Attendee count (max): 21
09:01 - Randy and Copter
No 3.5.3 release
Things seem to be going OK with the release
Planning on 3.5.3rc candidate this week
With GPS issue a 3.5.3 might just be pushed out quickly
To be followed by a 3.5.4
Could also just do a big commit
09:04am - GPS and bugs
Mid-flight shutdown
Mob in New Zealand
Reproducible
SBF GPS driver
Septentrino
Long packets from the device with a specific data pattern would cause a parse bug and cause memory access bug
E.g. Lots of satellites
128-byte wrap-around
Very much depends on data within packets
These messages are disabled by default
If the 6th byte had a specific value, then it could cause a crash
Normal configuration may actually create this
The ReceiverStatus message can be >128 bytes
Will also increase some buffer sizes
Plane 3.8.2 will go out soon to cover this
Philip has notified PH2 users and suggested not flying with SBP at this point
09:23am - Plane future
3.8.2 will have GPS fix
Mavlink command set home vs origin
Auto-trim at boot has already gone away
09:26am - Amilcar
10,000 euro copter at 1m above ground
Low-altitude fence!
Dronekt-python control loop
High-quality GPS
Tlogs to sync stuff in real time
Nice to have nav coordinates moving around
Origin never moves
Except origin’s altitude
Baro drift thing
Home should only be moved explicitly
09:29am - SafeRTL/SmartRTL
A couple of people have tested it overnight
Still have some pruning issues
Big improvement in latest version is that it deals with overlapping loops better
Small loops in big loops are not a problem
Overlapping loops are a problem
CPU load has been reduced a lot
Batches of points as they come in
50 points == simplify
Then a loop detection
Then doesn’t touch them again
500 points meant a lot of CPU
Special thanks to Luis for the testing and video and mission file
Luis thinks we should discard points e.g. static, doing loiter
Randy points out the existing parameters should keep addition of points down
Name still needs determining
SmaRTL?
SmartRTL?
SafeRTL?
09:38am - Rover
Pivot turning may require both treads to spin at a high rate
To overcome friction
Tridge’s pivot steering used to allow maximum
Randy cut it back to 50%
To avoid saturating motors when driving normally
09:40am - tridge and Plane
Test aircraft was the FireFly Y6!
https://www.youtube.com/watch?v=ulHYegkfSEE
Original motivation for tilt-rotor support was this aircraft
Prototype was in a long time ago
Thanks to A Trench for the airframe
Problem with second landing
Aerodynamic effect?
Weathervaning backwards?
Our custom URL stuff has disappeared a day after it was created (and tested working)
Another bug in plane where it descends after first waypoint in mission
Tridge thinks he’s found it
Line segment in 3D space which it is trying to fly
The height of this line segment when we’re at a negative offset in the path proportion between waypoint 2 and 3 then it tries to fly at a height lower than the first point in the leg
Height demand goes way down
Tridge thinks he can fix this
I2c compass support
Already in Plane 3.8.0
Randy to determine if it goes into Copter
Next Plane release soon
09:50am - APSync
Something out with APWeb in it
Hopefully late this week or early next
Edison-based board is still in demand
Probably because the price is close and it might be a feature you want
Edison will be released on old script rather than the process Lucas suggested
Multi-camera support will probably require change
10:02am - Issues
https://github.com/ArduPilot/ardupilot/pull/6702#issuecomment-326677986
landing/terminating vehicle
Issue with documentation / stuff appearing when it shouldn’t
Peter has commits that fix the documentation issue
https://github.com/ArduPilot/ardupilot/pull/6819/commits/53d0e48ff4c3fae8b4172a6e6662cdaa5b2c0610
https://github.com/ArduPilot/ardupilot/pull/6819/commits/9bc900c83c174096a9c231ff626440b0e01bcbae
Tridge suggests just linking to the wiki in the descriptions is good for now
10:08am - Philip and discussion of separate build for cube
Parameter reset on current builds can set params to values which are bad for the specific autopilot
Separate builds for cubes would allow parameters to be correct
Tridge is happy with this
.apj?
Rename directories on the firmware servers from PX4 to STM32_NUTTX
Special ones for the Cubes
[10:18 AM] To Weekly devcall: Would be nice to have the GCS update the manifest.json to make this transition easier
Reusing the builds is important
If only parameters are changing, most of the things should be reused
Solo parameters
10:23am - Randy
Sent out some partner resubscribe emails
Feedback so far has been good
10:24am - Interdrone
Presentation on Wednesday night
Craig needs slides
What’s been going on with ArduPilot for the last year?
Please send him media!
Jaime has live-streaming of stuff covered
10:25am - event in China?
21st through 23rd of October
Should go into AP calendar
Canberra again in February?
Craig, Randy and Jani interested + a general clamour
Proposals for travel costs?
Yes, a matter of how much can we afford
If 10 people applied that would be a year’s funding!
If people are sponsored by a company then that company should pay…
Limited funding
Mr Rowse:
California to Shauming $470 return
Need to ask club if we can use the field
Dates last time were because Randy was in town
But the time’s pretty good
So February/March
There’s a linux conference in Sydney…
Unconference-style again
Need to decide scale
How many users get invited
10:32am - tradheli wiki?
Needs work!
Would be nice to have someone step up and fix this up with more modern configuration procedures
Tridge flew Heli on the weekend
Tridge might do a video in realflight of setting up a heli
Luis has offered to add words around the video
Tridge never uses MP, ‘though
Apparently wizards are “out” on Facebook…
Defaulting to advanced view for everything in MP seems to be a desired thing
Jaime dissents
Commercial point of view he doesn’t want users to fiddle
Parameter screen also needs a lot of help
Very, very long to load