rroche
November 24, 2021, 6:00pm
1
December 01, 2021
Join us
Agenda
Community Q&A
Project Updates
High priority queue
In-Depth discussions
Community Q&A
@JulianOes Found an issue that sometimes makes CI hang: If the simulation gets killed in the wrong moment then the main deamon gets killed but the individual processes started from the startup script still try to run and make CI time out. Making the startup script fail with the first failing command seems to already fix it but @bkueng proposed to kill the entire process tree instead to not have the entire startup script fail in case a command once fails. We should revise the exit strategy of SITL. Instead of a hard kill exit everything cleanly.
Rarely the hardware test rack fails on a submitted pull request.
Servo calibration user experience issue on master found by @Paul_Riseborough
@dagar is checking the suggestion. The way to trim will change soon with the new actuator configuration.
opened 12:27PM - 30 Nov 21 UTC
bug
fixedwing
**Describe the bug**
@priseborough was setting up a flying wing on master and wโฆ as asking me how to calibrate the servos specifically why there are no `PWM_MAIN_TRIMx` parameters while there are `PWM_AUX_TRIMx` parameters.
**Additional context**
I did a quick debug and am pretty sure it is a race condition were the mixers are not loaded yet when the first parameter update is done. So the condition that was added here prevents the parameters from being discovered:
https://github.com/PX4/PX4-Autopilot/commit/089c962d9208b3616ba505ab927fa399554035d4#diff-ff52cb1224c92cc672316dab2f93a98e7db09fb08ccdfbb9fb0015e8d86f1387R859
I tested and after changing a parameter and reloading the ground station while the autopilot is still running I can see the parameters.
I'll create a pr to remove it but I'm not sure about side effects. @dagar would be nice if you could check.
**To Reproduce**
Steps to reproduce the behavior:
1. Flash pixhawk 4 with latest master (2021-11-30 01db98fccf7944a98f5ee8b14693a58d1fca34be)
2. Boot with an airframe e.g. quad_x 4001
3. Look for parameters named `PWM_MAIN_TRIMx` in the editor
4. See all the `PWM_MAIN_{DISx,FAILx,MAXx,MINx,REVx}` but no TRIMs
**Expected behavior**
`PWM_MAIN` and `PWM_AUX` should offer the same parameter set.
**Log Files and Screenshots**
![image](https://user-images.githubusercontent.com/4668506/144046478-b1353371-73d8-4e9d-aeb0-7b023ab2341f.png)
Project Updates
@MaEtUgR Proposing to have data link failsafe available in manual modes. Discussion results in not having this enabled by default and having a bitmask to opt into failsafes that are not hard requirements.
PX4:master
โ PX4:data-link-loss-all-modes
opened 02:20PM - 01 Dec 21 UTC
**Describe problem solved by this pull request**
When a pilot flies:
- manuallโฆ y in position mode
- BVLOS
- conventional remote e.g. SBUS receiver
- tablet to running QGC and relies on either QGC telemetry or even a video feed
and then he loses the data link he's basically flying blind but there's no failsafe reaction even if he enables the data link loss failsafe using `NAV_DLL_ACT` because the data link is not considered a strict requirement.
**Describe your solution**
I enable executing the data link loss failsafe also in manual modes if it's enabled.
For the use case where the pilot takes over in a manual mode during the data link loss reaction I had to introduce a flag at least temporarily. I want to solve this by allowing an event based reaction in the state machine where in this case only if you lose data link a reaction gets triggered and not all the time while you're in a mode and have no data link.
**Describe possible alternatives**
See last sentense above. It definitely needs improvements.
**Test data / coverage**
This was tested in simulation, in SIH with a lot of cases and flown in real world.
High priority queue
Discussion based on board: High-Priority Queue ยท GitHub
In-Depth discussions
For smaller groups expanding technical discussions, stay until the end and follow up.
If you have any feedback or corrections please comment on this post.