Fix Wayline Cycling

[emerydc8]

Hi Hardy,

Can you take a look at this situ. https://drive.google.com/file/d/0B-WRh0Hf7VdZN0Q5RVpsWUlKSFk/view?usp=sharing

Maybe Peter can chime in here too, but I am pretty sure that if you extend the CF and then turn inside it, such that you never cross the CF wayline, it will not cycle to the FF and then to the RW. In this case, if you did a go-around from an ILS approach, when you pressed LNAV at 400' it would give you guidance back to the CF. In the situ above, I installed and extended the CF as I was southwest of JFK, so I never got past the wayline for ZACHS. Watch it cycle from ZACHS to MEALS while on a base leg.

The 21-mile rule applies when you are abeam the fix (crossing the wayline). In order for the fix to cycle you have to be abeam and within 21 miles of the fix (just like enroute cycling). In the terminal area, you are almost always within 21 miles of the fix if and when you cross the wayline, so the big issue is crossing the wayline itself. In this situ, I never crossed it, so it should not cycle. This is why you have to extend the FF in this case -- so the fix you are going to is in 1L. Thanks.

Jon D.



_____________________________________________________________

[Hardy Heinlin]

Hi Jon,

so it's about sequencing the active waypoint of a direct-to extension.

Your quoted text does not define the exact criteria for this case. I don't think that the 21 nm rule is still valid here. PSX uses 3 nm in this case -- and --- as always, the aircraft track must be within +/-90° of the leg course.

This means, when PSX approaches the leg from offroute (e.g. with HDG SEL for LNAV capture), and the intercept track is greater than 90°, the waypoint cannot be sequenced. This implies the fact that you need to pass the "wayline" in the direction of the route, and not in the opposite direction as that would be an intercept angle greater than 90°.

So, in PSX you can never sequence it by passing the "wayline" even in the correct route direction and even if you are on a 19 nm offset as long as your crosstrack error is greater than 3 nm.

I think this 3 nm limit is important as some procedures include complicated teardrop curves and zig-zag routes, and may otherwise sequence the leg when you are crossing the "wayline" 19 nm abeam the active waypoint. Only in cruise outside a terminal procedure the 21 nm tolerance is no problem.

Consequently, those criteria in PSX will also sequence ZACHS in your situation when you are past ZACHS and your intercept angle is less than 90° and your crosstrack error is less than 3 nm.

The question is whether this is a consequence on the real ship too. Regarding "wayline" crossing in the route direction: This criterion is applied on the real ship and in PSX. But how is this criterion actually programmed? There are multiple methods. Each method has its specific side effects.


Regards,

|-|ardy

[emerydc8]

Hi Hardy,

Your quoted text does not define the exact criteria for this case.

I edited and posted more information on this while you were writing your reply. It is definitely a consequence in the sim because one of the favorite tricks check airmen like to do is vector you inside the CF and then give you a missed approach if you don't realize the CF never cycled.

[Hardy Heinlin]

I think it's reasonable that the aircraft has to be before the wayline in order to arm the sequencing. I can implement this additional condition.

But I doubt that the condition re crosstrack limit is still 21 nm in this case. Imagine you are 20 nm abeam the active approach leg, making a large turn and you cross the wayline out there for a second. It would sequence the active waypoint. In my opinion, that would be less reasonable than sequencing it when you are within 3 nm on an intercept angle past the active waypoint; in that case it would be obvious to the FMC that the pilot wants to sequence it.

The 21 nm limit must be reduced for this case. If not exactly to 3 nm, then to something very close to it.


|-|ardy

[emerydc8]

I think it's reasonable that the aircraft has to be before the wayline in order to arm the sequencing. I can implement this additional condition.

Thanks, Hardy. That might work because the 21-mile rule virtually never comes into play in the terminal area unless, like in your example, you are on a base leg and toeing a fine line very close to the wayline where it might or might not cycle.

[Hardy Heinlin]

You think in a terminal procedure too it should sequence when the wayline is crossed while the crosstrack error is 19 nm?

I have allowed this tolerance when I experimented with it some years ago; it's too sensitive, it may create chaos when you get vectors. It may even sequence multiple legs one after the other if the routing is formed like an arc.

Does your text really mention all conditions? Or just some?

[Britjet]

I don't know about the cross-track tolerance that triggers it but the idea that you could turn between the fixes and it wouldn't cycle is true. We also would do this in the sim from time to time when flying an LNAV approach to encourage pilots to check the leg cycling was correct. If the leg didn't cycle to the FF then an LNAV go-around wouldn't work, which was the main issue.
For the benefit of other readers, you should always monitor the LEGS page to ensure cycling is correct if you turn into an approach centreline at or inside the CF..
Peter.

[Hardy Heinlin]

Re crosstrack tolerance: Note that the wayline is very close to the active leg corridor if the next leg course change is close to 90°. The difference between leg course and wayline is then close to 45°. If you are a little bit offroute there is the risk that you cross the wayline many miles before the active waypoint, especially on a go-around when re-entering an approach in the FMC with a relative 90° direct-to base leg in front of you; a little turbulence combined with a slight heading error, and 30 miles of routing ahead of you will be sequenced just because you touched that 45° diagonal wayline in the critical direction for a second. (With the sine of 45°, the wayline is much longer than 21 nm.) To compensate this problem, I added that 3 nm condition for direct-to legs.

[emerydc8]

I have allowed this tolerance when I experimented with it some years ago; it's too sensitive, it may create chaos when you get vectors. It may even sequence multiple legs one after the other if the routing is formed like an arc.

Is there any way to limit this non-cycling issue to just the CF (or maybe what shows up at 6R when you select the approach)? This is such a limited scenario. I don't think you'd ever have to worry about cycling when turning inside the FF because you are usually established by that point anyway.

Thanks for the input, Peter.

Jon

[Hardy Heinlin]

It is limited to direct-to legs.

I can limit it to all aproach legs as well, regardless of direct-to or not. But not to the CF alone. Either all approach legs or none.

[Hardy Heinlin]

Problem solved in PSX 10.0.9a:

http://aerowinx.com/board/index.php?topic=3685.0


|-|ardy

[emerydc8]

Thanks, Hardy. In 10.0.9a, when you cut inside the wayline for the CF, it (correctly) won't cycle the CF, but when you do a missed approach, it starts to cycle the fixes (FF, RW, then eventually the missed app. proc.). So, LNAV will not give you guidance back to the CF because after pressing TOGA, it starts to cycle through the waypoints it flew over.

[Hardy Heinlin]

This is because this new feature works only for approach legs, not for missed approach legs.

[emerydc8]

I guess that's close enough. At least you can demo that the CF will not cycle if you don't cross the wayline with the idea that if you actually did go around from an ILS in the real plane that it would turn you back to the CF when you selected LNAV at 400' because that's what would be in 1L.

In the case of an LNAV approach, you wouldn't even get that far since LNAV would just stay armed (it would not capture) when you tried to intercept the final course. You would also get a NOT ON INTERCEPT HEADING.

Here's a video I did. The first shows a turn inside the CF wayline. The second shows normal sequencing.

https://youtu.be/FdNE0lO20Yg

[Britjet]

At the risk of throwing a spanner in the works here Jon, I'm pretty sure the LNAV will capture even if you miss a waypoint?
Peter

[Hardy Heinlin]

In PSX, the crosstrack error always refers to the active leg and its great circle extension around the globe, and LNAV always tries to reduce this crosstrack error and maintain the course, no matter where on the planet. The reference is always the active leg, and not the previous or next leg, even if the aircraft is closer to them than to the active leg.

When the aircraft is past the active waypoint and the waypoint is not sequenced, the crosstrack error remains a function of the abeam distance to the extended line of the active leg line. The crosstrack error will not become a function of the distance to the active waypoint.


|-|ardy

[emerydc8]

At the risk of throwing a spanner in the works here Jon, I'm pretty sure the LNAV will capture even if you miss a waypoint?
Peter

This is interesting. I have not tried this in the sim, but was just going by the material I posted above with the assumption that the check airman tried this in the sim and it wouldn't capture.

He writes: "If the fix is not captured: . . . If not in LNAV, the CDU scratch pad indicates NOT ON INTERCEPT HEADING. LNAV remains armed and will not capture." (Emphasis in original.)

Are you sure you were doing an LNAV approach and not an ILS when you tried to go around with the CF in 1L?

Jon

[Hardy Heinlin]

I think it makes sense that LNAV won't capture and that NOT ON INTERCEPT HEADING will be displayed.

I'll implement this effect and keep it until someone proves it's wrong or version specific :-)


|-|ardy

[Britjet]

Jon - yes, definitely an LNAV approach - that was the whole point of the training item...I see the logic that is being put forward here - yet.....:-)
I have seen this in the sim, I am sure - and in fact I remember it happening to me on one of my own checks (you tend not to forget!). It was quite tricky to vector tight enough to get inside the IF without it cycling - but if you managed to do it the crew would perform the LNAV (with capture) but the active waypoint would be behind them..
I am fairly sure that LNAV will engage here in the approach instance, but we will see. Socks at the ready..:-)
Definitely one to test if I get chance - tricky to do, though..
Peter

[emerydc8]

That would be great if you could test this one, Peter. If you  turn inside the CF wayline and you arm LNAV as you are on a heading perpendicular to the course, which way will LNAV turn if/when it intercepts -- towards the FF or towards the CF? Hopefully we can find out. Thanks again.

Jon