Dev Call July 26, 2021

Issues & Pull Requests





Attendee count (max): 15

UTC2300 -

  • Too much spam
  • Sid has already fixed it

UTC2302 -

  • Support CAN read and write in scripting
  • Sid tracked down issue
  • Merged!

UTC2307 -

  • Move tailsitter to class and hide params
  • Merged!

UTC2308 -

  • Atmospheric tables for high altitude flight
  • Improves errors even at low altitudes
  • Standard tables uses in general aviation
  • So we’ll be using the correct maths as normal airspace
  • Complex curve
    • Tridge tried fitting a polynomials up to 7 order and couldn’t get a good fit
  • Tables aren’t actually correct
    • But all aircraft use the same incorrect tables
  • This uses the standard US tables
    • Linearly interpolated at gaps of 1000ft
    • Table is in metres, this is mil 3103
  • 9:16 AM] Peter Hall: -15000 feet?
  • [9:16 AM] Charlie(TunaLobster): Yes. Higher pressure days in death valley
    • Only a few entries to go this low
  • 150,000ft is 10 times harder than 100,000ft
  • International organisation for flight actually did decide to go to metres
  • EAS2TAS into AHRS?
    • Could be getting eas2tas from somewhere else….
  • A review on this would be good

UTC2321 -

  • Acceleration is actually a proxy for lean angles
  • More investigation needed on this

UTC2327 -

  • If rc failsafe is ignore (continue in auto)
  • And then you change mode to stabilize via GCS
    • Vehicle has no RC in stab and won’t failsafe
  • User has kind of asked for this
  • This PR re-evaluates
  • Add check in start of set mode?
  • Maybe shouldn’t allow you into stabilize if you don’t have RC?
    • But stab is a mode we almost always say you can go into
    • What should we do for throttle if we do allow you into stabilize?
  • Radio suppressed should go to false before calling radio failsafe on event?
  • Edge-triggered where condition becomes applicable rather than when RC failsafe occurs
  • There’s a small race condition where we could end up with full throttle
  • Init in start of flight mode
  • Need additional check for RC input
    • Has_manual_throttle
      • And possibly drift mode
  • On edge of RC range?
    • Change mode is edge triggered on switches?
  • If we just refuse the set mode we’ve lost configurability of the failsafe option based on which mode you’re in
  • Current deferred setup does allow you to get into stab because once actioned the deferred failsafe goes away
  • Maybe allow into stab but center the sticks and hover throttle in case of RC failsafe
    • And potentially re-trigger the failsafe
  • Move from edge to level-based failsafes?
  • [9:43 AM] Alex/ARACE: For plane. I have and know others that have beed in situations had RC failsafe and needed to switch to FBWA to land it with not RC control without waiting for the battery to die. Albeit those are mostly hobby fly situations where you may not have reliable GCS.
  • [9:46 AM] Alex/ARACE: @Henry Wurzburg I specifically remember the video you showed me. You aligned the plane and let it land in FBWA, because you did not have enough battery to get home. But you know you will lose RC when you get close to the ground.
  • [9:47 AM] Henry Wurzburg: it was with RVOSD, but that was the action…lose RC while in stabilize, it just glided to the ground
  • [9:48 AM] Alex/ARACE: yes. but translate that to Ardu if this PR was in. Anyway it is very Fpv mindset
  • [9:48 AM] Buzz: why not allow both options, with a param? “ALWAYS_ALLOW_STABILISE_EVEN_IN_FS”
  • [9:49 AM] Henry Wurzburg: yeah I had changed the fs action to do nothing while I was still trying to fly back using the on screen parameter menu
  • [9:49 AM] Alex/ARACE: For copter I am not worried. Just plane needs more thoughts
  • [9:51 AM] rmackay9: I think we should raise an issue so we don’t forget this potential fly-away … assuming we’re not going to merge this immediately.

UTC2353 -

  • Auto RTL mode
  • Auto reports as this flight mode when we’ve done the DO_LAND_START thng
  • Merged!
  • Mission rejoin is coming
  • Need to get back on the line

UTC0000 -

  • Symbols in vehicle when no scripting enabled…
  • Nmfind to find references to symboles
  • We’d like to know why this is making a difference to anything as the symbols should be being elided anyway
    • Use nmfind to find users of those symbols…

UTC0005 -

  • Jump the other way and update AOA/SSA as part of the loop not the accessors

UTC0011 -

  • Separation of vehicle frame and autobody frame
  • Why have trim at all?
    • Cope with velcro mounts
    • Trim should really just be coping
  • PB; Should eliminate trim and just make users use custom rotations
  • Need Paul’s input on this

UTC0020 -

  • 45-degree orientations
  • May make compass orientation detection problematic
  • We have +/- yaws
  • New rule: don’t detect any 45-degree angles?
    • Just do the 90s
    • New enum entry for doing all angles?
  • Solo leg is a pain
  • Shouldn’t auto-orient that one!

UTC0023 -

  • Nmea output up
  • Can be merged once rt is checked

UTC0033 -

  • Switch compilers in master
    • And then for 4.2
  • New compiler makes for smaller binaries
  • Should really be using same compiler across all boards
    • Something for later
  • Beta6 copter went out so this could be merged today….
  • Some bugfixes also went in here
  • Saves 10kB!
  • Fastmath would also save us about 10kB
    • Magically turns division into multiply
    • Ieee floating point doesn’t allow this
    • Says we do assume maths is associative
      • Also turns off some nan and inf checks and whatnot
        • Can turn those back on

UTC0045 -

  • Need to sprinkle some enabled() methods
  • Stop using compass through AHRS

UTC0101 -

  • Plane beta is going well
  • All of the double-precision stuff is working well

UTC0102 -

  • Beta6 is going out
  • Aligned with Rover and plane so basically the same code
  • MissionPlanner auto-analysis is out-of-date
    • Need to check the one in the AP repo
  • Guided mode is not ignoring yaw input on takeoff
    • There’s an option which isn’t being obeyed
    • Full yaw on takeoff when the params say don’t do that
  • Goku autopilot
    • Qmc compass not recognised
    • Rebranded matek?
  • Rapid pos oscillation when using gps for yaw
    • Need to get Paul get look at this
  • Motortest main-loop-stuck issues

UTC0106 -

  • New beta
  • GPS for yaw is in!
  • Rover driving at offset?
    • Tridge will have a look at log
    • Maybe an existing configuration issue
    • User changed GPS setups
    • Mission lat/lng might need to move?
      • Our handling has improved
        • [11:08 AM] jmachuca: When did that show up?? (offset in navigation with gps yaw)
        • [11:08 AM] jmachuca: I have seen an offset too, but I thought it was my tunning
        • [11:11 AM] jmachuca: yeah but no mike
    • EKF applies offsets for GPS
    • No RTK fix on base GPS
      • Only a float fix
  • Heli guided mode fixes are coming!

UTC0112 - GSoC update

  • Will is volunteering at a Summer camp
  • Arsh is working on serial objet use with tridge
    • Tridge has it working in SITL on his system
  • Shorter GSoC == running out of time
  • Rishabh’s precland is going well
    • Changing over to use new features of the position controller
    • Input shaping
    • Any change in the target don’t change large input angle
    • Blog post to come!

UTC0120 -

  • No DMA channel for crossfire? What do you do?
  • Scheduler for next week…

UTC0122 - OSD parameter issues

  • Peter is still looking

UTC0123 - close