Dev Call Nov 26 2018 2300 UTC

Development Team
1

Issues & Pull Requests





  • Extending the battery monitor library to be a more general power monitoring interface

Plane

  • Update

Copter

  • Update

Rover

  • Update
2

Attendee count (max): 25

Pre-agenda:

  • DJI API concepts

  • Tridge is keen

  • Spend a day to scope out what would need to be added

    • Basic ping style function to learn enough about API to see how big a project it is

  • DJI API has some mavlink hooks in it already?!

    • Can do some mavlink ops through their SDK?

  • Another serial protocol on a serial port?

UTC2304 - removing plane update event scheduler

  • Merged!

UTC2307 - removing NuttX

  • PR has been created

  • Passes CI!

  • FF has reservations about losing VRBrain support

    • Queries have been answered on the issue

    • FF wants the Makefile build system gone

    • FF: We haven’t given fair warning

    • Tridge: who to?

      • FF: VRBrain

      • Tridge: I gave them a year!

        • Private discussions

      • Tridge is comfortable removing VRBrain

        • It won’t be a surprise to the VRBrain guys

      • They were supposed to send tridge a schematic and a board so he could do the port

        • They haven’t

  • Publically we haven’t set a firm date

    • We have announced that 3.7 will be ChibiOS only

      • This PR is preparing for that

      • 3.10 for Plane as well

    • Tridge: Shall we set the date as the next dev call?

      • FF was thinking 6 months…

  • Hal F4 Light also requires Make

    • Too many people using Make

      • On gitter

  • “Make” actually works well now in tridge’s PR; it guides people through the use of waf

  • MdB isn’t comfortable with the timeframe

    • Too many things popping up

    • MdB flies master

      • Missing too many things in Plane stable for QuadPlane

    • E.g. i2c lockup

      • Behaviour under NuttX was worse!

    • Flying px4-v3 master

  • “As long as the old one is available people will keep using it”

  • Tridge doesn’t fly NuttX for releases

    • He only tests ChibiOS

  • Not doing this is holding up doing HAL cleanups

  • Nate points out the old NuttX IO firmware was in use until recently

  • [10:18 AM] To Weekly devcall: @MdB what would you consider a good timeframe?

  • [10:19 AM] (Channel) MdB: Peter: I still want the dual release with parity, but apparently that can’t happen

  • [10:19 AM] (Channel) MdB: I’d be a lot happier with the end of year

  • Master is supposed to be our next generation

  • Tridge’s talk in February said it was going away

    • MdB said it stated there would be parity and that’s now not happening

    • Tridge thinks it was actually “dual release”, not “parity release”

  • Tridge believes we’re now at parity

    • MdB thinks that we should have releases in this state

  • Randy is happy for the next release of Copter and Rover to go out without NuttX and would prefer this argument to stay about Plane

  • MdB wants a Plane release off master with both platforms in it

    • Tridge doesn’t want to do that

  • [10:27 AM] To Weekly devcall: @MdB I’m really not sure what a dual-release off master would gain. People will just keep flying NuttX on it.

  • Craig to MdB: can you fork from master and support customers from that for a while?

  • MdB doesn’t have time to fly ChibiOS at the moment

    • Tridge points out the opportunity to fly it has been there for a very long time

    • MdB points out it didn’t have feature parity until recently

  • Tridge wants time to restructure the HAL

  • 10:31 AM] (Channel) MdB: I think you’d see a wider commerical side but I’m just the one here

  • [10:25 AM] To Weekly devcall: 14,000 line delta between master and ArduPlane-3.9.2

  • 10s of thousands of APM downloads/month

  • Would make an archived directory to hold APM and NuttX builds

  • Nate would like a ChibiOS-only beta soon

  • Delayed another week because we haven’t managed to resolve on the call

UTC2341 - semaphore passing when it shouldn’t

  • Tridge will fix

    • Actually, tridge’s testing shows that it does fail

UTC2341 - https://github.com/ArduPilot/ardupilot/pull/9844

  • Add separate radius for landing vs normal approach

  • Ensures you’re far enough from WP to start approach

  • Separate land airspeed

  • Automatic would be nice

    • MdB thinks he has the maths floating around

  • Nasty p1 hackery

    • Fix determined

UTC2351 - power monitoring

  • Battery seems to be where this stuff is done

  • But as we grow and want to add things like overpower is this the right way to do it?

    • If we grow a function indicating what the power source is used for it might be a nice cleanup

    • Nate thinks so

    • [10:53 AM] (Channel) eosbandi: Dont forget other power sources such as generators or tethers

  • How about a “watt_max”?

  • Watt limit should scale with voltage

  • Discussions of rejigging things to be energy-based

  • Fuel-gauging

    • Temperature

    • Air density

  • Need to separate current-limiting from energy management

  • Amp limit per battery will be put in – MdB

  • Need to make sure tethers fail means copter lands

  • Some tethers appear as smart batteries, some don’t

UTC0004 - UAVCAN dynamic node allocator

  • FF has a few issues

    • Stated on issue

  • Storing IDs in filesystem isn’t great

    • Having an open file descriptor all the time isn’t cool

    • This will be replaced in the future

    • opening/closing in-flight is BAD

      • Put a limit (say 100) on open/close and then start to reject allocations

UTC0017 - Plane update

  • Nothing happened in stable

  • PixHawk4 crashes at particular flying field

    • PX4 project also has crashes with PixHawk4

    • Possibly a HW bug

  • Tridge has improved robsutness of s.bus parsing and dsm parsing

    • Much better now in ChibiOS vs NuttX

  • Solo doesn’t follow the DSM standard

    • Fixing involved extra robustness checks

  • 12 actuators in RealFlight FlightAxis simulation

UTC0019 - Copter update

UTC0021 - dataflash

  • Hw tridge bought isn’t suitable

  • Still looking for fixes

UTC0019 - Rover update

UTC0025 - lots of links for microsd to SD

[11:29 AM] (Channel) MdB: https://www.newegg.com/Product/Product.aspx?Item=9SIA7257AS5778&ignorebbr=1&nm_mc=KNC-GoogleMKP-PC&cm_mmc=KNC-GoogleMKP-PC--pla-Best+Price--Card+Reader%2FAdapter-_-9SIA7257AS5778&gclid=EAIaIQobChMIx6Csxajz3gIVCzlpCh2YEgx6EAQYBSABEgLwlvD_BwE&gclsrc=aw.ds

[11:29 AM] (Channel) CE: https://c1.neweggimages.com/NeweggImage/ProductImage/ABV7_131687527074608308PnAgLTekrK.jpg

[11:30 AM] (Channel) CE: https://www.amazon.com/LANMU-Extension-Flexible-Monoprice-Raspberry/dp/B01D9JIUU0

UTC0031 - does anybody have a setup to test a fix for a race condition in the statustext frsky stuff?

  • Nate