Altitude capture mode.
 

Why was the the auto pilot control law for altitude capture designed to be apparent to the pilots (FMA “ALT*”) and not transparent - the aircraft transitioning from the vertical path to capturing the selected altitude with the FMA changing directly to “ALT” ?

Don't you want to know when it's about to level off?

It probably goes back to the days before the altimeter had a target bug on it. Also, when the VS is high, ALT* may engage well before the target altitude.

Absent a failure indication (or inappropriate FMGC programming) the aircraft will level off at FCU selected altitude. The current indications: altitude window flashing yellow when approaching within 750 feet of the FCU‑selected altitude and FMA change already indicate the aircraft will level off at the FCU set altitude.

ATC assigns a vector 90 degrees off the STAR. I input that heading and pull. The aircraft lateral mode immediately goes to HDG even though the aircraft is yet to be on that heading. In this case we make the assumption the aircraft will turn to the selected heading and roll out.

Why is there not a “HDG*” indication until the aircraft actually rolls out and “captures” that heading?

A turn to a heading is flown at more or less only one bank angle, and you'll always expect to see it start rolling out at something like to 5 to 15 degrees before it's reached. And it's very obvious when it begins reducing the bank.

Climb rates prior to an altitude capture can be anywhere from a few hundred to a few thousand FPM, and which would lead to capture initialization anywhere from 50 feet prior to the selected altitude, to more than a thousand. Moreover, pitch/VS changes are much more subtle for a given change, and much more easily hidden inside turbulence, AP chasing speed, and things like that. In short, harder to see. Therefore it's more valuable for the airplane to explicitly tell you that it's begun capturing. So if I'm climbing at 3000+ FPM toward a leveloff, it's 800 feet to go, it's subtly pitching down, I don't have to ask myself "is it doing it?" (since it can also be subtly pitching down for the other reasons I stated). "For this climb rate should it have started already? Or should I give it another couple of seconds?" Etc.

Also, heading mode is extremely simple logic and basically the same across every plane with a heading bug and an autopilot. But as far as leveling off, not all planes will necessarily level off at the selected altitude. Some, you need to push a separate button to arm that behavior. So that raises the potential for confusion through misapplication of old habits, so I appreciate the airplane telling me explicitly what it's doing instead of forcing me to infer it from assumptions.

Good question CaptainMongo. I noticed especially on the light A319 climbing at more than 3 to 4000 feet/min the ALT* will engage very early and the speed will often drop quite significantly as it looks like the ALT* takes priority on the speed and assumes you keep the same energy until you are close to the altitude target. I can't find it anymore in the FCOM where it's written but there was a note about this in case you lose one engine during ALT*, the speed might drop significantly as the priority is to the ALT* mode and not the speed mode.

Any more information about it would be cool. =)

The altitude capture phase (Airbus and Boeing) is a result of computation of ROC (ROD) and vertical distance between current altitude and FCU/MCP target altitude, hence the wide range of ALT* observed. BUT the really important thing is that once this phase is engaged the autopilot and/or FDs will now follow a geometric flight path or trajectory as calculated at the moment of capture. i.e. it will follow that path regardless of all else and that include a loss of energy for any reason (e.g.windshear or engine failure) The aeroplane will sacrifice speed if necessary to maintain that path. Some aeroplanes do have protection e.g Airbus but not until speed has bled off to about Vls (B777??). Earlier a/c had no such protection. If the speed loss is significant and automatics are not recovering then pilot intervention is a must. And so it is very important to know when capturing (ALT*) or when maintaining (ALT) a target altitude.

For an extreme example look at A330 crash at Toulouse during test flight 1994 A330 Crash Toulouse

Many factors involved but its this feature which sealed their fate even though they recognised it and monitored it, they delayed too long (FCU set to 2,000 ft with simulated engine failure and very very early ALT*)

Thank you for that input Starbear. Do you have any reference? I don’t doubt your saying, it makes sense. Would be interested to know more about it. Thank you.

Now that is more of a challenge these days but I will try later (just going out). My knowledge comes from many years of teaching this stuff combined with research. I use to have access to a wonderful set of manuals which were the maintenance engineers course notes (the Maintenance manuals also useful but sometimes too unwieldy for this stuff).

In the meantime if you happen to be due in the simulator on almost any "modern" aeroplane, ask if you can set up an early Altitude capture by inducing a high ROC withA/P engaged and then when Alt * or equivalent is annunciated, reduce a thrust lever to idle and watch the aeroplane's behaviour. If in an Airbus wait and see what happens close to Vls

Alright! Unfortunately I just did my sim early this month, but if I have a chance to go in the FTD, I will try that! Thanks again! = )

As stated earlier the key to "capture" mode is that it is a speed controlled by thrust mode. If for some reason the thrust become limited the APFDS will sacrifice speed until the low speed protection kicks in.

Engine failure during the capture mode can be very interesting and specific handling notes may be required.

For example:
"A Flight Director (FD) guidance issue has been identified when the aircraft enters into ALTS CAP at a low selected altitude, during a high rate of climb, and at the same time an engine failure occurs. With one engine failed, the ALTS CAP guidance continues to be based on the rate of climb corresponding to both engines operating, resulting in guidance that could leads to a loss of speed below V2. There is no low speed protection based on pitch in ALTS CAP.
Velocity Decay during ALTS CAP occurs in OEI (One Engine Inoperative) conditions, either during Take-Off (TO) or Go-Around (GA).
ALTS CAP mode of operation:
 Altitude on pitch
 Speed on thrust (in OEI, thrust command is saturated)
During ALTS CAP, FD is flying Pitch/G command to level off at target altitude (i.e. set through Pre-Selected Altitude) and speed control is on thrust command.
This issue can occur with:
 Autothrottle engaged---the Autothrottle (A/T) under speed (USPD) sub-mode gets displayed on the A/T Flight Mode Annunciator (FMA) in spite of not being effective due to lack of thrust command authority in this specific case
 Autothrottle not engaged ----when the A/T USPD is not displayed on the FMA"

As a general rule of thumb, the A320 will enter Alt* when the altitude left to climb/descent equals about 1/3 of the ROC/D. Eg climbing at 1500fpm to FL100 you’ll get Alt* approximately FL95. Another good tip for those new to the aircraft is that if you want to get out of Alt* using the autopilot due to too high a climb/descent rate then you need to use ‘push to level off’. Simply selecting a lower V/S will just see Alt* re-engage after a second or so as the conditions remain the same (high instantaneous V/S and a reducing height to level off altitude). By selecting V/S 0 the aircraft initially ignores the commanded Alt and you can then wind the V/S to a more appropriate rate as the actual V/S reduces. Very useful as a ‘get out of jail’ card to avoid a nuisance TCAS.

Thanks for all the inputs, especially starbear, now I have a better understanding of the alt capture mode.

One is never to old to learn something new...

As indicated earlier, providing sources for my explanation is easier said than done but the attached documents may help with understanding and support my assertions somewhat. If you don't want to trawl all of the attachments, the Flightglobal articles summarises very well both the accident and how the a/c behaves in ALT*. I realise the majority of these papers are heavily weighted towards the accident rather than the actual request for source information but its the best I can do these days, I do believe taken together,they should provide reasonable confirmation.

1) Combined articles from Flightglobal soon after the crash.
Attachment 4559

2) Article from Pprune is an English translation of the BEA preliminary report (I though a final report had been published but cannot find at all now).
Attachment 4555

3) A330 FCOM extracts
Attachment 4558

4) Deeply technical extract from B757 MM but explains how their ALT CAP works (ALT* Airbus very similar) it explains how it adjust vertical rates in that phase to basically follow what I perhaps erroneously refer to as a geometric path (trajectory is perhaps the better word)
Attachment 4557

Pretty good references Starbear.

On both Airbus (post 300/A310) and Boeing (post B767) altitude capture is actually done by using V/S mode. It is hidden by either ALT * (Airbus) or the ALT alert box (Boeing).

The mode change is triggered by reaching the calculated ROC for the trajectory intercept of the altitude. It is is approx equal to 10% of the ROC or ROD.

It then follows a parabolic path to the level off pont. The required parameters for the first segment thereafter are first cruise speed, pitch attitude for that speed, required thrust (all derived from the FMC and associated performance databases). The AFCS trims as required.

Although SPD my be displayed in the A/THR channel, the auto thrust is actually monitoring Vmin/Vls and Vmo/Vfe -5. (see your FCOM for which is which for your manufacturer.

The actual V/S mode was invented by Douglas for the DC10 AFCS. It made its way to Airbus and Boeing.

It can be an automation trap if you do not know what you are doing.

:8:8

How does this V/S mode differ from traditional PID control system design?

On some older autopilot systems, changing the vertical mode during capture will disarm the alt and if not caught you end up barreling through the cleared level. Knowing that the capture mode is active is critical in these types.

Flybe -Q400 altitude capture required careful watching as well.

Not necessarily. Some time ago I had an A320 bust an altitude during a descent. AP on, everything in managed, nothing abnormal going on. I (PM for the leg) called thousand to level off only to later see the aircraft passing thru 17000', with all the proper mode switching mind you! We stopped it at 250' below the cleared altitude, even the altitude alerter went off. I've never fully trusted the A320 automation, but I do it even less ever since.

I was about to say this too. In heading mode there's no mode reversion if you switch the target value near "capture". However, switching the altitude target in Alt*/Cap will revert the pitch mode, hence it's quite useful to know when capture is occurring so you don't revert. The A320 is now a 30yr old design so I guess it qualifies as an old aircraft now, but this same logic was true for the DHC-8 (another old aircraft by now)

As a side note, I had noticed the "target over speed" preference in Alt* as I've seen the aircraft get below the speed in certain circumstances during capture whereas it didn't during the rest of the climb. Nice to see my suspicion was correct

Just a note that not all Capture modes use VS commands.
Some systems use a constant g command.

As starbear says. ALT* tells me to watch it like a hawk as the only protection I have is VLS. As an aside I am unlikely to let it capture an altitude with a high ROC as I have witnessed to many instances of energy loss not to many that most airspace is to busy these days and I want to avoid a TCAS. A small V/S ROC does a nice job at high levels.