Issues & Pull Requests
- update on watchdog code (including in-flight reboot successful test on a plane)
Issues & Pull Requests
Attendee count (max): 22
Sending jump information to GCS
Should we be sending so much text to the GCS?
Two good reasons for it
Currently no way to track jumps
Has confused users before
Jump counts of zero
Should include mission item count ‘though
[9:06 AM] (Channel) MdB: I hate status texts, but this is the only insight, which is fairly relevant for state tracking…
[9:06 AM] (Channel) MdB: avoiding #'s in the string would be good though
[9:06 AM] To Weekly devcall: As in literal hashes?
AndrewTridgell: Jump 17 4/-1
[9:07 AM] To Weekly devcall: s/-1/Inf/?
Only call into this if we actually do the jump
Need to do it all the time, even if we’re not going to take the jump
[9:10 AM] (Channel) AndrewTridgell: Jump to 17 (4 of 5)
Other conditional jumps should include details
Tridge hasn’t had a chance to look at this yet
Qacro with rate mode wasn’t setting pitch/roll values for logging
Minimal patch to set from multicopter controller
This also does qstab
Rate mode stab is done in qacro so it calls stabilize methods
AttitudeControl routine being modified should only be being called by QuadPlane
There’s various other methods in AC_AttitudeControl which are QuadPlane-only
Maybe do a ccache -C as last part of the run?
Are we going to move the gcc forward from 4.9?
Template patch was removed
UTC2336 - rangefinder prearm checks
Tridge will try to get to this this week
Restore plane throttle min behaviour
Plane: qstabilize throttle midpoint and expo
Only save gains when landed in autotune for quadplane
Minor cleanup for functions suggested by Randy
autotune wasn’t always a flightmode
What happened to saving gains with a switch?
It’s in the docs
If you take off again in autotune it will use the new gains
Going low-high on the switch should use new gains?!
We should test this again
Will raise an issue
landing/disarming when doing single axes?
Peter gave a bad method name - will need to be changed
Should we make it take a uint8_t?
Check error return values are consistent
Maybe return a bool and a reference to return value?
Bugs fixed and tested
Flag in sim aircraft to give kinematic consistence
Rebased and pushed up
Looks like a good fix from MdB
Tridge doesn’t like having a single mode.h file
Neither does Peter
Massive number of files
Does this clean up really help?
Need to slow down the rate of change for stable releases
Added a failsafe for moving channel values
Tridge adding a comment and closing
Files are going away
MdB is unhappy with scripting modifications
Randy will try to test this
Probably a mission item ordering problem
don’t publish frozen compass data
Requires 30milligauss of noise from compass
If you have a really good quality compass it will be marked unhealthy
E.g. arming and taking off with really good compass could become a problem
If there’s a particular compass that exhibits the problem we should look at why it is happening
Are we checking return values in qmc driver?
Maybe this has already been solved by one of tridge’s patches? (see issue)
Respect condition yaw during landing
Assigned to Randy
Replaces an old one from Jon
fix TECS true airspeed rate
Do we want the filter on the true airspeed rate demand?
Can probably be closed
Tridge will think about it
Peter’s brain left out to lunch
Will need to get back to this
UTC0044 - baro vs gps checks
Use an atmosphere model
Need some hysteresis?
Should we mark a baro as bad if we seem really bad temperatures?
People could arm at very low pressures and high alts
Should check temperatures one against the other
UTC0049 - GSoC update
UTC0050 - notch filters not working?
Fft is on raw data, not the notch filter
Wiki update required?
UTC0051 - Plane update
New stable release
Stm32 watchdog timer
Can’t change protocol because of NuttX support ATM
Does correctly reset safety state etc and updates mixer and gives a flyable Plane
Setting default rate to servos is broken
Probably still have a flyable plane
But we can do better
CAN timeout needs work
Watchdog stuff is in master
Ongoing discussion as to when Copter will get it
Definitive logging of watchdog event is the good bit
Distinguishing brown-out vs CPU-lockup is the tricky bit
Septentrino GPS patch will go into next beta
Tested lockup in flight - in flight
Can restore altitude and attitude and rearm it after an in-flight lockup
[10:58 AM] (Channel) LV: armed with right rudder stick ???
[10:59 AM] To Weekly devcall: @Luis: yep!
[10:59 AM] (Channel) LV: so plane flew right because of the arming procedure ?
[10:59 AM] To Weekly devcall: Yep. But I rolled it left a little
The mavlink command - should we have this?
Should we tight-loop?
We tried it on a Copter too
Not so good
Would have needed to be at 150 or 160 metres
Heli would work
UTC0101 - Copter update
Randy’s been looking through the issues a little
Funny report from Marco this-morning on an unexplained crash
Falling a little behind Plane on releases
Almost ready to start beta testing on next Copter stable
UTC0102 - Rover update
Oneshot issue still needs looking at
Should be a beta this week
Bug in omni3 frame type?
UTC0104 - Randy has a RealSense
Randy’s hoping to get stuff documented on non-GPS navigation
And some work on object avoidance
Jaime’s also been working on this
10Hz shaking bad
Pointing at corners bad
UTC0105 - Roadmap published
If you’re a dev and you want to add to the list, ping Randy
UTC0107 - End