I’ve been seeing this problem for a while, but it’s not reliably reproducible, it’s random, and I don’t have enough testing evidence to submit it as a bug. I wondered if anyone else has seen it.
While on autopilot, in pretty much any aircraft (all avionics types except steam gauge, since I don’t fly steam), the aircraft will suddenly bank to the left. This is not a course change or a waypoint de-synchonisation issue (though I’ve had those too). If left uncorrected, the aircraft will continue to roll left (always left, irrespective of p-factor, including in non-piston aircraft) until the nose drops and then it will, if left uncorrected, descend into the terrain. All while the autopilot is indicating that it’s still on. If I correct it straight away using the yoke, and if I do so gently enough to not force AP off in aircraft that can do that, then the autopilot will resume proper course and continue to work (until it happens again); so it’s not that the AP is becoming deactivated while showing as on.
This happens mostly in HDG mode, but also sometimes in NAV mode. It has been a left bank only up until after SU13, but now I’ve also seen it push down the nose until the aircraft is pointing straight down.
I’ve checked all my controllers, control bindings etc, and I can’t find anything wrong there. This just happens. It’s similar to the ‘autopilot taking you around in circles’ issue in early SUs, but it seems like the autopilot stops correcting the attitude of the aircraft until you force it to by either going AP disengage / re-engage, or manually bringing the attitude back to what it should be.
It’s random, but once it starts happening on any flight, it tends to happen several times in short succession, and then may just stop for the rest of that flight, or come back later.
Has anyone seen anything similar? Any ideas? Or am I going to have to go the whole hog and remove all mods etc to test it until I get submit an actual bug report? Hopefully not
I’d have to just video my flights until it happens - could be immediately on turning on AP, could be not at all for hours. I’ll give it a go but it might take a while to capture it.
Edit: it is basically the same behaviour you’d see if you disengaged the autopilot and didn’t trim out or touch the controls. So I wonder if it’s somehow that the autopilot code thinks it’s changing the attitude of the aircraft but actually isn’t. But as I mentioned, if I immediately correct when I see it happen and bring the aircraft back to level flight, with the AP still on, the AP goes back to doing what it should be doing.
My recollection is that this only began with SU11 for me, but I can’t be 100% sure.
Understand it could be a long video, so as an alternative, can you do a screenshot when you start, including your flight plan showing on the map, and then push active pause when the issue starts, and grab another screenshot? Might take a few screenshots at the start of flights until you hit a flight where the issue happens. That would give us a beginning state and a shot of anything different at the point it happens.
Regards
I seem to remember this happening to the A320, which use to keep turning to the right with the autopilot on they have fixed it now. Just rename the community folder and fly and see what happens, but I would also check all my USB connections as well as the hardware it self, because some time it is difficult to tell faulty hardware when it is sending erratic feed/signal to the MSFS app.
Interesting. I had a similar unexplainable autopilot incident in the A2A Comanche today. I was flying over mountain ranges in Alaska on a clear day with little wind in HDG mode. The plane kept turning off course and not to where I had the heading bug set. Thinking I had a flight control failure, I disengaged AP and hand flew it and it was fine. Back on AP and it had a mind of its own.
A few hours later I flew the same route on AP and it was fine. I have no idea what had occurred. GPS jamming by Sim terrorists? Solar flares? Strange.
Have had similar to this issue frequently, had actually posted about this but was moved to a specific category forum.
Plane flying great then suddenly starts going in circles at a very steep bank or circles then a spin out of control to the ground.
Always on auto pilot and I mainly fly an IFR route.
I use the stock A320 mainly and is a shame as had been having trouble free flying a month or so ago.
I use Xbox series S.
Interesting to hear there are more people who have this problem or similar.
I am as certain as I can be that I don’t have anything wrongly mapped to the aileron or rudder axes. The last time it happened I was flying without a flight plan, so I don’t think it’s a case of the aircraft trying to return to a waypoint out of sequence, which I have seen happen.
I’m going to do the testing steps required to be able to submit this as a bug, and I will see if I can get it on video.
Update: I’m putting together a video from footage I just shot of three different examples of the bug happening.
Interestingly, when I came up with a random scenario using the C172 and EGCB Coventry Airport, every time I tried it (4 times in total), the problem occurred immediately after activating the autopilot, and I have that on video. Then when I tried a new scenario (Caravan at KSFO), I took off and went onto AP successfully, but within about five minutes the bug happened again, and I have that on video too. Does this suggest it’s somehow location-related? No idea.
All tests done with an empty community folder, nothing other than my yoke bound to any control surfaces (I even removed all the keybinds to make sure it couldn’t be a keyboard issue), and no external applications running on Simconnect.
This has to be an autopilot bug. It seems like when it happens, basically the autopilot stops controlling the control surfaces, and it’s as if the aircraft is untrimmed without anyone at the controls, and it behaves as it would in that situation. That said, I’ve had this happen to me on both the Vision Jet and the HJet, which would not naturally pull to the left.
Once I have the video done and uploaded I’ll make a bug report.
I will do some extensive testing flying with the AP in HDG mode this weekend and reply back. As I read some of these replies I may have experienced the same thing but wrote it off as I must have inadvertently done something that I did not intend.
That’s interesting. I could fly with a monitor on the rudder trim value in Simconnect and see if it changes when this happens to me.
(Obviously the AP should adjust all the trims in order to stay stable, so I’d expect rudder trim to move, but presumably to cause the issue I’m seeing it would have to move a long way, which would be noticeable.)
I experienced similar behavior a while back for months. All my settings within MSFS axis bindings appeared to stay centered - even though I slightly increased the dead zone within MSFS axis settings, the uncommanded roll still happened.
I finally realized that the cause for my issue wasn’t MSFS or any addon:
I use a Thrustmaster Warthog
I had unchecked the deadzone within the Trustmaster software, and that caused inputs to register in MSFS, even thought MSFS didn’t show those inputs when looking at the axis contol settings.
After re-enabling the deadzone in the Thrustmaster software, I could then remove the deadzone within MSFS. I’ve been flying without deviation since.
Interesting. I don’t use any 3rd party software with my controllers other than Spad.neXt and none of the flight axes are mapped through that; and I checked to make sure there was nothing directly in the device profile XML files. I have Thrustmaster kit but I don’t use their TARGET software.
Edit: also, I ran my tests without Spad.neXt running, or any other external software including FSUIPC. All were disabled. Still happened (as you can see in the video).
I don’t “use” the Thrustmaster software either. I only used it to update my firmware so I could use the F-18 grip. I had unchecked the deadzone box when I updated my firmware because I thought, “Hey, I fly helicopters too. I don’t want any deadzone of any kind”
I only ran the software again to correct my error and re-enable the deadzone checkbox for the embedded firmware deadzone.
Hmm. I’ll check. But my TM joystick is generally not mapped to flight surfaces most of the time; I have a profile for it for when I’m flying a sidestick plane like the Vision Jet. Otherwise I use a PFC yoke.
Lots of good ideas on this thread, though… keep them coming. Perhaps we can solve our own problems. But maybe it is a WT issue with the autopilot code. Maybe I’ll report at the WT Discord as well.
