ArduPilot / ardupilot

ArduPlane, ArduCopter, ArduRover, ArduSub source
http://ardupilot.org/
GNU General Public License v3.0
10.99k stars 17.52k forks source link

Copter: 4.1.0 issues list #16478

Open rmackay9 opened 3 years ago

rmackay9 commented 3 years ago

This is a list of known issues especially blockers for the 4.1.0 copter release. (also see Rover-4.1, Plane-4.1, Wiki-4.1, 4.2 issues)

Issues to be resolved asap:

Nice to have for stable release:

Existing 4.0 issues:

Reports requiring investigation:

SCurve known issues:


Unresolved issues that will not be resolved in time for 4.1.0


Resolved issues that will be included in future releases (but not 4.1.0)


Resolved issues:

SCurve resolved issues:

andyp1per commented 3 years ago

@rmackay9 from my perspectve its pretty important that https://github.com/ArduPilot/ardupilot/pull/15984 goes in

IamPete1 commented 3 years ago

I think we should also bring scripting up to date with master. Mostly just bindings and one or two helpers.

rmackay9 commented 3 years ago

@IamPete1, 4.1 will branch from master so it will have everything in master.

amilcarlucas commented 3 years ago

I vote for GPS ordering bugfix and AC_fence support PRs

rmackay9 commented 3 years ago

@amilcarlucas, @andyp1per, I've added two PRs mentioned to a new "before stable release" section. We can expect the beta period will be up to 3 months (depending upon how quickly we investigate and resolve issues) and we often rebase on master for the first few release candidates so I think there is time to get these into the stable release even if they don't make it into the first beta release.

@amilcarlucas I wasn't sure what the AC_Fence PRs are but in any case, if we can get them into master then they'll go out with 4.1.

amilcarlucas commented 3 years ago

The AC_Fence stuff is https://github.com/ArduPilot/ardupilot/pull/15288

joshanne commented 3 years ago

@rmackay9 #15288 (AC_Fence for Plane) is in a state ready for people to look at, review, comment, etc.

pompecukor commented 3 years ago

@rmackay9 @IamPete1 @pkocmoud With the help of Peter Hall, we will try to investigate the Pixracer Pro unhealthy gyro issue on 4.1 (3rd item on the above list). Peter has suggested to see if turning off/on each IMU will help identify the source. I will do that tomorrow and report here.

anbello commented 3 years ago

I think also #16842 should go in.

amilcarlucas commented 3 years ago

@rmackay9 the Fence has been merged now. 4.1 will have low altitude fences, and ArduPlane Fences will work in QGC.

JoshWelsh commented 3 years ago

Hi @rmackay9 I see that the S-Curve "climb-away" has been marked as resolved, but I was curious if there was a discussion somewhere that detailed what the fix/finding was? Thank you!

rmackay9 commented 3 years ago

Hi @JoshWelsh, the issue was a rally point had been setup with an altitude of 90m so because that point was closer than home the copter started climbing to get to that alt (or maybe 15m above that alt). I reproduced the behaviour both with 4.0 and 4.1 and it's unchanged. Climbing to 90+ meters is scary if you're not expecting it but the copter would have been OK and would have gotten to the rally point safely.

peterbarker commented 3 years ago

On Wed, 14 Apr 2021, Josh wrote:

Hi @rmackay9 I see that the S-Curve "climb-away" has been marked as resolved, but I was curious if there was a discussion somewhere that detailed what the fix/finding was? Thank you!

User (me) error :-)

JoshWelsh commented 3 years ago

Hi @JoshWelsh, the issue was a rally point had been setup with an altitude of 90m so because that point was closer than home the copter started climbing to get to that alt (or maybe 15m above that alt). I reproduced the behaviour both with 4.0 and 4.1 and it's unchanged. Climbing to 90+ meters is scary if you're not expecting it but the copter would have been OK and would have gotten to the rally point safely.

Great thanks Randy! Great to hear it was an easy find!

On Wed, 14 Apr 2021, Josh wrote: Hi @rmackay9 I see that the S-Curve "climb-away" has been marked as resolved, but I was curious if there was a discussion somewhere that detailed what the fix/finding was? Thank you! User (me) error :-)

You should see the stuff I've done! Glad you got the copter back.

lthall commented 3 years ago

17345 and #17399

Hwurzburg commented 3 years ago

-[ ] new bootloader fails on trad heli firmware

andyp1per commented 3 years ago

MOT_PWM_MIN/MAX not being set to 1000/2000 when MOT_PWM_TYPE is set to a DShot type, but SERVOBLH* parameters not setup. Looks like a regression from 4.0 is now fixed

andyp1per commented 3 years ago

Extant Dshot issues:

lthall commented 3 years ago

low LOIT_ACC_MAX leads to incorrect loiter speed (issue) Fix here: https://github.com/ArduPilot/ardupilot/pull/17815 Guided with high rate position target updates don't work (see issue #17795) Fix here: https://github.com/ArduPilot/ardupilot/pull/17539

andyp1per commented 3 years ago

RunCam split 4 video start/stop not working discussion

rmackay9 commented 3 years ago

@andyp1per, txs for this. Maybe keep a single list in your comments from 22 days ago (a few comments higher in this issue..)

lthall commented 3 years ago

18190 to address the lean angle limit for auto.

rmackay9 commented 3 years ago

@pompecukor @IamPete1 @pkocmoud,

I've marked the "mRo Pixracer Pro gyro & accel health failures" issue as resolved above because in a quick test by Tridge we couldn't reproduce the issue. This is a very old issue from early on in the beta and I haven't heard anything about it since so I suspect it was a one-off problem with just one user's autopilot or we've resolved the issue as part of other fixes in subsequent beta releases.

Of course I'm very happy to re-open the issue if anyone can reproduce it with a newer beta.

hendjoshsr71 commented 2 years ago

I checked off one that i fixed via, https://github.com/ArduPilot/ardupilot/pull/19263 "AP_Periph HAL battery monitor default for Matek board (discussion, discussion2)"