VNAV Path Construction When Extending from CF

[emerydc8]

Hi Hardy,

I was trying to set up a situ today to demonstrate one of the three features of on-approach logic.

Aside from allowing you to speed intervene on a VNAV descent and stay in VNAV PTH, and allowing you to move the MCP altitude through the aircraft's altitude without it leveling off in VNAV ALT (both are necessary on VNAV approaches), the third feature is that the FMC should shift automatically from VNAV ALT to VNAV PTH if you are below the path, but above the altitude for the next fix (intercepting the path from below). If you are not in on-approach logic, this shift will not occur and it will stay in VNAV ALT. 

To demonstrate this, I set up the following situ:
https://drive.google.com/file/d/0B-WRh0Hf7VdZUDg4NUFLUDVFLVE/view?usp=sharing

In the situ, I am in SFO at 6000' on a modified left base leg for the RNAV RW 19L. I extended from the CF (BERKS), which has a crossing altitude of 5000'. So, on-approach logic is not active at this point. When the situ opens, I am 15 miles from BERKS and it shows me 630' high. I would think that this far out, and with only 1000' to lose, I should actually be below the path.

My intent was to do the setup for a VNAV approach (LVSA) and set 5000' in the MCP until reaching BERKS. I was trying to demonstrate that as you intercept the path from below (somewhere around three miles prior to BERKS), there would be no automatic transfer from VNAV ALT to VNAV PTH, since you are not in on-approach logic until reaching BERKS.

The FMA correctly indicates VNAV ALT after I do the LVSA, but the problem I'm having is that during the entire situ, I was never below the path, so I couldn't do the demo. It just started around 630' high and eventually got to around 1000' high when crossing BERKS. I agree that the 1000' is correct over BERKS, but being 15 miles away from BERKS at 6000' shouldn't show that I am 630' high.

I am curious what it is using to determine my position with relation to the path when I extend from BERKS. Shouldn't it draw a 3 degree path out from BERKS, so that when I am 15 miles out and only 1000' above BERKS the VPP should indicate I am below the path? I could lose 1000' in 3 miles.

Thanks,
Jon D.

[Hardy Heinlin]

Hi Jon,

I'll try to analyse your situ later this year. I'm on summer holidays; these days I can answer just questions that are trivial :-) Please bear with me ...


Regards,

|-|ardy

[emerydc8]

Thanks, Hardy. Enjoy.

PS: Glad you don't think it's trivial.

[Hardy Heinlin]

I just discovered a special part in my code that makes it impossible to get an E/D when the aircraft is in idle descent mode (which is always the case when starting the DES phase) and the runway is the first waypoint during this idle descent.

Now I have two options.

a) Cancel the idle descent mode whenever the runway is the active waypoint.

b) Cancel the idle descent mode whenever the runway is the active waypoint and the distance to the runway is less than XX nm. -- If so, how great is XX?


|-|ardy

[emerydc8]

Hi Hardy,

Hopefully Peter and others can chime in here too. I think the only situation that you would be extending from the runway fix is on a 2-engine approach and the idea is never to be more than 15 miles from the runway in this situation anyway. So, if you made it 25 (second option), I think that would work. Plus, if you are more than 25 miles from the runway fix, on-approach logic will not be active and you won't be able to speed intervene and descend in VNAV PTH anyway. On a 2-engine approach, I don't think you will be doing many idle descents and even if you have to do it, the A/T won't work anyway.

Jon D.

[Hardy Heinlin]

I'm getting closer to a solution.

I already have the approach logic becoming active when within 25 nm from the runway. What I'll have to do is cancel the idle descent mode when the approach logic is active.

Normally, the idle descent phase is over after passing the first descent constraint. However if the runway is the only descent constraint, the current version remains in idle descent mode until passing the runway, and there will be no accurate approach path guidance, just raw idle descent guidance.


|-|ardy

[emerydc8]

Thanks. I don't know if anyone here has actually extended from the runway fix on a normal flight. I can tell you that it's virtually never done in the real airplane because you would always select some sort of approach.  And in the sim, the only time I have extended from the runway fix is for a 2-engine approach, because this is where they disable the ILS on you. So, the A/T FMA would be blank.

[Hardy Heinlin]

Why don't you fly the normal ILS approach profile (with VNAV) instead of deleting all fixes before the runway? Is it because you don't want to level off with 2 E/O?

Anyway ... in my current tests it's working now.

[emerydc8]

Why don't you fly the normal ILS approach profile (with VNAV) instead of deleting all fixes before the runway? Is it because you don't want to level off with 2 E/O?

Good question, Hardy. Flying the ILS approach profile is essentially what you are doing but there are some things that could cause trouble if you don't extend from the runway. The idea is to capture the glidepath at whatever altitude you level at as you maneuver for the approach, so there are no level segments at all once you start down.

First, in the sim, they fail the second engine (engine fire) at 1,000' during the 3-engine missed approach, so you may decide to level off at only 1500'. The CF or FF altitude may be 2000' or 1800' and you may never climb that high. That causes problems if your first fix that you extend from is 500' above you and you are trying to set up for a VNAV approach. If you just did a missed approach, you are in CLB mode. If you never get to your CF or FF altitude (e.g., ZACHS at 2000' or MEALS at 1800' for JFK 31L), installing the approach and extending from either of these fixes while you are at 1500' is not going to take you out of CLB. Your option is to re-cruise or climb up to that altitude then extend the fix from there. This is why it is important to have the FMC go out of CLB mode anytime you install and extend an approach fix in 1L -- be it CF, FF or RW -- and your actual altitude is at or above that fix altitude.

Second, if you extend from the CF (or even the FF) while you are on a downwind and then turn your base inside that fix, you will not be able to get LNAV or VNAV to work correctly. In order to sequence that fix and let it cycle to the next fix in front of you (CF to FF or FF to RW fix), you have to go out (extend your downwind) past the wayline for that fix, which is a line drawn across that fix and perpendicular to the course. If you don't cross that wayline, the fix will not sequence and it will stay in 1L. If you do your LVSA on a base leg and you have cut inside the fix you extended from (you didn't cross the wayline), LNAV is going to give you guidance to that fix when it goes active -- not the fix you want to fly to next. Also, descending in VNAV PTH with the speed window open requires on-approach logic. But on-approach logic doesn't go active until sequencing the fix you extend from (CF or FF). As you said, if you extend from the runway, on-approach logic is active immediately as long as you are within 25 miles and you don't have to worry about a fix not sequencing because the RW fix is always in front of you. So you will have correct LNAV guidance and on-approach logic by extending from the RW fix.

Third -- and maybe this is a sim glitch -- if you extend from the FF, there is a possibility that when the FF sequences and on-approach logic should become active, the processor may no sequence fast enough when it crosses the FF. Thus, the FMC will not go from VNAV ALT to VNAV PTH -- it will stay in VNAV ALT. In this situation, pushing the MCP altitude button would be required to start the descent to the DA (nominal 500' for a 2-engine). By the way, on a three or four-engine approach this is why we wait until after passing the FF before setting the missed approach altitude. If you had already set your missed approach altitude before the FF, you wouldn't be able to get it to descend by pushing the altitude knob. Anyway, if the processor doesn't sequence fast enough, you would have to altitude intervene (with your DA set in the MCP) and the FMA pitch mode will go to VNAV SPD (presuming you are already more than 150' above the path) and it will dive until it's within 150' of the path, then revert to VNAV PTH, after which it is important to rotate the altitude selector to the missed approach altitude or it will capture the DA and go to VNAV ALT. When you extend from the RW fix, you don't have to worry about whether on-approach logic will activate at the FF or CF since you are already in on-approach logic as soon as you extend and execute.

So, to answer your question, it's because of (1) the possibility of being below the fix altitude that you extended from and not getting out of CLB mode (it will not descend in VNAV while in CLB mode); (2) not having on-approach logic and correct LNAV guidance because the waypoint didn't sequence; and (3) the possibility that the sim may not shift over to on-approach logic at the FF and you will go high if you are still hooked to the A/P (it will stay in VNAV ALT), it is easier to just extend from the runway fix. And that's how Atlas, Polar and Kalitta teach the 2-engine approach.

[Hardy Heinlin]

This stuff is now implemented in PSX 10.0.9-beta5:

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