Capetonian; yes but it's not just AF is it? Colgan, whoever the finless A310 belonged to, and the carrier that recently crashed the 777 at SFO were not AF. This is a general problem - not enough time is spent training and practicing this. On the ATPL course, we did recovery from unusual attitudes* In my case; in a PA28. I have never done that in an Airbus SIM. (I've never flown Boeing).

pontifex; Yes, I couldn't agree more - A vs B is very tiresome. They both design and produce fantastic machines of the highest quality. (Unfortunately, they are nowadays operated by pilots with very variable ability and training).


*One is told to close one's eyes and take hands and feet off the controls. The instructor then puts the aircraft into an unusual attitude, for example, 30 degree bank, nose high - approaching a stall, and then one is told to "open your eyes: you have control". You have to instantly assimilate the situation and attitude and quickly return the aircraft to straight and level flight with the appropriate power, configuration and attitude using just your instruments. A very good exercise.

I still feel that there's a fundamental flaw in expecting humans to monitor computers.. Computers make an excellent job of monitoring.. they'll happily monitor away for a million years..
Now tell a human to take a look at his watch.. and tell him to carefully monitor it for the next 3 1/2 hours to see that it doesn't falter...:bored:
Pilots need to be designed back into the loop.. with the computer monitoring us.

On the basis of this statement:

That would still be you. The flight control computers are part of the FBW system, which as many have explained over and over on here is a completely separate system from the autopilot.

In the most simple terms, all the FBW system does is replace the old cable and/or electromechanical/hydraulic connections with electronic ones. FBW simply translates control input from the pilot (or the automation) to the flight surfaces and engines, it is not an automated system in and of itself. Furthermore, the envelope protection aspects do not "overrule" the pilot, they just keep the aircraft within the safe limits of the flight envelope of the airframe.

Sometimes a small excursion outside the safe limits may be required Mother Nature doesn't know the limits. Boeing recognise this airbus doesn't

Yes, but we are not monitoring computers per se, we are monitoring the flight path. In other words, the computers should be following what we have programmed them to do, and one of our many functions in the cockpit is to monitor the flight path and the instruments to confirm that the computers are in fact doing so.


A bit off-thread, but: That half-wit who wrote to the Times about pilots falling asleep because of the autopilot missed the point, as so many non-airliner pilots do. We are not sitting bored and going to sleep because of the computers* - there is more than enough for both pilots to do throughout a flight**. The computers may be controlling the flight surfaces during the cruise, but we, the pilots are still doing the flying, navigating and communicating required.

(An analogy is: a cruise control in a car looks after the speed, but you the driver are very much still involved in decisions and considerations, and are checking your mirrors every 10 seconds and keeping tabs on other traffic around you etc.).



*fatigue caused by poor rostering - long hours and minimum sleeping time between duties - is the cause of sleep incidents, despite what Europe may think, and is something that urgently needs to be addressed if we are to avoid any more Colgans etc.

**fuel and fuel system checks (are there any leaks?); navigation checks; keeping an eye on all the aircraft systems; radio work; position accuracy checks; avoiding thunderstorms by 20-30 miles and flying round them, plotting one's position; checking the weather conditions and diversion airfields along the route, (in case of engine failure, pressurisation failure, or medical emergency for example); checking position and that of others when beyond radar coverage; ditto when overflying states with minimal air traffic control facilities, and/or navigation beacons; avoiding turbulence, special checks and considerations during ETOPS phases; checking the pressurisation and air conditioning etc. etc.

Oh yes, there's plenty to do, and using an autopilot increases safety because that way we are doing all those other tasks as well and not using our total concentration to keep an airliner within +/- 200' of it's cleared altitude, at 37,000' doing 500mph, while being just 1000' feet above or below other airliners coming in the opposite direction!

What are the other tasks ?
Piloting the aircraft ? :rolleyes:

jcjeant,

Try VERY carefully reading Uplinker's post prior to yours!

Cheers ;)

Uplinker please know (or indeed remember) a fact about the A330 when discussing the performance of the First Officer that night, the A330 stall warner is suppressed below 60 knots because the computers believe the aircraft to be on the ground below that speed. It is incredibly unfortunate because when the First Officer was pulling back on the stick the audible warnings were going away. Therefore he (wrongly) believed he was doing the right thing at the time. From what I read about the crash the picture in the flight deck was incredibly confused and I don't think they really trusted anything instrument wise. Would anyone conceive that an airborne A330 could be travelling at an airspeed LESS than 60 knots? Incredible but all so sadly true in this case.

Yes the crash was poorly handled as we know. But don't treat the guy like an idiot because the only one who sat in his seat at the time was him alone. It is highly possibly he believed the aircraft to be overspeeding given the lack of airspeed information and an audible warning that went away when he applied back pressure. Which of us could honestly say that we definitely wouldn't have been led into the same trap given the same set of unfortunate circumstances in the middle of the night on a long haul flight crossing time zones? It would be incredibly arrogant to rule it out.

They just needed the real pilot up there at that time who would have handled it but he was taking his rest break. No pilot who had a clue what he was doing would pull the nose up over 20 degrees at that altitude and not know he was going to stall. These guys were not qualified to hand fly obviously and killed everybody. Putting out the stall warning alert at 60 knots should not make a competent pilot feel all is well with that deck angle at that altitude.

Pilots need to be designed back into the loop.. with the computer monitoring us.

Bring on the dog; ah, but it's there to keep us out of the loop. Tacho's in the cockpit? They've already thought of cameras; add a motion sensor and an alarm; job done.

It is highly possibly he believed the aircraft to be overspeeding given the lack of airspeed information and an audible warning that went away when he applied back pressure.

This has been hashed about already. It's a very modern a/c with ground speed read outs. The lack of air-speed is a red herring. What was the ground speed readout? Back to basics. If you are not sure what is going on you put the a/c into a known state and wait & watch; generally level flight. Pitch & power and assess. If you don't know the basics it's impossible. There in lies the root cause. BASICS!
It's like the various crashes with blocked tubes. The instruments might work in the wrong sense. Back to Basics. Level flight at a known power setting. Gently move the stick and watch what the instruments do. If they are obviously daft then assess, but don't start pushing & pulling is gay abandon hoping to finds the exit to the maze of confusion.

I think it is a fact that the crew did not do what they where supposed to do to recover from the stall.
You can only conclude that they were unsufficiently trained to recover from a stall.

Coroner's Inquest - the facts
 

Just so everyone is fully informed of what happened:

The Coroner reviewed the full accident report, the relevant medical and pathological evidence, then called the AAIB as expert witnesses to explain the sequence of events and the findings of the BEA report.

Evidence was also admitted relating to Thales pitot tube problems on A321 aircraft in 2012 (see AAIB Bulletin 9/13), equipped with the same later mod tubes that were specified in the AD for the A330/A340.

After all of that, and some questions from the families, the Coroner handed down a considered narrative verdict. He did a thorough and even handed job, and the families were satisfied with the result.

(Coroners in the UK are judicial officers charged with conducting Judicial Inquests into unnatural deaths, not to be confused with US Coroners.)

Bubs - they were all real pilots. For one thing the much-maligned PF was a highly-qualified sailplane pilot and would probably have had more handflying experience than his peers (albeit at relatively low levels). The PNF was clearly concerned at the start of the sequence, but seemed to be overcome with indecision as time progressed. When the Captain arrived, even he couldn't make head or tail of what was going on until it was too late.

As I've said many times, the old "no pilot would X" claim is something of a convenient excuse, because if the circumstances conspire in certain ways, then even the best and most experienced pilots can make horrific mistakes (e.g. taking off without clearance, retracting high-lift devices too soon, attempting a landing in a thunderstorm that exceeds minima etc.). As we covered in the original threads in Tech Log, the circumstance here seems to have been a startle response that got out of hand - combined with a lack of high-altitude manual handling training and a lack of recurrent stall recovery practice.

"They just needed the real pilot up there at that time who would have handled it but he was taking his rest break. No pilot who had a clue what he was doing would pull the nose up over 20 degrees at that altitude and not know he was going to stall. These guys were not qualified to hand fly obviously and killed everybody. Putting out the stall warning alert at 60 knots should not make a competent pilot feel all is well with that deck angle at that altitude."

Not having flown anything this big yet - would just one sensor can make the aircraft go into ground mode? Wouldn't it be airspeed+squat switch+(stuff)?

@Yancey Slide:

I know it's a bit of a big ask, but I recommend going through the Tech Log discussions on AF447 before risking starting the old hamster wheel up again. The <60kts IAS limitation on the validity of AoA data has nothing to do with ground mode, but if you want more detail, then the threads are there for your perusal.

The long and the short of it is that the aircraft did not go into any unusual modes, and the only technical problem was the icing of the pitot tubes. Even that had cleared prior to the aircraft stalling. It should be pointed out that the stall was allowed to develop to a point where the aircraft was so far outside of the flight envelope that any instrumentation would likely have given readings that didn't make much sense.

I'll go dredge the tech log, thanks. The question wasn't so much re 447 as it was a generic inquiry about redundancy in ground/air mode logic. I wouldn't have thought one sensor was sufficient.

Yancey - to save you having 'hours of fun', accept that the designers expected (not unreasonably!) the a/c to be 'on the ground' at 60kts or less and therefore designed the software to close the warning. They now realise, thanks to AF, the error of their ways.

@Yancey - You're right, it isn't and there are several conditions that need to be satisfied for the systems to switch to ground mode. It's important to understand that the <60kts rule was added to the spec regardless of air/ground mode though - it's purely because the AoA vanes' certification states that they cannot provide useful data below that IAS.

There was a lot of discussion over the relationship between that limit and the stall warning in the Tech Log threads, and the only thing that could be considered a consensus opinion was that it is a very tricky problem to solve. This is because a behaviour that may be valid in some scenarios would not be valid in others, and could be downright dangerous in many of them.

The take I came away with is that no matter whether your instrumentation is of the old "steam gauge" variety or of a more modern vintage, it can only provide useful information when inside the flight envelope, or to an extent when slightly outside. The further outside the flight envelope you go, the less useful some of the instruments and warning systems become.

EDIT : @BOAC - see my first paragraph. The <60kts limitation is not related to ground mode, and the designers made no such assumption. What they couldn't take into account was a scenario in which the aircraft was as far outside the flight envelope as AF447 became.

- and I didn't say it was?

@BOAC:

I may have misinterpreted then:

As you seem to be equating the AoA <60kts inhibition with being on the ground as a design feature, when it is not. It is an implicit side-effect of the design, but it was not intended to apply only on the ground.