Attendee count (max): 26
09:03am - Randy and Copter 3.5
Still testing 3.5rc7
Planning on pushing rc8 out later this week
No surprise reports in last week
If anyone has seen anything then speak up!
Couple of reports remaining
Spectrum not binding
PixRacer / DSMX producing crazy values on higher channels
Randy has ordered the same hardware so will attempt to reproduce
Still have a delay in waypoint turning
Randy will be working with Leonard to attempt a fix
Still have EKF-OK issue
Landing gear issue
Patch in place to not put out any PWM until switch toggles
One user says it is great
Maybe we should add a parameter to set the default output instead
What do they do if they don’t want any value at all?
So on power-up servos are never strained
Need to make sure we can actually do 0 pwm individually for all channels
Would be good to be able to demand 0 on a channel to stop PWM going out at all
Perhaps allow a trim value of 0 says “no outputs”
We don’t use servo min/max for landing gear
Now we have output split we can eliminate the special parameters for land/deploy
Generally trying to eliminate using raw pwms and move to scaled values
No-change-in-state is probably preferable
Landing gear might partially deploy if knocked?!
Or be deployed but not holding position?
No pulses or a specific output would seem to handle these concerns
Prearm check that landing gear is in a set state?
We’ll go with the 0-means-no-output thing, but maybe for 3.6
09:27am - copter control issue
Maybe a bug in SITL?
Rob’s got a neat video showing an oscillation issue with loiter in Copter
Copter doesn’t stick where it is supposed to be
Could be to do with the time scaling
EKF10 was a workaround for a delta-angle/delta-velocity time-scaling bug
It was actually a real bug in the flight code, not in SITL!
Should try with built-in SITL or XPlane helicopter rather than FlightAxis
Randy and Peter will install FlightAxis to try to replicate
Robert is enjoying using RealFlight for testing, everybody should use it!
NDA required to use properly with SITL
09:41am - UAVcan and baro
Unhealthy when no Baro
09:42am - safety switch for flashing?
Needs a look over as the only patch found doesn’t seem to address the problem
Update to previous release may not solve transition-to-3.5 problem
Someone will need to dive into px4 code
09:43am - Craig arrives
Avoid being poked in eye with blunt stick
General agreement on this point
09:44am - top contributor
Email will be sent out with a vote associated with it
09:54am - Tom and mission authorization
CC prearm check
Ask cloud whether it is OK to fly
Utm is traffic management
Altitude Angel etc etc
Should the request be broadcast or should it ask a specific ID?
Maybe we should have a key instead
challenge/response / shared secret
Pair of mavlink2 messages
One to request authorization
One to provide authorization
Just re-use our existing key
Send a nonce, expect a hashed response
Require signing to be enabled to use the feature
Need to handle very long delays in telemetry
2 or 3 minutes
10:13am - https://github.com/ArduPilot/ardupilot/issues/6413
PX4 use a block method
We use a klt algorithm
After doing some testing (in 2014!) testing showed klt was best close to ground, poor lighting condition, poor contrast, …
Poor error reporting but EKF can sanity-check
PX4 stuck with block method
We’re no longer compatible with the firmware
Can’t really switch back without testing as the previously results show klt was much better
Do we have a method on checking focus on the flow sensor?
MP shows a little video of what the sensor is seeing
Don tried to be able to support both sorts
We identify as PX4 which is problematic
Fixing the identity string seems easiest
Randy will give this a go
10:22am - tridge and Plane
Tridge is making progress on Plane documentation
Fairly major rewrite of Plane docs
Already getting good input
Tridge will continue to work on the docs this week
First release of APWeb web server
In own repository https://github.com/ArduPilot/APWeb
Will be in APSync
Built on top of a mavlink to JSON gateway
Should get this onto e.g. Solo
Will work even if you can’t get to the internet
Pouchdb for persistent data
Would be nice to put onto an esp8266
10:43am - MdB and LOG_DISARMED default to 1?
0 by default
Should it be 1
Always have a round-trip when doing support asking the user to set LOG_DISARMED to 1
Problem is: enormous logs
LOG_DISARMED value 2
Start logging when an arming request comes in?
Most issues tridge deals with he doesn’t see the problem of no-information-before-arming
Log at a lower rate while disarmed?
Statustext overlay on graphs is awesome
10:54am - Jaime
Annual conference in Mexico
About 1 month from now
Jaime is giving a keynote
Wants to do it on Open Source / ArduPilot
Wants to do a remote flight like tridge has previously done
Needs a volunteer!
Keynote is at 10pm
10:58am - statistics resetting with Amilcarlucas
When loading a parameter file we reset the statistics
Shouldn’t do that at least without confirmation!
See also BRD_TYPE
Read-only metadata for parameter
Should use the system identifier (e.g. from stm32) to see if you’re loading the data back onto same vehicle
Peter will comment on the issue after investigating it further
11:15am: off the floor
GSOC update would be nice next week!
Miguel says OK!
Seabstian says OK!
Amilcarlucas says there’s an issue with Emlid Reach GPS units
PRs are in place to fix it
Don’t fly before these are in!