Autopilot won't fly in a straight line..

[Paul Ramsay 8]

Please login to display this image.

 

So as you can see this is my flight from Sydney to Brisbane. I have a 51 knot crosswind from 275 with my current heading of 354. However it keeps swinging left to right. It's actually must worse in game. It's pretty constant, the above image shows my flight on Event tracker.

 

 

[jfwharton 6]

I have had a similar experience although not all of the time.  Most of the time after takeoff when the aircraft is setting up the initial first leg intercept it overshoots, over corrects, back and forth for a while until it finally settles down.  After that most of the time it does well but there is an issue with this autopilot over correcting.  Even when the CDI is only slightly left or right of course the autopilot makes 20 to 30 degree turns to correct instead of small 10 degree or less turns.  Also, it waits until the last minute to turn on course which causes the overshoot and the toggle back and forth.  It would be very nice if an update could be done to the autopilot to correct this issue.

[JeremyGolf70 2]

Actually I have experienced similar issue when cruising with Working Title CJ4 in crosswind. I guess it is because of the wind turbulence simulated by MSFS since during the cruise, the plane was really unstable. 

[Paul Ramsay 8]

update please...

[kvk1 17]

Please login to display this image.

Please login to display this image.

[JRBarrett 675]

3 hours ago, kvk1 said:

Please login to display this image.

Please login to display this image.

I don’t see anything particularly wrong in your screen capture, other than the fact that the autopilot is not in 1/2 bank mode, which should have happened automatically climbing above 31,600 feet, and would be annunciate at the top of the PFD.

[kvk1 17]

58 minutes ago, JRBarrett said:

I don’t see anything particularly wrong in your screen capture, other than the fact that the autopilot is not in 1/2 bank mode, which should have happened automatically climbing above 31,600 feet, and would be annunciate at the top of the PFD.

 

That constant left-right weave between DOFFY and CAMJO is normal?

 

I wanted to include as much visual detail as I could, in case it pertains to any solution.

 

The right picture includes the weather/wind/etc. coming out of the problem area for the LNAV.

[JRBarrett 675]

2 hours ago, kvk1 said:

 

That constant left-right weave between DOFFY and CAMJO is normal?

 

I wanted to include as much visual detail as I could, in case it pertains to any solution.

 

The right picture includes the weather/wind/etc. coming out of the problem area for the LNAV.

Can’t tell from a still screen shot, unless you intended to post a video.

[JRBarrett 675]

On 3/23/2021 at 7:40 AM, Paul Ramsay said:

Please login to display this image.

 

So as you can see this is my flight from Sydney to Brisbane. I have a 51 knot crosswind from 275 with my current heading of 354. However it keeps swinging left to right. It's actually must worse in game. It's pretty constant, the above image shows my flight on Event tracker.

 

 

On most of my flights with a crosswind, once established on a leg, the aircraft assumes a fixed wind correction angle and holds it with almost no variation. But I have seen several reports now (including yours ) of constant S turning with a wind, and it seems to be affected by direction - in particular when the course is very close to north.

 

I am a CRJ tester, but I have no specific insight into the FMS software for the CRJ. But, having written navigation software for my own flight planning uses, I can definitely see where a a course very close to 359 degrees combined with a crosswind could possibly cause instability because of the way that the most common mathematical algorithm used for calculating direction between two waypoints of known latitude/longitude works.
 

If the course passes from 359 to 000 degrees or back, the sign of one of the terms in the equation would oscillate from positive to negative, (or negative to positive) which could cause the correcting back and forth. 
 

I will try setting up a flight plan with a course close to north with a crosswind, and see if I can get the aircraft to to display the continuous S turn behavior. 

[CRJay 427]

As an aside note... The heading/course should never read 000, it should be 360 instead. Not to hijack the thread, but perhaps it has some impact on this issue as well. 

[piotrmkg 20]

I have the same issue, straight leg, 15knots crosswind and plane just sways left and right, even with 1/2 bank on it turnus with 20deg bank angle. FBW320 was swaying left and right for some time, but it never had any problems flying in a straight line in crosswing. I cannot believe it never was tested before release.

[Chaxterium 119]

6 minutes ago, piotrmkg said:

I cannot believe it never was tested before release.

 

I'm sure it was tested. A lot. But you can't test for every scenario and you can't test for every system.

 

I've had issues when the AP can't keep the plane straight but for the most part it's done a very good job. I've noticed that it tends to have issues on sharper turns while in cruise with a strong wind. Aside from that it seems to do pretty well.

[Braveheart76 0]

I have had a similar experience although not all of the time when the plane have autopilot mode and NAV was selected.  Most of the time after takeoff when the aircraft is setting up the initial first leg intercept it overshoots, over corrects, oscillation (Wobble) from side to side and more when there is cross wind.

I will like to solved or an update could be done to the autopilot to correct this issue?

[StanTGM 38]

Same here: todays early morning flight from Riga to Helsinki was with an 90 knts (!) crosswind at FL260. In NAV, the CRJ began to oscillate as described above. Changing to HDG solved the problem (I had to make some small adjustments along the way to get the crosswind compensation right). Once I descended below FL200 wind calmed down a bit and I switched back to NAV, which now again followed the flight plan without problems.

 

I think this oscillation behavior indeed happens only with very strong crosswinds. Maybe the programing logic  simply can‘t cope with it  

Would be very interesting to hear from a RW CRJ pilot if the FMC in the real plane has the same limitations.

 

[CRJay 427]

13 minutes ago, StanTGM said:

Would be very interesting to hear from a RW CRJ pilot if the FMC in the real plane has the same limitations.

 

Nope. The main 'issue' with the RW CRJ autoflight is undershoot or overshoot on an ILS intercept due to wind. That is usually very minor. Two days ago I was flying nearly direct North (track 002 to 004) with an 87kt direct crosswind and it tracked rock solid.

[renatodmc 13]

2 hours ago, StanTGM said:

Would be very interesting to hear from a RW CRJ pilot if the FMC in the real plane has the same limitations.

It does not. Every single jet I’ve flown (2 with Pro Line 4) does the job precisely no matter the wind. 

[DGH 0]

It happens at certain crosswind ANGLE.

 

Also not in heading mode ... just in nav mode (calculating crosswind to stay on course .... so wind related ) 

 

 

[Gerwil 7]

I can confirm that it happens when flying in a northern direction. I did severeral flights between EDDK and LFKB. Going south the CRJ flies the LNAV nicely, the return flights always results in a lot of S turning. Altitude doesn't matter that much. Shortly after departure, while still on the SID it started to lose track already. It happened with a almost headwind and also slight crosswind of 42 knots.

[nfinger 3]

I have seen this, left and right banking issue with large crosswinds components on climb, a few times and the way I was able to correct it was to override the next WP so CRJ flies direct from current position to next WP.

[Nepomuk 10]

This is really funny, the same issue was there with Hans Hartmans ATR72 and was never solved:

Strange LNAV Problem - ATR Discussion Forums - Page 1 (flight1.net)

 

[CRJ Pilot 15]

Yes I can confirm. The autopilot is broken. It won't hold altitude correctly - about 800ft out. Speed hold during v sync is not controlling the throttles correctly. And autopilot is constantly singling left to right every few minutes in FMS 1 hold. 

 

Needs a patch I'm afraid. 

[JRBarrett 675]

1 hour ago, CRJ Pilot said:

Yes I can confirm. The autopilot is broken. It won't hold altitude correctly - about 800ft out. Speed hold during v sync is not controlling the throttles correctly. And autopilot is constantly singling left to right every few minutes in FMS 1 hold. 

 

Needs a patch I'm afraid. 

The throttles in the real and simulated CRJ are 100 percent manual they are not controlled by the autopilot in any way. I have seen some instability in course tracking, but in 200+ hours test flying the sim version I have never had any problems with altitude hold.

 

I have seen the displayed altitude suddenly jump by plus or minus 250-300 feet when using Live Weather due to sudden shifts in the barometric pressure being injected into the sim environment, but that is an external issue - not directly related to the CRJ

[BremyG 2]

Similar observation on a flight today: slow but significant roll oscillations in cruise (up to +/-8 degrees) as the autopilot slowly hunted for the proper wind correction angle. This was observed with crosswinds in cruise of 40-70 knots along the route. With a 40 kt crosswind, autopilot roll oscillations were less frequent, but with a 70 kt crosswind, the autopilot was constantly rolling the jet back and forth, oscillating around the desired ground track.

 

 

Please login to display this image.

[Chaxterium 119]

17 minutes ago, BremyG said:

Similar observation on a flight today: slow but significant roll oscillations in cruise (up to +/-8 degrees) as the autopilot slowly hunted for the proper wind correction angle. This was observed with crosswinds in cruise of 40-70 knots along the route. With a 40 kt crosswind, autopilot roll oscillations were less frequent, but with a 70 kt crosswind, the autopilot was constantly rolling the jet back and forth, oscillating around the desired ground track.

 

 

Please login to display this image.

 

I've had some luck with engaging heading mode, letting the plane settle down, and then putting the plane back on an intercept heading. Once everything is scared away I re-engage NAV mode and it seems to work well. If this happens again give that a shot.

 

I know this doesn't solve the problem but it should help to allay the symptoms until there's a fix from Aerosoft.

[JRBarrett 675]

27 minutes ago, BremyG said:

Similar observation on a flight today: slow but significant roll oscillations in cruise (up to +/-8 degrees) as the autopilot slowly hunted for the proper wind correction angle. This was observed with crosswinds in cruise of 40-70 knots along the route. With a 40 kt crosswind, autopilot roll oscillations were less frequent, but with a 70 kt crosswind, the autopilot was constantly rolling the jet back and forth, oscillating around the desired ground track.

 

 

Please login to display this image.

This definitely appears to be most common in a situation where the desired ground track is on one side of north and the required wind correction angle is on the other side of north, which appears to be the case in your screenshot. 
 

Because of the way that the standard mathematical algorithm for calculating a great circle bearing works, I can see how this might happen, but I need to do some more research while in flight.

 

During testing I have had several east-west flights with 100 knot+ crosswinds where the aircraft held a rock-steady wind correction angle without any oscillation at all.

 

I have a northbound test scenario all set to go, but because of the current “packages out of date” problem in MSFS, I can’t enable Live Weather at the moment (or any other online streaming for that matter). I am hoping this will be fixed on Tuesday when the WU 4 update is released.