F28 enters spirals when trying to capture radial VOR beam
-
Hi everyone. My Fokker is broken and I'd love some help. I have two issues with the autopilot that make it unplayable to me.
Firstly. The aircraft misbehaves in BEAM mode, which is now completely useless. Attempting to capture a radial (any range, any intercept angle) the aircraft will confirm A/P capture and then start a turn to the right or the left (depending on whether I am north or south of the VOR, I think?) which never ends, sending me spiralling. It looks as though the AP accepts the first roll command to capture the radial, but it won't accept the second (to turn into the radial) or the third (to stop the occasional overshoot) roll commands.
I have found one other person who reported identical problems, but no mention here on the forums that I could see.
https://www.reddit.com/r/flightsim/comments/152c2mw/justflight_f28_vor_issue/At other times the plane joins the radial correctly but merely fishtails. This might be related to the second issue, which I think corrupts some roll inputs.
You can see a few slides of the fishtailing and how it looks on the panel here at imgur.com/a/JaN85oC but it's probably not very out of the ordinary for you.The issue was present before 1.2 and remains after 1.2. I put it down to an initial bug and put the addon away, but now that it persists after 1.2 I'd like your help on the matter.
The second issue presents itself with all modes ROLL, HDG, and BEAM. The aircraft will roll back and forth without ever stopping. When making horisontal turns the aircraft (judging from the yokes and judging from the view outside) will pull to the side, then it comes back just a tiny bit, before pulling to the side again. It feels like a cyclical event over the ordinary curve. Rudder is not visibly affected, but aileron and elevator trim is. Can't detect if aileron trim is also influenced. I think the yoke is moving at roughly 0.5-1 Hz. There were 70 lefts-and-rights in 60 seconds last I timed it. Or to put it another way, it had about 35 right inputs in 60 seconds.
For the roll oscillations I've found no solution. For the radial spiralling however, the GPS is sooort of a solution.
To make the product work* I have to
-Enable the GPS panel
-Transfer NAV control (CDI button on GPS panel. GPS now shows VLOC.)
It is now possible to tune VORs and use BEAM mode to fly the radials.*However in this mode it's not possible to switch between NAV1 and NAV2. If you try to do that the HSI will spazz out and control reverts to GPS. This is not a tenable solution.
I think this has to be an issue with your automation. If I put a GPS into the aircraft and use that to steer, the aircraft turns smoothly. According to your changelogs the aircraft should not have many issues remaining anymore with jumping between AP with or without the GPS present. However I encounter a whole heck of a lot of them if I activate/deactivate GPS navigation, so I wonder if 1.2 simply hasn't been properly installed. Or if there are panel save state configs somewhere that persist?
I appreciate that the first answers to this inquiry might be regarding the FD, but I'm not confusing the two. In fact the FD often knows how to get onto the radial whereas the autopilot never will.
I have tried to
uninstall, repair-install, and reinstall the addon several times.
remove my controller (Saitek X52 Pro). Issue persists with keyboard only.
run the addon without any other community addons
change assistance options (gyro drift off, turbulence realistic-medium-low, no AI pilot assists)
enable all aircraft options in the EFB
disable all aircraft options in the EFB
change Nanovg for XML changes to OFF. It's back to ON now.My Flight Model settings are all 100%.
Cockpit Interaction System is Legacy.
EFB says I have version 1.2Finally I'd like to commend you on what is obviously an amazing addon, and express my admiration for your commitment to this plane's recreation in the virtual world. Judging by most people's experience it is a solid product.
Thank you in advance for any help.
-
I'd like to add this illustration of another attempt I made to capture the phenomenon.
Intercepting a radial with a low angle (approx 25 deg), at 70nm, outbound, I see the F28 roll left, cross the radial at an angle of more than 90 degrees, and then roll right. It continued to roll right, its apex at 1.5 dots off the line, until it crossed the radial, once more at an angle of more than 90 degrees.
Then I intercepted a radial with a HIGH angle (approx 60 deg), at 80nm, outbound. Once again the F28 rolls left. It begins to smooth its roll (red section) as though it has intercepted the radial (despite doing so in the wrong direction). But this time it just -barely- overshoots the radial, putting the course at more than 180 degrees. It begins to roll left again, its apex at 1.5 dots off the line, until it crosses the radial once more. This time, you notice, with more than 90 degrees but less than 180 degrees to the course. It rolls RIGHT, hits its apex at 1.5 dots off, and then continues right. It begins to level off (red section), but does not cross the radial this time, instead spiralling in a nonstop right turn.
These are not exhaustive tests. I can't be bothered spending my whole day trying :D Still I think the difference between fishtailing and spiralling is linked to whether the plane crosses the radial at more or less than 180 degrees to its course.
-
Not disputing what you've experienced, however, I watched V1-simulations on YouTube, he's a Airbus pilot but used to fly pretty old aircraft with really rudimental autopilots like the Fokker has. He's streamed a lot of the Fokker.
I remember him saying some the oscillation in beam mode wasn't uncommon in old skool autopilots so much so that pilots would use the HSI and fly it in heading mode, using the HDG to track the VOR, babysitting it with adjustment as you went along as the AP would just oscillate along the VOR otherwise in beam mode.
Not sure if the update has more realistically replicated the AP to be even more rudimentary (This is a 1960s Autopilot) or is it a bug.
I've linked his video, was with the release version of the Fokker by the way, maybe different in v1.2
-
@spurious said in F28 enters spirals when trying to capture radial VOR beam:
Not disputing what you've experienced, however, I watched V1-simulations on YouTube, he's a Airbus pilot but used to fly pretty old aircraft with really rudimental autopilots like the Fokker has. He's streamed a lot of the Fokker.
I remember him saying some the oscillation in beam mode wasn't uncommon in old skool autopilots so much so that pilots would use the HSI and fly it in heading mode, using the HDG to track the VOR, babysitting it with adjustment as you went along as the AP would just oscillate along the VOR otherwise in beam mode.
Not sure if the update has more realistically replicated the AP to be even more rudimentary (This is a 1960s Autopilot) or is it a bug.
I've linked his video, was with the release version of the Fokker by the way, maybe different in v1.2
Hi Spurious
Thank you for your reply.In my tests this issue goes far beyond ordinary oscillations and I don't think we need to babysit an intercept of 20 degrees. Here is Just Flight's pre-1.2 demonstration flight, timestamped where the pilot puts his aircraft on an intercept heading of 45 degrees. https://youtu.be/_yf7IKD9zYk?si=9a3fsin9CwYWCglN&t=320
Which has been a perfectly sound method since before this aircraft existed. :)After the intercept the aircraft in the video overshoots, it establishes itself on a new intercept at roughly half that angle, and is (by the magic of video editing) then back on course. Several more intercepts occur during the video, such as at 10:30 when it intercepts a radial at 20 degrees. Never does the aircraft turn to beyond 90 degrees of the radial. Nor is the yoke flapping about.
If I put myself square on a beam I am able to follow it. However if there is any meaningful amount of roll in the intercept, the AP goes haywire. I am able to intercept at angles of less than 10 degrees, but not 20.
You can see at imgur.com/a/DzCcTYS that I have annotated a sequence of tests. If you have the time, please make the test for yourself: Can you intercept a radial at 35 degrees as in the 4th and 5th image there? My aircraft rolled left to intercept it at a 70 degree difference, then rolled right and blew right through the radial again.
On the video that you linked, at 1:05:22, he intercepts a radial. He engages BEAM mode at approx 40 degrees during a turn. The AP tracks this beautifully. The "hunting" then commences with less than 5 degrees of difference. What he is remarking upon regarding smoothness is correct but also, when you are not concerned with your reputation among colleagues, superfluous.
In fact at 1:12:50 he has some issues that aren't too dissimilar from mine.
Unfortunately I must be an edge case in some respect otherwise more people would have this issue.
-
@scaflight said in F28 enters spirals when trying to capture radial VOR beam:
You can see at imgur.com/a/DzCcTYS that I have annotated a sequence of tests. If you have the time, please make the test for yourself: Can you intercept a radial at 35 degrees as in the 4th and 5th image there? My aircraft rolled left to intercept it at a 70 degree difference, then rolled right and blew right through the radial again.
(quick edit after the hour-mark has passed: You can see at imgur.com/a/DzCcTYS * that I have annotated a sequence of tests. If you have the time, please make the test for yourself: Can you intercept a radial at 35 degrees as in the 5th and 6th image there? My aircraft rolled left to intercept it at a 70 degree difference, then rolled right and blew right through the radial again.
*imgur buggered the order. The second image should be the last one in the series.) -
Check your content manager (or whatever its called in the english version) for updates and install them.
Updating the Working Title GNS fixed this issue for me.
Quick edit: I know it might sound weird updating the GPS to be able to fly the aicraft withouth it a GPS, but this did fix the beam issue for me so its worth a try.
JF Support also reckoned it they've had more reports of nav issues coming from an outdated GPS mod on severall aircraft.
Cheers Luc
-
@kenzo You crazy f********* bastard that solved everything except the oscillating roll/tipsy yoke.
Imagine the trouble I'd have saved if you'd updated your thread eh???? ;D /S
Cheers mate
edit: Someone answered me in the reddit thread -immediately- after I posted, I just didn't see it. Oh I could've spent the whole DAY flying arghhhh
@JF the issue with the tipsy yoke remains.
Maybe ought to put somewhere prominent that you're using libraries or something from WT. 6_9 -
@scaflight you’re welcome! I’ve had some though days after the release of this amazing addon finding that out. I should’ve posted the culprit at my own thread after finding it out for others to find like you.
about the oscillations in roll, if they are what I think they are, that has been introduced in the latest 1.2 update. I’m also affected by, in my case, very tiny roll movements left and right. I’d rather not have it, but it’s better than the pitch waves at certain CG values in 1.1.
-
@kenzo @scaflight
Despite the reported "SaveState" issue i haven't seen this issue in many of my flights.I'd say its either an issue with your install or hardware.
Are you using the Windows feature to de-install the aircraft ? If not do soYou may also make sure the nothing (but the work folder) is left in :
C:\Users\YourName\AppData\Local\Packages\Microsoft.FlightSimulator_8wekyb3d8bbwe\LocalState\packages\justflight-aircraft-f28aca: "M-Files"
Oliver
-
Hi all,
Just to clarify a few things that have been mentioned in this thread.
The VOR "spirals" issues isn't something we have ever been able to replicate, but I have personally seen a very small number of users reporting similar issues. As mentioned above, making sure the Working Title GNS530 is fully up to date should solve the issue. As much of a chore as it can be to keep everything up to date, we would recommend always keeping any GPS mods fully up to date to minimise any chances of issues appearing. It's probably the number one cause of autopilot issues we get through our support channels across multiple aircraft.
The autopilot roll oscillations are a known issue when no GPS is selected (weather radar visible). We do have a solution for this that is currently in testing, but we need some time to ensure the fix works across all possible scenarios. As soon as we are confident in the fix, we will push a new update out for the F28 Professional. The difficulty we face with the F28 (and our PA28 range) is when the Working Title GNS530 is selected, it tries to take control of and override all of our autopilot code, so we have to effectively code the aircraft systems and flight model so that it works with multiple autopilot systems, which gets very complicated very quickly! One last-minute improvement to the Autopilot behavior with the Working Title GNS530 selected has likely caused this knock-on effect with the basic Autopilot functions which went unnoticed ahead of the v0.1.2 release. As a short-term solution, you shouldn't have any roll oscillations when the Working Title GNS530 is selected on the EFB.
If anyone is experiencing any issues with State Saving in v0.1.2, if you can contact Just Flight support we should be able to provide a fix for that: https://www.justflight.com/support
Mark - Just Flight
