If I remember correctly those Air Caraibes guys who managed their loss of speed indication, also had the good idea to throw that switch from AUTO to ON and that probably helped to regain correct speed indications

@barit1
The device you describe would be an ice detector which on some aircraft would bring on a warning only.

IMO it's just plain stupid that a multi-million machine with humans inside have to rely on a "stupid" analog tiny tube, even worst without redundancy or a reliable alternative.

If you're about to invent something better than a pitot-tube you'll be a rich man. Even the space shuttle measures its speed with pitots while in the atmosphere...

Aircraft have been flying around for decades with pitot tubes which haven't iced up - how come there's been a problem with these ones?

Hi,

It's not new ....

Would there be a forum such as this if we truly believed in being cautious in interpreting findings until all relevant information has been examined?

pitch + power
 

Using the standard practice of pitch + power to maintain safe flight does not work if the aircraft is ALREADY stalled, which was the case here.

@ RobBamber

Didn't the PF acknowlege reaching level 100?

Quote from http://www.bea.aero/fr/enquetes/vol....ai2011.en.pdf:

At 2 h 13 min 32, the PF said "we’re going to arrive at level one hundred".

That's roughly 25000' from the level 2 min before (350):

"At around 2 h 11 min 40, the Captain re-entered the cockpit. During the following seconds, all of the recorded speeds became invalid and the stall warning stopped.(...) The altitude was then about 35,000 ft, the angle of attack exceeded 40 degrees and the vertical speed was about -10,000 ft/min. The airplane’s pitch attitude did not exceed 15 degrees and the engines’ N1’s were close to 100%. The airplane was subject to roll oscillations that sometimes reached 40 degrees."

The aircraft pitched up steeply and decelerated from cruise speed to a virtual standstill in the first 30 seconds or so, then went from about 80 mph up to over 100 mph down. None of that would have been interpreted as normal horizontal flight.

The Guardian has cut ‘n pasted directly from an earlier post in this thread.... Air France crash inquiry details pilots' battle for survival | World news | The Guardian Attributed to "Anonymous A330 Pilot”

Judging from the reports of what happened in the cockpit, one is left with the impression that without automation both pilots were seemingly bemused at events and helpless. This is not surprising, since with many airlines basic instrument flying ability has been relegated to low priority in favour of competency at autopilot operation. Even Jet Upset training is not a serious subject in simulator training. Five minutes at the most and once a year is the norm.

Unless I missed it, nobody seems to have mentioned that the original worst-case theory--that the airplane flew into the mother and father of all ITCZ thunderstorms because it was radar-blanked by a line of cells between it and AF447--no longer seems to hold water.

Certainly 447 flew into weather that created an unusual kind of supercooled, high-altitude icing that suddenly affected the Pitots, but the CVR transcript doesn't mention anything that might be interpreted as turbulence upset--only the apparent impossibility of dealing with multiple fault warnings and anomalous displays.

Too Many Laws
 

Seems like too much pilot training is spent on understanding five control laws: Normal, Alt 1, Alt 2, Direct, Abnormal. Are there more?

Makes me wonder if more than two are needed or desirable: Normal and Direct. I don't understand why the AP/AT had to drop out of Normal for flaky airspeed. At least not in cruise. It could hold pitch and power for awhile without dropping out, giving the pilots time to analyze the situation without having to hand fly, too.

In the sim if all the pitots are iced at the same rate the aircraft will stay in Normal Law as there is no disagree. In a climb the aircraft will eventually overspeed and an nose up input made leading to an uncontrollable climb leading to a stall and very low airspeed. Not a very pleasant scenario.

stall warning
 

Why would the stall warning be programmed to stop if the AOA indication is deemed unreliable below 60kts? Doesn't it seem intuitive that if a 200t jet transport has an IAS of < 60kts (and WOW says the plane is in the air), then it MUST be stalled, regardless of the AOA reading?

Then again, we have our iced pitot scenario, with IAS of less than 60kts.

Der Spiegel on AF447
 

New article on Der Spiegel in english, including a quote from captain Dubois after entering the cockpit, that was not in the BEA-report, but comes from sources close to the investigation team.
Doomed Flight AF 447: Questions Raised about Airbus Automated Control System - SPIEGEL ONLINE - News - International

Right Way Up, about as likely as 3 engines failing at the exact same time in the exact same way, not very likely.

"The aircraft pitched up steeply and decelerated from cruise speed to a virtual standstill in the first 30 seconds or so, then went from about 80 mph up to over 100 mph down. None of that would have been interpreted as normal horizontal flight."

There is some confusion among posters here, and in the press, about the flight path. The pitch attitude is reported to be around 15-16 degrees nose up, and the angle of attack to be 35-40 degrees. This translates to a flight path angle of -20 to -25 degrees. If the vertical speed is around 10,000 fpm (~100 knots), then the slant speed has to be the vertical speed divided by the sine of the descent angle, or somewhere around 250 knots. This is the true airspeed. The airplane was not fluttering down vertically like a leaf, and its foward speed was not 60 knots or 107 knots or whatever. As the flight track graphic in the interim report (vertical dimension exaggerated) shows, it was in a stalled glide with a comparatively moderate angle of descent and a good deal of forward speed.

Slightly O/T....
I agree with your basic remarks.... Not for nothing do glass engine instruments still show replicas of the original 'clocks'.
But the altimeter is a bad example, with the old "hours-minutes-seconds" clock scale (you know what I mean) not being all that fast and easy to read, and being implicated in several incidents (I don't remember accidents off-hand). Which is why the numerical 'drum' scale was added, even before we went to 'glass'.
Unfortunately, with the conversion to 'glass', just about all the moving pointers disappeared.

Humans are analog oriented.... For all the non-pilots on here.... just try to imagine playing a video game where all the game information is in the form of numbers on the screen.....

No, it is just as likely (or more so) to be an instrumentation error.

I absolutely do not wish to start an Airbus v. Boeing thing - no interest in playing manufacturer top trumps at all.

However, I would be genuinely interested to know if and how this accident scenario might've been different - in terms of aircraft behaviour and what the pilots were presented with - if the aircraft involved had been one of Air France's 777s rather than one of their 330s?

Specifically in terms of autopilot/autothrust disconnection, stall warning being silenced <60KIAS, lack of AoA information, trim issues? Do you think the 777 would've been a more helpful/intuitive aircraft to grapple with in those desperate moments?

Again, not trying to stir it, just interested to know how much the aircraft type and design philosophy is to blame here (if at all) and if an alternative design might've helped those guys more...

Didn't some Airbus fanatic write a book in which he said of Sullenberger's accident "It was the aeroplane that saved them. It cradled them all the way down to the river." Has he said anything about this accident?

I'm a scientist, not a pilot.

It seems that the biggest source of uncertainty is what did the pilots know and what was their assessment of the situation.

There seems to be a lot of debate as to whether they were aware of or had reliable data about speed, pitch, aoa, etc. But, we know that they were aware of their altitude and rate of descent as they passed through 10,000 ft. Even so, they continued to pull back on the stick. Is it possible that they believed they were in a steep dive rather than a stall?

Is it possible to get a stall warning when pulling out of a dive?

Isn't a dive is more recoverable than a stall? With "no other indications" is it not possible to fool oneself into believing the more hopeful scenario?

Because without that information, some of the protections can't work. Any protection out is an immediate drop to Alternate Law 2, which puts the pilots in near-absolute command of the control surfaces (any remaining protections can be overridden with sufficient control deflection).

It may surprise many, but it was part of the original specification for the systems that drive the A320 and her descendants that if, for whatever reason, the computers think they can't behave as they should, they defer to the pilots - on the understanding that they have more information available to them, especially in daylight when there's an external reference as to which way is up. The human brain can at least try to work with partial information to rationally attempt a solution to the problem at hand, whereas a computer can't.

Also, I think you may be getting the FCU (Flight control unit - i.e. the FBW part of the design) confused with the FMC (aka A/P and A/THR), which is much like that of a conventional airliner).

This would add unnecessary complexity to the system. In fact the A/P can be engaged again in Alternate Law, but with airspeed data out, it's not considered a good idea. I know I keep harping on about this, but if you look at the Birgenair 757 incident, you can see what happens when an autopilot tries to fly the aircraft with a blocked pitot tube (and that was on a calm night, with little or no convective activity for miles).

I've lost count of the number of times I've heard "Switch the damn computers out and give control back to the pilots", or words to that effect. That is *exactly* what happened in this case. Unfortunately, with no airspeed data over water in the middle of the night and threading their way through patches of very inclement weather, even the best and most experienced pilots can be overwhelmed.

The main issue seems to be the uncertainty about the human factor, it appears. For that we'd need to hear the entire CVR contents to draw any conclusions. I'm quite sure there is a lot of stuff on there which would be immensely helpful, but which we are unaware of at this point.

FDR plots would be nice too, of course :)

Hi,

We know ?

No evidence of cb's?
 

"For the umpteenth time, there is no evidence the crew flew through a Cb. In fact, based on the last BEA report, the crew knew very well what was coming,weather-wise as they turned to avoid and briefed the cabin crew.

Had they then encountered mod or severe turb in the top of a Cb, I'm sure the BEA would have mentioned it."

NO EVIDENCE OF THUNDERSTORMS? IN A MAX WEIGHT CONDITION BARELY ABLE TO KEEP LEVEL AT 35,000 THE PLANE SUDDENLY ZOOMS UP TO 38,000' AT THE SAME TIME ALL THE BELLS AND WHISTLES START GOING OFF DUE TO FROZEN PITOT TUBES -- WHAT ELSE WOULD MAGICALLLY CAUSE THIS?

EVERY ONE ELSE IN THE AREA DEVIATED EXCEPT THEM. WHY?

IT IS ALSO WELL DOCUMENTED THAT A LARGE LINE OF CBS WERE DEAD AHEAD.

From pointers to drums... From dynamics to numbers...
 

The addition of the 'drum' scale to a single pointer was a good thing. But that is where the improvement should have stopped. Removal of the 'last' pointer completely degraded the 'altitude trend situational awareness'. Especially on a rough ride...

So, to close the circle again, for the removal of the 'trend situational awareness' the trend vector was 'invented'... euhh...? Which 'as such' is a nice additional feature to have... Not as a replacement...

Automation paralysis? Inexperience? Temporary Insanity?
 

The Flight Data Recorder recorded airspeed and altitude as displayed on the left primary flight display (PFD), and on the integrated standby instrument system (ISIS).

"There was an inconsistency between the speeds displayed on the left side and the integrated standby instrument system (ISIS). This lasted for less than one minute."

Even if the pilots were confused about momentary airspeed deviations and had no longer trusted the displayed airspeeds: It's a mystery as to why the Pilot Flying would pull back on the stick and climb from FL 350 to FL 380 during multiple stall warnings. Elementary, basic flying instinct learned from day one in flying school, should have made him do just the opposite. That is, to get the nose down, not up during a stall warning.

Curiously, the captain in the Colgan Air DH-8 crash in upstate New York had also pulled back on the yoke during the stall warning and active stick pusher...

Inexperienced pilot graduates from the same school...? :{

To override the NU of the THS, in Alternate Law, one pulls back on the SS.

Yes? Was there confusion about the origin of the chronic NU on the way down?
I would assume so. To clear this, does one pull back to the stop and release?

"during" and "continuous" mean different things, yes?

If the cg was even partially ng, could the PF have confused the a/c insistence on climb with a 'stuck' "Law" command? One which he continually tried to 'Clear'? With intermittent full back ss to override??

see touch 'n oops below

nitpicker330 BACK TO THE BOOKS.

For someone who is "supposed" to know their aircraft, you really do spout a load...

What is that supposed to mean???

Don't you mean:
Alternate WITH protections
Alternate WITHOUT protections

AUTOTRIM IS AVAILABLE IN ALTERNATE LAW TILL GEAR DOWN WHEN FLIGHT CONTROL LAW TRANSFERS TO DIRECT!!![/COLOR] :D

AUTOTRIM IS NOT AVAILABLE IN DIRECT AND MECH BACKUP.

NOW LEARN YOUR **** BEFORE YOU START BASHING OTHERS!!!

YOU MUST BE WELL KNOW FOR BEING A RIGHT ONE AT YOUR COMPANY :rolleyes:

We know almost nothing. So little data has been made public that making any assumptions and conclusions is quite premature, if not simply stupid. Nothing can be said for sure without going through the entire CVR transcript, while at the same time looking at the FDR data.

I have an unrelated question. One of the error messages transmitted by the aircraft was
Is there any explanation why TCAS would fail? It seems completely unrelated to airspeed...

@jcjeant - I'm just going by the bea rpt. The PF says "we're going to arrive at level one hundred." The report also says "the inputs made by the PF were mainly nose-up." It doesn't say that the last, simultaneous, inputs were nose down, so I assume they were nose up.

Still, my question is whether they might have thought they were in a dive?

Guildenstern:Absolutely.

It's called "high speed stall" or maybe "accelerated stall", and it occurs any time you're asking the wing to work beyond its max CL. The attitude of the aircraft is immaterial.

In fact, what we usually call a common stall (1.0 g) is a special case; in general a stall occurs whenever (mass x g) exceeds the wing's aerodynamic lifting capacity.

There is ample opportunity for confusion, since at different points in the report ground speed, AIS, and the erroneous speeds on the displays are mentioned. To understand what they are saying you have to keep the context of each separate in your mind.

As far as the forward speed, it was 107 knots when last recorded, according to the report. I assume that's GPS derived, so it should be accurate. However, that was ground speed, not the speed through the air. For your 250 knot estimate of airspeed to the true, that would imply a headwind speed of around 120 knots, which is unlikely. It was a thunderstorm, not a category 4 hurricane.

The report also mentions that the AOA in the final moments of the flight was always in excess of 35 degrees. It did not say how high it reached, nor what it was at the end of the recording. Thus, your assumption of only 40 degrees is probably an underestimate.

To put a perspective on it, if the wind was calm, then the actual flight path would be about 45 degrees down, AOA of about 61 degrees, with an actual speed through the air of about 151 knots.

As for TCAS failing: might be completely unrelated to events. We sometimes have it momentarily failing in the CRZ and then rapidly coming back soon after.
However (speculation) it could be due to enormous amount of static (St Elmos) around the direction sensing receiving antenna etc.

In the grand scheme of things it's completely irrelevant to proceedings.

There but for the grace of God.

ok you're in the cruise and the pitots ice up. airspeed/mach will stay at current values. Crew decides to reduce to turbulence speed., thrust reduces to achieve this. As actual a/c speed reduces autopilot starts to slowly pitch the nose up to maintain altitude and whilst doing so applies nose up trim. Maybe the thrust even reduced to idle at this point.

This continues until ths is at aft limit (remember the indicated speed/mach are still at original cruise values but the actual aircraft speed is much lower. Autopilot and autothrust disengage and a/c reverts to Direct law giving immediate pitch up due trim and a/c gains circa 3,000 ft and then enters dynamic stall. Presumably this zoom would give a slight reduction in static pressure, therefore the indicated speeds would increase.

The aspect that really interests me is that neither crew member noticed the increasing pitch attitude as the speed was reducing. What would the flight director be commanding throughout all this? If during the zoom upwards it's showing fly down to regain the selected flight level this might explain the initial side stick forward to regain.

If neither pilot wasn't paying any attention to the attitude then where were they looking?

Coming back to the issue of pitots getting iced up I still don't understand with modern design why this is a problem. There are plenty of other jet transports flying around the world where, as far as am aware, this has never happened. What is intrinsically different in the design of the system on this a/c?

one of the things we learn in primary flight school is "hazardous attitudes". One of them is, iirc, "that can't happen to me". I read a lot of that one between the lines here. Scary.

I am fairly certain that either of the three pilots on board knew, at least in theory, how to recover from a stall.

btw: has anyone ruled out severe icing on the wings and stabilizers?

TCAS needs to know how fast they're going. If it doesn't it only has half the information (i.e. where they are) required to transmit .