CraigElder
(Craig Elder)
1
Issues & Pull Requests
ChibiOS
GSOC
- 6 students chosen
- Announce students on 29th
Plane
- release plane 3.8.5 today
Copter
- Copter-3.6.0-rc1 has started beta testing
Rover
- Rover-3.3.0-rc1 has started beta testing
Attendee count (max): 23
UTC2300 - Craig not present for next two weeks
- Or MdB
- UAVSI on as well
- Likely to be a light-on meeting
UTC2305 - https://github.com/ArduPilot/ardupilot/pull/7877
- Board orientation options
- Looks good to FF
- These look OK to merge
- Tridge is happy
UTC2308 - https://github.com/ArduPilot/ardupilot/pull/8148
- [9:09 AM] To Weekly devcall: FWIW I’d like to use final elsewhere
- Does final infer override?
- Should be using override
UTC2314 - https://github.com/ArduPilot/ardupilot/pull/8125
- Peter will test today and push in if it is good
UTC2320 - https://github.com/ArduPilot/ardupilot/pull/8033
- Looks good, just need semaphore to pass
- Michael_duBreuil: I hit it
- [9:20 AM] (Channel) Michael_duBreuil: I don’t know why config sometimes take an hour
UTC2320 - https://github.com/ArduPilot/ardupilot/pull/8197
- A few problems crept in with some late-changes
- Good to go past those changes
UTC2320 - https://github.com/ArduPilot/ardupilot/pull/8213
- Slew rate functionality in RC
- Manual mode
- Passes directly through to motors except for the motor mixing
- Spin joysticks around and see motors spin
- Rapid joystick movements can tear vehicle apart electronically and physically
- We already have servo slew limiting in output?
- Needs a servo function to use this
- Seems wrong to tridge to do this on input
- Auto to manual mode would be a really sudden twitch
- Put the slew-rate-limit in the motors backend?
- Would also affect logging
- Multicopters already rate-limit motor change rate?
- No controllers involved in manual mode
- Slew-rate limit on the demand to motors?
- [9:31 AM] To Weekly devcall: Makes more sense to be in motors as it might be able to redirect thrust requirements to some other motors .
UTC2333 - https://github.com/ArduPilot/ardupilot/pull/6912
- Is this a valid change?
- Make sure it’s not a problem for compass calibration
- Big lumps of red and green where stuff is moved
- Some issues need to be addressed on the issue
- Probably an unrelated check, but should be confirmed
- Tridge has un-PR-ed changes to fix this sort of waiting-for-nav-checks by doing an EKF reset when on the ground and disarmed
UTC2338 - https://github.com/ArduPilot/ardupilot/pull/7819
- Is this OK to go in?
- Gps location vs ekf/ahrs location
- Still adds useful functionality even if it just GPS
- FF was only asking why GPS, not AHRS, not dictating!
- Devo doesn’t stipulate GPS, FrSky does
- So long as no harm is done outside we should accept new functionality
- AHRS will never give you garbage data
- Tridge is thinking he’ll go with the get_position
- Could always send relative altitude…
UTC2347 - ChibiOS
- Lots of new contributors to ChibiOS support
- New ports, new hardware, …
- KakotF4 Sparky2
- 16-bit timers
- New RCInput method, interrupt-driven
- Efficiency improvement on timers
- Please test!
- Plan on using integrated OSD?
- Need to properly support these boards
- Tridge has one of these on the board
- He’s planning on getting ChibiOS support soon
- SPI-connected
- F4Light already does
- frontend/backend split required
- Tridge’s RTF
- Integrated dataflash logging
- OSD
- Do we rename DataFlash?
- Flashlog backend to AP::Log?
- New upstream ChibiOS?
-
https://github.com/ArduPilot/ardupilot/issues/8109
- Not all of these are high-priority…
- Analogue stop-pins
- Two analogue sonars both pointing out the front of your Rover and you want them to interfere with one another
- HW has a “stop pin” so it stops transmitting
- [9:58 AM] (Channel) TP: I also vote we rename “fmuv2” to “pixhawk”
- [9:59 AM] (Channel) FF: Why? It’s not just Pixhawk…
- [9:59 AM] (Channel) FF: It’s a hardware spec
- [9:59 AM] To Weekly devcall: PixHack, PX4Pilot, …
- [9:59 AM] (Channel) TP: right, but it’s a massive part of our market and it’s considered a class of processor.
- [9:59 AM] (Channel) TP: those are all “pixhawk compatible”
- [10:00 AM] (Channel) TP: and fmuv3 can be “cube” or whatever. using “fmuv3” is horrible from a user’s perspective
- [10:00 AM] To Weekly devcall: It also means we’re not taking sides, 'though
- 4.0?
- Reserve for wiping the eeprom?
- [10:02 AM] (Channel) MdB: eprom wipe +1
- Add stuff to the document (#8109) if you spot stuff that is missing
- It would be nice to switch between ChibiOS builds and normal builds with a button
UTC0002 - Randy and Copter/Rover
- Both betas went out a few days ago
- Good feedback already
- SmartRTL needs a fix
- No control of vehicle until very very late
- 2m before ground
- Very last minute
- Change to some sort of cone?
- Maybe 10m from ground?
- Stick-mixing works very nicely on the SkyViper
- Whether it works on larger frame…?
- Pivot turns are improved
- Still slow, ‘though
- Need to increase the angle-error-to-desired-rate gain
-
UTC0007 - GSoC
- Announcements have been made
- Students have been informed of our decisions
- https://docs.google.com/document/d/16pqxA2fe3_daTBq82iJ2sdc4Bu8V7gdxlIz_aS2Pf9g/edit?usp=sharing
- We can now blog and announce it
- Dmitri is on the call! Welcome!
- [10:09 AM] (Channel) DV: Thanks for having me!
- We’ve had balance-bots twice before, but not to the point of getting it into master
- Jacob would like to help with the live-video stuff
- Dmitri is planning on helping with UAVCAN
- There’s lots of companies out there that want to make uavcan ESCs
- [10:11 AM] (Channel) Randy: eLab has a balance bot being built for Ebin
UTC0012 - AUVSI
- Need a flyer to hand to potential partners
- Cut-down version of existing document