pull vs. push
 

A mixup (PF wants to push and pulls instead) is out of the world.

During my instructing air combat maneuvers in the F4 it never was very difficult to get the students to pull on the stick hard enough, but it was work to prevent him to exceed the allowed g-limit.
On the other hand it was hard work to train the students an effective extension maneuver with close to zero G. At the beginning most just relaxed back stick somewhat and felt unloaded with still 1.5 G on the frame or thaught they had a good unloading maneuver going with 3/4 G.

Same for the duration of the maneuver. Whilst students pulled all the way into the buffet and staying there all day until speed was gone, i never ever had to tell one to terminate his unloading maneuver. Those where most times too short and therefore not effective enough. We taught to unload for 5 seconds, most ended after 2 seconds.

Situation as described above where agrevated by a nose high position, where a hard pull disipated more speed rapidly and the unload had to be way longer until the nose was below the horizon.

It is natural behaviour to pull more and longer than needed and not to push hard and long enough to get a desired unload of the AC, at least in fast flying jets.

How often does an air transport pilot push (not relax) on the SS and feel less than 1 G? And how often in an unplanned situation under stress?

Hi franzl,

Thanks. Very interesting comments regarding pilot sensed g's.
- Never in the simulator, where he did his approach to the stall training.

In the simulator, apart from mild turbulence, there is no sensation of vertical acceleration (motion legs only have about 3 m of travel). I have witnessed over controlling during TCAS events due to the lack of the sensation of delta g. The "sinking feeling" on entry to the stall would be something the pilot had never experienced on type or in the sim.

Push the trackpoint, pull the joystick...
 

Hi JD-EE
I was fortunate to spend a year working in a college near Buffalo, NY. One of the professors there was a proficient touch-typist in both QWERTY (his laptop PC) and DVORAK (his desktop PC). All he had to do was to hit an 'alt+key' sequence to switch between the two modes and start typing. Hence, I don't think a pilot is going to confuse his side stick with his track point. That said, however, I am concerned about the switch from the LHS seat to the RHS seat. What amount of 'hand flying' experience are the pilots given in each seat?

Sensed motion. Presumably the initial falling sensation could have been misinterpreted as turbulence, after that once at terminal velocity they would sense near enough 1 g wouldn't they ?

JD-EE
This is nothing to do with circling (thermalling) in a steady moving air mass (which was the question).

It seems to approximate a description of dynamic soaring through an altitude range with a wind gradient i.e. the flight of the Albatross (which wasn't the question)

=== Retired F4 ===

A useful observation (about low to zero 'g' sensations) I think. I also think there is no chance the pilot got his 'ups & downs' & 'push & pulls' wrong - if he did there is one massive question to be raised over the whole airline industry for years to come :ugh:

This is exactly the situation with weight control Vs aerod control (bar Vs stick) aircraft.. one pushes, the other pulls for the same effect.
I found when teaching some stick-centric pilots would get it wrong... whereas I never could imagine a situaton, in an aircraft or on a computer sim of getting it wrong, despite flying h/gs and powered h/gs lot, despite initially flying with a stick (from age 16) how could one not make that switch immeditaely and correctly... ? Proviso, as long as one kept a picture in your minds eye.. so are we saying there may be people flying by rote, like learning things parrot fashion instead of understanding the basics and working up from there, the exam tick box generation OMG

No, that has to be ruled out, PF meant NU, UP, CLB, SLOWER or he had already given up focusing on the task of speed & pitch stabilisation by then.


As RetiredF4 suggests between the lines, unless you have been in a stall and been taken through the ND inputs, wait for pitch and speed to come back, steady (constant 'g' if poss) recovery and pullout etc THEN I suppose, how would you know what was required, or indeed, how long it might take.. :hmm:
IMHO - It is unonscionable that pilots have not acquired high manual handling skills in conventionally controlled aircraft before being accepted for Commercial Flying of almost any nature.
Fiddling with the stick for a few seconds to see what happens with the stall warning, if that is what we think may have happened, is the sign of trainee, not an experienced airline pilot

I am not suggesting PF nor PNF had or hadn't those skills

Hand flying
 

andianjul

From the responses that I have had, so far, " not if I can avoid it."
Some earlier generation training aircraft had the throttle in the left and stick in the right hand - others had their hands and controls the opposite way round. I do not recall that it was a problem. Some of the transport aircraft had nose wheel steering from the Left seat, others had auto-pilot controls out-board of the Left seat. Whichever pilot whose " leg" this was, would sit in the left seat. Until, in the UK in the mid sixties, we became "seat-rated". Training Captains and few Senior F/Os, who could act as Second Captains were allowed to sit in either seat, checked on alternate Base Checks. Flying with a very Senior Captain ( non Training) who was operating from the Right hand seat, I asked him if he was Right hand seat qualified. I knew that he was not.
He told me to B#### - Off. ( Should I take offence, NOW ?)

Having had to do my hour at a time hand flying, because the A/P seldom worked on some of the aircraft that I flew in the fifties, I still tended to hand fly to F/L10.0 when not too busy. Approaches were all manual, except for just two in very clear conditions - the aircraft only had a single channel. ( A " Practice Emergency G.C.A.", down to touch-down, whenever practicable would be good for all of us.)

JD-EE;

The sidestick position symbol is only displayed on the PFD during takeoff.

I know about the small "stick" in the middle of a laptop keyboard - I think the HP has the same arrangement.

I think there is no connection between the way another control device might work and either the yoke or a sidestick in terms of muscular or cognitive habit which would confuse or alter a psychomotor response, even under stress.

If anything, the response would more likely be exaggerated, but in the correct direction. In fact I have always thought that that's where the initial two stall warnings came from right after the AP disconnect...a strong, quick pull on the stick, followed by a relaxation, (unloading, to use Retired F4's term). The continued aft stick after the stall entry is another matter I think. One considers the Airborne Express DC8 accident when thinking about this. (That said, we will never know why the Colgan captain pulled instead of pushed the stick; the explanation has usually been the notion that on that particular design, the T-tail can stall in icing, (several (NASA?) videos were referenced) and the unloading of the tail (stalled tail) would pitch the nose down. The response was to pull back, (not sure why, if the tail was "stalled" but there it was)).

I've flown in both seats while teaching and there is a difference when switching sides if one doesn't do it often. We've had the discussion about who was sitting where and why, (who replaces captain, licensing issues, First BEA Report, etc).

While possible, I don't think that played a decisive (primary) role here. But because sitting in one's familiar physical place does have a psychological and even physiological effect, (IOW, it does make a difference), I think the question of who was sitting where is still important for secondary, cognitive reasons. When one is not in one's normal seat, one needs to intellectualize responses somewhat, (think about the hand's position when reaching for a switch on the overhead panel, etc etc), rather than operating out of pure habit.

The poster who worked the mouse and typed with opposite hands makes an interesting point. There may be some minor similarities between this example and cockpit psychomotor skill but I think that that example doesn't fully translate in terms of analyzing or fully explaining cockpit behaviours. It doesn't take into account thorough training and the establishing of behaviour through long and frequent repetition both of which greatly reduce the effects of high stress and not being in one's familiar place.

That isn't meant to dismiss the effect; there is the need to examine it. My view has always been, the loss of airspeed information wasn't itself an emergency but it quickly turned into one due to the pitch up and we need to know why that happened. Large physiological responses are natural when under stress; is that what happened here? Perhaps the next BEA Report will address this, if so.

Part of the reason training is so thorough and realistic (up to a point, as sims can't reproduce 'g' or aircraft behaviour in unusual attitudes), is to reduce the effect of the sympathetic nervous system "fight or flight" response and provide ways built through habit, of controlling the stress response.

It is known that the stress response has a natural "pace" or curve where heightened senses and acute physiological responses quickly diminish under high, continuous stress. IIRC, the report on the Alaska Airlines MD80 stabilizer jackscrew accident off the California coast near LAX mentioned this. Whether this is a factor here would be something for the final report to comment upon.

To your question, I think an instinctual response which "confused" push and pull is not likely, given the complete/total absence of such a control movement/requirement in all aircraft. In other words, there are no QWERTY and DVORAK aircraft! ;)

kind regards.

grity, about that right hand left hand thing you mentioned ...

Ten years ago, I was made aware of the risks of carpal tunnel syndrome.

I use computers a lot.
I am right handed.
I chose to begin to use my mouse left handed. (Still do)

It took very little time to adapt to left hand.

I had decided to play a favorite computer game (about eleven years ago) that I used to play right handed, left handed. It was very mouse intensive. (For those interested, it was a dungeon crawl called Diablo, not a flight sim) I found that while I initially was a bit clumsy, it wasn't long before I was able to run the little computer animation around the screen and activate mouse button commands with little problem.

Due to the set up of power and stick on most planes I flew, I was never able to try and teach my left hand how to fly that way. I did however, get to be good at gently flying with my left hand while writing things down on my knee board (right handed) very early in my flying career.

Put another way, with enough practice, left handed flying seems to be a teachable skill ... but you need to practice it. How much time does one get to do that? I've yet to see any of the SS pilots complain about having to fly left handed on the stick. Is this really a problem? (I taught a number of left handed people how to fly, in the RH stick LH power set up, and they seemed to do just fine). I doubt very much that PNF not taking the controls sooner had anything to do with "I don't fly very well with my left hand" concerns.

PS: I broke my right hand in 2004, punching something (no, not someone). I still had to go to work and type reports. So, I did, with just my left hand, for a few weeks. It didn't take that long to adapt, but it was slower.

Conlcusion? Adaptability seems to me the rule, not the exception.

PS#2: JD-EE.

Helicopter pilots who initially flew the V-22 reported some non-intuitive monkey skill issues with the flight controls of that aircraft. Had a long talk on that score some years ago with a test pilot who had flown with me in another squadron. Bottom line, took a bit of getting used to, but adaptation wasn't major. You can probably do a search for the V-22's teething problems, and maybe using "ergonimics" or "flight control differences" search terms to find some of the articles written about this a couple of decades ago. Might answer your question.

The test pilots at Patuxent River and the Marines in New River did revise twice, at least, the proposed training pipeline mix of rotary wing and fixed wing (from initial training) for their prospective Osprey pilots. The multi-engine training balance was increased on one of those reviews, as they felt the helicopter bit was over emphasized.

I am not sure what they have done since, been some years since I was (tangentially) involved with such programs.

Also worth noting:

The initial flying in a helicopter, if one began in fixed wing, could be confusing since power forward, push with left hand, in a plane is the same as power reduction, push with left hand, in a helicopter.

Howerver, you usually translate into helicopters in a VFR training environment where pull is up and push is down with the left hand. That, and being taught a collective isn't a throttle, doesn't take long to adapt to.

By the time you are in flight above translational lift, or in instrument flight regimes, you have already learned what the push or pull does, and the confusion does not arise. If it had ever been present,,and remained, the instructor would probably suggest you go back to flying fixed wing. :)

stick and momentum
 

- We never had a left-handed side stick.
- We never had problems with lefties using a stick mounted just above and to the right of one's right knee

Many of us were worried about our newbies ( Ell Tees right outta pilot training) forgetting how to fly a "real" plane. It didn't happen, and many went to other jets and commercial airliners. The Reserve unit I helped check out had many commercial airline pilots, and they flew the Viper and a host of heavies every week. We always thot the Viper would be the easiest plane for someone to hijack, as the average pinball wizard could simply "point" the thing. I have a feeling that the 'bus is harder to fly than the Viper due to momentum.....

I know my assertion of duplicating the stall entry is brash. However, JD and others have done energy calculations and such that seem to support my assertion of entering a stall before the confusers can react. Unlike the Viper, the 'bus appears to command less than one gee according to pitch. e.g. 30 degree nose up would be a 0.87 gee command ( also commands more than one gee if in a bank). Our little jet used a pilot-commanded gee. So Retired's example of unloading is a good one. Several of our pilots would trim for zero gee before entering a fight. Letting go of the stick resulted in a 'perfect' zero gee "extension" to gain energy. Conversely, at extreme pitch attitudes and a one gee trim, the jet would slowly raise the nose to achieve one gee.

The point of "feeling" the reduced gee may play a role here with the pilot back stick input. All the pilots here KNOW, they KNOW that you cannot trust your senses in IFR or even a dark night. So the old saw about pitch and power procedure has legs.

Lastly, most of us with lottsa hours have prolly had a static or pitot failure due to freezing. Easy to recognize and to apply the "power/pitch" law. Some of the new jets don't have pneumatic tubing directly connected to the displays or "meters" - the displays are electronic so the raw air data is converted to electrons along the way. So is it an electronics problem or a real lack of pressure in the tubing? Worse, the confusers use the electronic signals in any FBW system. Our system had separate pneumatic sensor systems - one for the FBW system and the other for navigation and weapon delivery purposes. Our basic airspeed indicator was a steam gauge!!! The HUD and other displays were electronic conversions.

Hi,

Can the pilot think he was in a inverted flight ? :)

I haven't been impressed by your earlier posts, but this one takes the biscuit.....
Quite apart from the fact that everything indicates so far that the attitude displays were functional.
"Weird and wonderful" remarks such as yours only "pollute" the discussions.


PS.... in (roughly) -1g inverted flight, he could have been blinded by the sandwich crumbs, dried Brazilian mud, a few pencils and pens, and other extraneous matter, all "rising" from the cockpit floor.
Nice theory, but no sigar.....

Not to mention the coffee splashing about, and perhaps even some old cigar butts that had been stashed in some corners ...

1) PROBE-PITOT fault is not the result of this ADRs self-testing (two by two), it is the result of EFCS and AFS functions monitoring their respective CAS imputs. They are working with imputs from all 3 ADRs. CAS was rejected by those functions, triggering the following ACARs sequence by pointing to this PROBE-PITOT fault (total pressure imput).

2) My concern about ADR self-monitoring is to explain why ADR DISAGREE wasn't triggered following the PROBE-PITOT fault even if the three CAS were rejected by those monitoring functions. My explanation is that it was a "triple ADR fault"; hence, it was not detected at internal level. No "outlier" ADR could be rejected, being all erroneous while passing (or failing) the comparative test two-by-two.

This test condition required to reject the first ADR (CAS) is: 16 kt difference during 10 seconds.
It means that ADR 1+2 / ADR 1+3 / ADR 2+3 all passed or failed this test when CAS dropped. Hence, ADR2 was also affected like the two other recored CAS from ADR1 and ADR3; all 3 speeds went down from coherent and valid values to erroneous values:
- passing succesfully the test, hence staying consistent (c)
- failing the test, hence being inconsistent (d).

Example (CAS_1, CAS_2, CAS_3); T_1 = T_0 + 10 seconds:
.... T_0 -> (274, 275, 275) => all valid and coherent values
a).. T_1 -> (61, 275, 275) => ADR_1 ("bad") would be rejected due to ADR 1+2 and ADR 1+3 test.
b).. T_1 -> (61, 275, 77) => ADR_2 ("good") would be rejected due to ADR 1+2 and ADR 2+3 test.
c).. T_1 -> (61, 70, 77) => all values consistents (no fault) but 3 ADR erroneous.
d).. T_1 -> (61, 99, 150) => all values inconsistents (triple fault) and 3 ADR erroneous.

In case of (d), ECAM messages (FLR) are compiled into Current Flight Report (CFR) sent by ACARS when they are linked to the same fault (having the same ATA). The priority would simply be PROBE-PITOT fault over ADR fault during the correlation window.
In this case, ADR 1+2+3, being not displayed (no BUSS), it would nonetheless trigger three ECAMs:
- ADR 1+2
- ADR 1+3
- ADR 2+3


3) FCOM:
"If one ADR is correct but the other two ADRs provide the same erroneous output
or if all three ADRs provide consistent and erroneous data:
The systems will reject the “good” ADR and will continue to operate using the two “bad” ADRs."
Like I said before about this quote, something is obviously missing here:
The system could only reject the "good" ADR in the first case (one is "good", two "bad").
In the second case, the system will continue to operate using the three "bad" ADRs. This should have been added to this sentence. The consequences are the same in both case (erroneous output could be used), but the logic would be restablished.

It should be our case to be considered excepted that CAS was rejected by external monitoring functions => CAS monitoring (EFCS) and ADR monitoring (AFS).

The combination of Alternate 2, turbulence and ice crystals - likely present - were IMO considerable factors.

The way human (and animal) mechanical and control mechanisms work is that the control is conditioned by feedback coming from the object of the action, from the motion, and the result of the motion.

A simple experiment/example is the opening of an umbrella. If there is no wind, I can open the umbrella with a quick, short action. In wind, I need a stronger, and prolonged opening action, until I defeat the wind's reaction, and see the umbrella opened.

A similar behavior is shown by an experiment with a dog, or cat,, for instance an attempt to pull a toy from my hand. There will be a stronger, and prolonged action of the dog, or cat, if I don't release the toy immediately.

It's clearly a reflex built deeply in our motion control systems.

Back to the A/C, the A/C was in turbulent air, and thus the A/C's response to a certain control, could be delayed by the countering effect of the random direction and force of the turbulent air motion.

Additionally, it was instrument based flying, and the perception of the A/C response was through the instruments. Was that as fast as normal visual perception reflex?

If the PF's first action on the stick encountered such a delayed perception of the A/C response, the natural reflex - referred above - is that the action will be stronger, and prolonged, until the A/C is responding to the stick action.

An additional element, which could amplify the effect of the reflex is how the duration of the action on the stick is translated into the actuating of the control surface action.

Is the excursion/amplitude of the actuating of a control surface proportional to the duration of a certain stick action?

If that's the case, it's clear, that the longer duration, is equivalent to a stronger action on the stick, which adds to the already stronger stick action driven by the reflex.

Deprogramming the reflex mentioned above, and reprogramming the pilot's hand control reflex mechanisms when is at the A/C controls only, requires training, and practice, which may be different from one individual to another.... How stress worked in blurring the new reflex, with the old/natural reflex?

An additional perception factor was the angle of the ice crystals - likely present - hitting the windshield. Where they in an angle giving the perception that the A/C is nose down?

I hope that a transcript of the full cockpit sound recordings will be made available with the next BEA report, along with data from data recordings.

We'll never know what they thought. Question is could the aircraft, through failure or some human interface factor in design, have convinced them they were inverted.

Very unlikely. As someone kindly pointed out to me a while back when I conjectured that they lost attitude: in the roll axis, control was maintained (albeit with some large excursions possibly due to stall). Without attitude indication at night / in IMC, I don't think they would have have kept wings anywhere near level.

flightglobal
 

Just found that attached (link) report on "FlighGlobal" in "Safety" and "Recent Accidents" on the page that opens it says on the right column - "Last AF447 News from the web"
An Airbus Captain’s Take on the Air France Disaster | Autopia | Wired.com

I think rather interesting to have a statement like this.
Looking foreward to Friday

Hi Annex14,
Until today, there was not a single press article quoting a "real" A330 captain (one being not an anonymous). This one (article) is no exception as I can't believe that a real captain would really know so few about his aircraft.
Hence, interest is very limited.

limited interest
 

takata

correct what you say and I too agree that if someone has to say anything about that accident before the facts become public, he/ she should stand it by name !
Seen al the mess that was published in the past by the media I understand the reluctance to any report that even looks like that "Quality Journalism"

Nevertheless I thought it worth to mention and show the link, but also that - hopefully - the puzzle is layed to a picture on Friday.

jcjeant

No. There was no inverted flight. Nor backwards flight, nor spin.

A different look at the descent?

"Mainly Nose UP". Did they know they were Stalled? They did.

Then why the "Mainly Nose UP"? Especially with that rod? 10,000fpm?

The nose never fell below 30degrees NU? YES. BEA say so, and I believe it.

Wait. "Mainly Nose UP" means there were other inputs to the Stick?

Yes, of course. Mainly Nose DOWN. But the Nose? Yes, unable to drop below 30 degrees; less Pitch was UNAVAILABLE.

PITCH was at its lowest with stick back, or stick forward? Stick back.
Every pilot will push the stick to get the Nose down, and if 30 degrees had been the 'lowest' result of stick forward, PF would maintain ND input. Instead, it looks like Stick back was the default (preferred) position. What did the a/c do with ND? Probably go NU? Buffet like she was coming apart? Increase descent?

Guess? Something BAD.

Some possibilities show these guys did everything by the Book. Even the procedure at the STALL Warning. Both of them.

Alarming to me that when presented with a possibility that mitigates the 'PE', or is critical of the 'platform', the popular drift tends to migrate toward the PE.

I predict that with new data, these pilots will be seen in a new light.

Wouldn't there have been also ECAM messages in case b) or c) if during the correlation window of 10s the speeds would have fallen more than 30kts (which was very likely the case) ?

Hi,

It was just to emphasize that all this discussion about confuse push or pull or the inverse (choice yours) is useless ... as we know (for 99,9999999 % sure) that is not this possible type of error involved in this crash ...
Maybe no "good sens" .. but certainly "humor sens" .
Some seem's to had lost this particular sens :) or the skill to detect it :)

Bearfoil

Your input on this subject has been prodigious, however, on your last, I have to disagree. They did very little 'by the book.' Inter alia:

On seeing the UAS prompt, the SOP is to set the pitch attitude to 5 degrees nose up, then consult the QRH for the appropriate pitch and N1 settings for weight and Flight Level. Why the much larger pitch-up and subsequent climb, with its associated loss of energy?

Had they recognised that the aircraft was stalled, the PF should have maintained nose down input on the sidestick, augmented by manual pitch trim.

There was at least one simultaneous sidestick input, which gives the algebraic sum of both inputs, and is completely forbidden. The correct procedure is to announce "I have control" and take control by pressing and holding the red pushbutton on the sidestick.

This is the most perplexing accident I can recall. There seems to have been much confusion over what information was being presented to the crew, versus what was reliable. Let us hope for more answers with the next BEA release.

airtrens:

Quote:

---

An additional perception factor was the angle of the ice crystals - likely present - hitting the windshield. Where they in an angle giving the perception that the A/C is nose down?

---

Given that the flight was being counducted in Night, IMC conditions at high altitude, in the clouds ... if they were looking out the window that would be an airmanship problem far more serious that if ice were hitting the windshield. That said, the most likely theory seems to be, at that temp and altitude, ice crystals forming, not ice hitting the aircraft in the form of sleet or hail. From the data so far.

Infrequent:

Quote:

---

Without attitude indication at night in IMC, I don't think they would have have kept wings anywhere near level.

---

Agreed. Easy to lose it, particularly if they were in turbulent air. (Won't digress into how one could finesse a partial panel sort of scan using only heading ... that way lies madness! :eek: )

bear:

Quote:

---

Some possibilities show these guys did everything by the Book. Even the procedure at the STALL Warning. Both of them.

---

As Nigel pointed ou, perhaps everything except the pitch and power thing for UAS ... if more CVR info becomes available, we may see how far into that procedure they got/went.

Nigel:

Quote:

---

There was at least one simultaneous sidestick input, which gives the algebraic sum of both inputs, and is completely forbidden. The correct procedure is to announce "I have control" and take control by pressing and holding the red pushbutton on the sidestick.

---

If I remember, this was in the end game, when someone took the controls, perhaps to overcome a losing situation. (Possibly they pushed buttons as the conversation of “your controls” took place.)
Wasn't this about when one pilot said they were hitting ten thousand feet ... absent more CVR info, we don't know all of the communication about controls. The second interim shows me that there was a verbal component to the last (and ultimately futile :{) passing of controls.

Neptunus

"On seeing the UAS prompt......." The first input was NU, are you saying that was not an attempt to "SET PITCH" (at 5 degrees)?

"Why the much larger (larger than 5 degrees) PITCH UP?" Beats me, you are saying the "larger PITCH UP" was PF commanded? Not known.

"At STALL, the PF should have maintained ND". Not at all. He is trained to prevent loss of altitude. Cannot do that with maintained ND. The STALL warning is not STALL, something he has never experienced in this a/c.

"augmented by manual nose down trim". NO. They were not in the correct LAW to have been prompted: "Use Manual Trim Only".

"At least one simultaneous side stick input, which is forbidden" You are saying they were not attempting to maximize ss input to be certain they were at maximum deflection (authority)?

Without CVR (further) we cannot eliminate that this dual input was not sussed as a possible remedy for PITCH authority? That it was discussed, and attempted in spite of (horrors) it was forbidden?

The 5°/CLB thrust... memory items apply if "safe conduct of the flight is affected". Else the procedure is to "level off for troubleshooting".

I hope they weren't trained for NU inputs, as that obviously doesn't work all that well to prevent loss of altitude... Procedure at that time called for pitch reduction to 5°.

No, they just were not in a law that left them only manual trim, the wheel was still there to use. But why trim for nose down while pulling up?

Zorin

The first input was nose up. You are certain the a/c was not pointed down at this point? A NU input would be an attempt to level, if the a/c was PITCHED NOSE LOW. You also do not know the decision was not made that the safe conduct of the flight was in jeopardy?

This also ennables a further explanation of what occurred next, a lag in a/c response to the NU.

(at STALL)

"PITCH reduction to 5 degrees" . PF may have reduced his back pressure, but been unsuccessful at reducing PITCH. The a/c at STALL WARNING was at 6 degrees PITCH UP, so holding altitude at that attitude would be read by the Computer as a "stick back". Remember, up to this point, He WAS holding NOSE DOWN inputs. (Until STALL WARN).

Lastly "Why trim for NOSE DOWN when inputting NOSE UP"? He was NOT necessarily inputting NOSE UP, he may have been attempting a hold on to the altitude.

Also, he was not looking for TRIM, he was looking for AUTHORITY. He did not know the THS was full UP, and that he desperately needed to move it DOWN.

This is a gripe of mine. At unusual Attitudes (sic), why is the THS allowed to (commanded to?) migrate into a position where, if setting is unknown, it may be causing the problem PF is trying to solve?

Neptunus Rex;
We'll soon know more about the pitch-up. I don't think a ND pitch attitude is plausible but given the loss of data there may be a possibility of a small indicated height loss to which the PF may have then responded - we just don't know yet.

However, I don't think any pitch-up is indicated in the memorized items under these conditions, (level cruise flight, aircraft not at risk). Loss of airspeed information is not an emergency nor is the safe conduct of flight impacted so one would go to the next main memory item, "Level off for troubleshooting". As I have said from the start, the first response is to "do nothing" (except ensure the aircraft is under control, then get the QRH out, then communicate with ATC/FA's etc), because both speed and thrust were fine just before the event and, within controllable variations (using attitude information and setting thrust to what it was before the AT disconnected),

From the First BEA Interim Report:

http://www.smugmug.com/photos/i-GnmP...-GnmPntp-M.jpg

From an A330 flight crew training manual, (2007):

http://www.smugmug.com/photos/i-hvNc...-hvNcXRD-L.jpg

Because the displayed information may be erroneous, the flying accuracy cannot be assumed. Incorrect transponder altitude reporting could cause confusion. Therefore, a MAYDAY should be declared to advise ATC and other aircraft of the situation.

PART 1: MEMORY ITEMS
If the safe conduct of the flight is affected, the flight crew applies the memory items. They allow "safe flight conditions" to be rapidly established in all flight phases (takeoff, climb, cruise) and aircraft configurations (weight and slats/flaps).The memory items apply more particularly when a failure appears just after takeoff. Once the target pitch attitude and thrust values have been stabilized, as soon as above safe altitude, the flight crew will enter the 2nd part of the QRH procedure, to level off the aircraft and perform trouble shooting. This should not be delayed, since using the memory item parameters for a prolonged period may lead to speed limit exceedance.

PART 2: TROUBLE SHOOTING AND ISOLATION
GENERAL
If the wrong speed or altitude information does not affect the safe conduct of the flight, the crew will not apply the memory items, and will directly enter the part2 of the QRH procedure.

Depending of the cause of the failure, the altitude indication may also be unreliable. There are however, a number of correct indications available to the crew. GPS altitude and ground speed are available on MCDU GPS monitor page and RA may be used at low level.

For faulty ADR(s) identification, the flight crew may, either level off and stabilize the flight using the dedicated table in PART 2, or, if for instance already stabilized in climb, use the CLIMB table given in part 3. The trouble shooting will be more accurate, using the level off table.

LEVEL OFF AND STABILIZATION (IF REQUIRED)
The table gives the proper pitch and thrust values for stabilization in level off according to weight, configuration and altitude.

It must be noticed that, if the altitude information is unreliable, FPV and V/S are also affected. In this case, the GPS altitude, if available, is the only means to confirm when the aircraft is maintaining a level. When reliable, the FPV should be used.

If the memory items have been maintained for a significant period of time, the current speed may be quite above the target

If FPV is reliable, or if GPS altitude information is available:
--Maintain level flight (FPV on the horizon or constant GPS altitude)
--Adjust thrust according to the table
--Observe the resulting pitch attitude, and compare it with the recommended table pitch target.

• If the aircraft pitch to maintain level flight is above the table pitch target, the aircraft is slow, then increase thrust

• If the aircraft pitch to maintain level flight is below the table pitch target, the aircraft is fast, then decrease thrust

When the pitch required to maintain level off gets close to the table pitch target, re-adjust thrust according to table thrust target.

This technique permits to stabilize the speed quickly, without inducing altitude changes. If FPV is not reliable and GPS altitude information is not available (no means to ensure level flight):

Adjust pitch and thrust according to table, and wait for speed stabilization. Expect a significant stabilization time and important altitude variations.

We don't know what he knew. Nor what he saw in his field of view.
That's a good question, bear. Here's an idea for you.

The unusual attitudes (UA) law looks to be a form of graceful degradation that is intended to ensure that automatic (robot) inputs no longer contradict a pilot trying to recover from an unusual attitude. (Note, current hints suggest this law state was not active in 447). Mikelour posted a few pages back on some UA scenarios he used to teach in the sim, some of which demonstrated that you needed to remember to get that trim wheel working to get the plane back under your control. (He also caveats that he wasn't sure if it was a "simism" or not).

The design thus (it seem to me) includes the possibility that a haywire robot may contribute to UA. If those out of normal parameters boundaries are reached, robot is told "sit on yer hands, you naughty boy" and the pilot flies as manually as can be in this aircraft, to include trimming the nose with his trim wheel -- he has to, as the robot has been put into the penalty box for a two minute minor. Don't know if there is a flag or alert for UA law that lets the crew know "robot is in the penalty box, you are flying a man down" or not.

I'll hope that Svarin, if he'd like, can pursue further his concerns on unexpected logic failures or failure modes.

Can't wait for the Friday report. Hopefully some of the fog will be burned off.

I think your hope on training is sadly misplaced. See Stop Stalling | Flight Safety Foundation [reference posted earlier by others]

It seems approach to stall training (and possibly all stall training) has become "lose altitude and fail check", leading to NU inputs, resulting in stall.

I suspect we will add 447 to the list of crashes quoted by the boeing guy. At least the problem has been recognised and will hopefully now get fixed.

[I'm still interested in what caused the initial climb into the stall though - like Bear says, the seconds either side of A/P disconnect are the key].

While I'm waiting for Zorin, I'd like to suggest a possible (sic) explanation for what we know.

I trust the Pilots were doing all they could. I trust they were highly qualified. I trust them easily as much or more than I do an a/c whose bread and butter is autoflight, and not UAS.

After the STALL, and what I take to be a proper and trained PF response, we find that the THS is 'held' in almost maximum NU. For this, the Pilot holds NU, well after his 'recovery' from STALL WARNING. What on Earth causes a Horizontal Stabiliser to be planted almost to the stops NUP, and in a STALL, with an extreme r/o/d?

Let alone a Pilot with full back stick? As I mentioned above, my opinion is NU stick is what gave the PF the best ND response.

Pardon? There is no other explanation except 'He didn't know about the Stall'. What else could this attitude be? Nose in the sky, tail first descent in average 10k fpm? No other attitude explains the basic environment, Panel or NO, Viz or NO, QRH or NO.

ACARS be damned, the trickle of 'data' from BEA aside, I think I know what explains this, if not the climb.

Also the "Duet" on ss at 10k.

Bearfoil

No. There could have been significant, and changing, vertical currents affecting the flight.

Same answer as above.

Direct Law will display the Manual Pitch Trim prompt, because that is all that is available. They were in Alternate Law, so both automatic and manual pitch trim were available. Manual pitch trim has priority, and is faster. The system follows up and reverts to automatic when manual is relinquished.

The CVR should tell us the answer. However, the commanded inputs were predominantly nose up, which is anomalous, to say the least.

Not so. At "Approach to Stall" the training was to prevent loss of altitude. At "Stall" the technique is to lower the pitch attitude and accept a loss of altitude. Most A330 flights take place in Normal Law, where α Floor will prevent the stall. α Floor is only available in Normal Law, and is triggered when full back stick causes the pitch angle to reach α Prot.

[/QUOTE]

Captain Moody (nor anyone else) had never experienced volcanic ash causing all four engines on his B747 to flame out. Captin Sully had never experienced a double engine failure on his A320 at low level; nor had he previous experience of ditching a jet transport. Captain de Crespigny had never experienced an uncontained engine failure causing power loss, substantial airframe damage as well as collateral damage to multiple systems on his A380. Every once in a while, valiant members of our profession become involuntary test pilots. Mercifully, it is very rarely: thankfully, most of them rise to the occasion and become our heroes.

The correlation window that I was talking about is the one lasting 1 minute for PROBE-PITOT 1X2/1X3/2X3 and the four other "FLR" ACARS that were reported during the flight. When a failure (FLR) is detected, independently of what the ECAM is displaying in real time in the cockpit, the Central Maintenance System (CMS) opens a 1 minute window during which all the correlated failures (same ATA) are compiled into the same CFR (Curent Flight Report) and a single ACARS is queued for sending when it is closed - in between, each message header could change during this compilation following a particular logic. The full details about such a failure message are only available by accessing the PFR (post flight report).

This system is for maintenance, not investigation. What really matter is that, at arrival, the maintenance must look at the PITOTs and not waste its time for troubleshooting ADRs or the flight Computers which reported in fact this pitot's failure.

This particular "FLR" ACARS was sourced/correlated with:
- EFCS2 & EFCS1 = Electronic Flight Control System 1 & 2 = Flight Control Data Concentrator (FCDC) 1 & 2
- AFS = Auto Flight System = Fault Isolation and Detection System (FIDS).
Hence, during this minute, those systems could have reported other "FLR" with the same ATA that could have been compiled within this reported ACARS.

The CAS falling suddenly triggered two other CAS monitoring functions different from this ADR self-monitoring discussed here. Beside PROBE-PITOT ECAM, the other Flags and Warnings time stamped 0210 were correlated with this sudden fall of speed, whatever ADR (external) fault was or wasn't reported. The second class 2 "FLR" at 0210 (27-93-34) points at FCPC#2 (PRIM2) OR a connection between FCPC2 and ADR1; BEA correlated also the two Maintenance messages (FCDC1 & FCDC2) to this fault. Maybe we'll got more details from their last findings about system fault logic and cockpit ECAM.

With the ongoing discussion over the what the crew may have heard, seen or felt in the stalled environment, I believe the following should be noted:-

• Once stalled the AoA was never less than 35° and finished at 61°
• Pitch attitude was 16°NU at stall and at end of flight, though occasionally less
• Aerodynamic flow was abnormal/turbulent
• Likelihood of rain/hail striking cockpit windows was low
• Buffet vibration generated around wings and engine pods complete with noise

http://oi56.tinypic.com/ebeip1.jpg

The overall effect may have been perceived as a Vmo/Mmo event. That of course doesn't explain exactly how the aircraft was delivered into that situation in the first place.

Taking a step back from the UAS upset, it is worthwhile noting that the flight plan provided for a climb to FL370 at SALPU. The forecast OAT at FL350 between ORARO and TASIL was -46°C, but with their weight of 205 tonnes and high OAT, that higher level request wasn't made to ATLANTICO.

Add ITCZ to the bolded comments above, plus an OAT value probably warmer than forecast, and you have a receipe that should have "rung a few bells" - but it didn't.

In the AF447 Thread on 27 September 2009 I wrote, "Icing that took place was probably 'rapid' as opposed to 'gradual'".

Following the AP/ATHR disconnect, the aircraft entered a "zoom climb". The Stall Warning sounded briefly during the pitch up manuever and some ND inputs were made and attitude was reduced to 7°NU but the AoA was still rising and the SW sounded again. Further NU inputs prevented any normal stall recovery and once the CAS became less than 60KTS the SW stopped.

http://oi52.tinypic.com/2lwap0l.jpg

As I stated over 2000 posts ago, the software needs to be changed to sense both the CAS and AoA direction of change and the effect of NCD values to provide an unambiguous "STALLED!" warning on the PFDs. There should be no need for the "big red button" marked "SOS", because if there is, it is only a cover-up for lack of training and/or understanding by those charged with the safe flight of the aircraft and all those souls onboard.

Also posted many moons ago was a suggestion that once a UAS event at cruise altitude is detected, that the AP/ATHR combo continue in a NORMAL 2 LAW where pitot dynamic/static related inputs are replaced by IRU data and a pseudo CAS is provided to the AP. This will also cover the 300 foot static baro loss often noted in these events, and with due warning to the crew enable them to monitor the situation and oversee that NORMAL 1 LAW is resumed when ADRs are in agreement.

You can, "Lead a horse to water, but you can't ....... "

Notwithstanding, there may also be some underlying physiological related events which will no doubt surface in the Final Report.

While we make our final comments before the interim report Friday the 5 degrees nose up and climb power will only work prestall. Once stalled and had a 61 degree angle of attack in the descent at 10,000 fpm descent rate they need at least 30 degrees nose down pitch to recover. The 5 degrees pitch up and climb power wasn't applied when it should before the stall and they went way above that causing a full stall. Hopefully they will tell us Friday how they did this.

Hi mm43,
I'll add also few additional notes for those who are following this discussion:
When the second stall warnings sounded, at 0210:51 (see mm43 graph):

- @ FL375, aircraft was NOT in "abnormal attitude": pitch (~7°), Alpha (6°) and roll (10/12°) were NOT excessive, airspeed (215 kt) was NOT totally lost.... and THS (3NU) was NOT at maximum (14NU)... and of course, the aircraft was NOT stalled.

Everything seriously wrong started at this point and during the following 10 seconds:
- TOGA was applied (NU limited effect);
- Sustained NU sidestick orders;
- Pitch increasing from 7 to 16°;
- Alpha increasing from 6 to 16°;
- another 500 ft were added, reaching FL380;
- Speed falling from 215 to 185 kt;
- THS starting moving UP, (but reaching 13°NU only one minute LATER = close to the end of mm43 graph)
- The aircraft stalled at the apogee.... while the STALL, STALL warnings sounded at least 40 seconds more during the stall.

bubbers 44

Absolutely right. As I said, 5 degrees nose up pitch was the correct response to UAS. Once in the stall, corrective nose down pitch must be applied.

Let's see what Friday brings to help us understand this enigma.

Neptunus Rex;

According to the A330 flight crew training manual and the UAS drill, a 5° pitch up at cruise altitude is not the correct response to the UAS memory items. The correct response is to maintain level flight and troubleshoot.

PJ2

Neptunus Rex

"5 degrees NOSE UP PITCH was the correct response". (To UAS).

Both Zorin and I disagree. Zorin claims it was "LEVEL OFF and troubleshoot". My claim is that with NOSE UP per PF, he could have been trying to do either five degrees NU OR LEVEL OFF, as the siituation demanded.

In Turbulent air (an accepted factor), AutoPilot could have trimmed NOSE DOWN to keep the a/c on altitude (UPDRAFT). We don't know yet the PITCH @ a/p drop. With UPDRAFT, the AoA is increased independent of SA, and depending on energy therein, the NOSE UP would have an additional bias NOSE UP, perhaps adding to the unexpectedly ambitious PITCH UP when the a/c responded?

"Once in the STALL, corrective NOSE DOWN PITCH must be applied."

PF had been applying NOSE DOWN PITCH. At 6 degrees NOSE UP, he heard the STALLSTALL, selected TOGA, and apparently, with back pressure, attempted an escape from the STALLWARNING, as he had been trained.

At this POINT, the STALL WARNING is important, in the sense that had it continued, it is likely the PF may not have sensed the NOSE FALLING, felt or heard buffet, or had other STALLED cues. We don't know except the BEA said the PITCH kept increasing. From here, BEA have no applicable statements except to say the PITCH did not decrease below 16 degrees.

There is no 'STALLED' prompt inSTALLED. :ok:

As before, I am offering possibilities to keep my mind open; if others have witheld judgment, so much the better.

regds

bear

I'm waiting patiently. If all it takes is UAS training to keep this from happening again then do the UAS training so even the automatic airplanes can be flown with automation inop with a qualified pilot. I never flew one that the throttles didn't move and you just monitored so feel fortunate.

My J3's and Twin Beeches, Lears, biplanes and Boeings didn't have much of that magic stuff. Had autoland at the end but never did one outside the sim.

I prefer yokes and sticks because you can see what everybody is doing and it feels really good when you hit a bump and if your hand moves an inch or two you are still stable. You don't have to worry about potentiometers and all that electronic magic.

Some news?
On a reconstitué l'accident de l'AF447 au simulateur - Le Point
Had they accessed some leaks, or pure journalist assumption?
The time data are strange: Publié le 27/07/2011 à 23:59 - Modifié le 27/07/2011 à 21:15 (French time: UTC+2).
Is it credible one can hear ice coming out Pitots bouncing on cockpit? ("Dans le cockpit, s'ajoute le bruit de la glace se détachant des sondes.": ~ In the cockpit, you have to add the noise of the ice coming apart from Pitots)?