Autopilot Irregularities

[shnt 11]

HDG mode works well, but LNAV is sometimes unable to hold heading.. yesterday I flew 2hours flight and last 30mins It went to snake mode, so I had to use HDG. I assume that its not issue due to a winds when HDG works well in same conditions. 

Please login to display this image.

[metzgergva 210]

the two modes are completely different in terms of correction.

HDG compares the current HDG versus set HDG and by that regulates on a angle difference to be zero.

LNAV compares a track position laterally to defined line. So it regulates by a distance to be zero.

 

The influence of wind and its modeled turbulences is different too. If you are hit from wind speed difference from the right side, the plane automatically counteracts with yawing to the right and helping the AP with the angle correction.

While on LNAV that correction happens too but at the same you get a bit drift to the left and only that is seen by LNAV and starts correcting once the difference exceeds a value.

[BubbaGump954 1]

So is it the nav mode that needs tunning or the winds cause im also going left, right, left, right and its doing it at 30 degree bank angles.

[BubbaGump954 1]

Also the course line for FMS1 bounces left to right it doesn't stay consistently in the center.

[JRBarrett 675]

34 minutes ago, BubbaGump954 said:

Also the course line for FMS1 bounces left to right it doesn't stay consistently in the center.

If the FMS course line shifts back and forth, it would certainly cause the Nav tracking to oscillate, but I have not seen this happen while testing - at least not when established on a path.
 

Significant direction shifts in the upper wind (if using Live Weather) might cause something like this. I have seen some pretty significant sudden changes in both wind speed/direction and barometric pressure when using Live  Weather, but it does not happen every time.

 

I just completed a flight from JFK to DTW at 36,000 feet with Live Weather with a steady 80 knot crosswind coming from the left, and the aircraft held the track with no problem. It did fly at a significant wind correction angle to the track line, but it never varied from the track at all.

 

Do you have a flight plan where this happens for testing?

[JerryJet 14]

4 minutes ago, JRBarrett said:

Do you have a flight plan where this happens for testing?

KCHA HCH VXV KTRI at FL215. The turn at HCH

[JRBarrett 675]

2 hours ago, JerryJet said:

KCHA HCH VXV KTRI at FL215. The turn at HCH

I just completed this flight. Takeoff on runway 33, landing ILS 23. Cruise FL210 at 280 knots. Live Weather gave a 45 knot left rear crosswind on the leg CHA to HCH, becoming a  direct tailwind on the leg HCH to VXV. 

Full disclosure - this flight was done using the update 1 version that will soon be released for all customers - not the original release version you are using now.
 

The airplane began the right turn 4 miles prior to HCH with a ground speed of 375 knots - It turned inside of HCH (which is correct), then intercepted and established on the leg HCH to VXV with two left/right correcting turns at 15 degrees of bank. It was on course centerline by 6 miles past HCH and stayed on it to VXV. (I also added the HMV VOR to the route beyond VXV to set up for the approach). 

 

Other than during the turn itself at HCH, the FMS CDI stayed centered on the leg from CHA to HCH, and HCH to VXV.

 

I am not saying the LNAV is perfect - there is room for improvement, but any changes have to be done with caution and a lot of testing, because the LNAV coding affects many different parts of the autopilot, including lateral RNAV and ILS approach tracking. The golden rule of software development is “don’t fix one thing, and accidentally break something else”.

 

Though I have seen the autopilot sometimes struggle with extremely sharp course changes combined with high airspeed and a very strong tail or crosswind, in general the performance has been good. I have not seen continual S turns going on for miles once the aircraft is established on a track, nor fluctuations in the CDI needle once on track.

 

Some of these problems may be caused by conflicting controller inputs - like a double assignment of aileron or rudder axes. I use an XBox controller in addition to my yoke, throttle and rudder pedals. The XB controller is only used for controlling the drone camera, When I first installed it, I found that MSFS automatically assigned flight control and throttle bindings to it in addition to the view controls. I didn’t realize that had happened at first, and it caused some very strange autopilot behavior in the default aircraft until I went into the controller settings and removed all bindings from the XB controller except the drone camera controls.