Or gaps in your graphs……
  • If your logs are way, way too big this might also be useful
  • 50Hz IMU data isn’t useful on a boat going a long way….
  • Do we need more visibility over dropped messages?
    • Make subsystem unhealthy?
    • Statustexts?
  • Need a “long distance” section in use-cases
  • Useful for getting Replay logs, too!

UTC0022 -

  • Move event logging up to AP_GPS
  • More consistent between vehicles
  • Should we have a statustext when you switch GPSs?
    • Too much stuff for the user?
    • Don’t emit stuff to user when things are working as they should
  • Can be merged once the -1 thing is done

UTC0028 -

  • Stop passing mission into the constructor
  • Merged

UTC0035 -

  • Make sure home stays restored once we restore it
  • Merged

UTC0037 -

  • Not enough time

UTC0036 - Plane update

  • Plane beta is going well
  • Did find a 60ms delay which is a bit concerning
    • No explanation for that one yet on a dual-IMU board
    • Definitely a known HW problem with these boards, but we can’t explain this one with that HW problem
    • May be a real issue lurking here somewhere
    • Two spots where it overran by 60ms

UTC0042 - Copter update

  • Beta6 is looking OK
  • Leonard found some issues on the control side of things
    • Guided mode doesn’t follow guided velocity above wpnav
    • Some safety fixes for divide-by-zero if people made some changes
    • Yaw-on-takeoff issues
      • Causes some yaws
      • Any movement would cause yaw
      • Fixed
    • Lean-angle adjustment
  • Some other blockers
    • Helicopter guided mode takeoff issue
  • Pxhawk4 LEDs are swapped on the board?
  • batterymonitor default parameter issue?
  • Lots and lots of issues require investigation
  • Randy needs some help on the system-level stuff

UTC0048 -

  • EKF attitude going bad if you leave vehicle for a long time
    • Hopefully resolved in beta6 already….
  • Position can jump and you can end up with rapid switching between cores
  • Very rapid switching between cores
    • Multiple times per second
  • Paul thinks there’s something wrong with the switching logic
  • GPS yaw is greatly improved in beta6
    • No report from Rover users saying they’ve used it successfully yet

UTC0052 -

  • Two new chips
    • Requested by various vendors
    • Scrabbling to find available chips
    • No hardware to test these on….
      • Merge anyway and wait for reports?
    • 128kB flash and low memory
  • Merged!

UTC0054 -

  • Stop segfaulting with bad command
  • Merged!

UTC0055 -

  • Tack scripts in git?
    • Need to be able to add them easily….
  • Merged!

UTC0057 -

UTC0100 -

  • Already discussed this one
  • Tridge will keep thinking about it
  • Might PR something to reduce the MON trigger time
  • Need people to look in logs for vestiges of this one

UTC0101 - GSoC update

  • Hopefully Will is back soon!
  • Great demo from Rishabh on AI-based machine-vision precision landing

UTC0104 - close

Did find a 60ms delay which is a bit concerning

may I ask, what is the definitely known HW problem with these MatekH743 boards?
I would be really eager to know that :slight_smile:
many thx

Servers by jDrones