AF 447 Thread no. 4

[bearfoil]

Thread no. 3 can be found Here. (http://www.pprune.org/tech-log/452836-af447-thread-no-3-a.html)

As this series of threads has become of epic proportion, the following observation from jpete regarding search filtering may be of interest -

You can achieve a search of only the 4 threads by using the following search string in Google

ths af447 site:http://www.pprune.org/tech-log/

this will search for mentions of THS in the AF447 threads of tech log only. Just change the THS in the string to whatever you want to look for. Adding the site:URL end part is the magic that restricts Google to only searching in the tech log on here



I take your view re: Boyd. So did Schwarzkopf. Your impression is anti OODA. It is also anti (airline) pilot, for the Colonel had other precepts that made OODA, as he taught it, revolutionary. Autonomy, Agility, Mobility. (FFS) are you in Wiki? Have you read Coram? John Boyd's ACM? His "Creation and Destruction"? Read Pierre Sprey on Boyd. The "Acolytes".

Agility is both a cerebral and logistical norm in OODA. As such, it is being applied in Leadership, Military, and economic models. OODA can be deadly without the tools innate in the trained Warrior (combat pilot). It is more than turning inside the scissors, and wasting his tail.

Applied to fbw, it is a disaster, hence my original question. The crew in any modern airliner, to include Boeing, gets immediately into trouble since his vehicle is a thing of mystery in certain regimes. and as PJ2 says, "Wait.....Wait....." etc. That is the antithesis of OODA, don't you agree? Wait until the enemy does something you recognize?

Our PF was reacting to a screen, which is his intel, and he acted. Downhill from there? He wasn't in the loop at all, and his adversary may have been his platform.

Boyd demonstrated new ways to fly in genII a/c, the F-100, specifically. I would dearly love to see an improvement in man/machine. Wouldn't you? I think OODA is a point too far. I would love to hear gums on this, though I think he was in the Branch that disowned Boyd to their distinct detriment. Are you A Marine Pilot? The Marines have a statue built in is honor.

Conclusively, I think OODA would be a disaster for Air Travel. It would be expensive, there would be high washout, etc. Should we start by re-introducing Man to his Machine?

Trimmable Horizontal Stabilizer. For anyone who wishes, take a peek at the pic Machaca posted of the Inner tail cone where the engineer was on a ladder working on the RCU. Look specifically at the Jackscrew and suss which direction it must move to orient the THS in a specific manner. In the pic, it is in ND. To acquire NU, the hydraulic motor turns anti clockwise and the threaded collet distances itself from the fixed motor. As it's Arm increases, the screw itself is in greater torsional stress, and I might entertain that with the enormous feedback of a maximum effort deflection, some damage may result in the threads. If the forces were intermittent, and opposing, the critical point may have been surpassed. HazelNuts39 has posted a graph which should give anyone the willies.

Anticipating a familiar (knee jerk) response from those who have blind faith in the integrity of this airframe, I will even suggest a method to disprove. Without question BEA know the performance of this system, and the actuals will out. The most important evidence for mechanical failure is already in the Public Domain.

1.The THS remained in its near max position for seven miles down.

2. Any pilot who honestly looks at the evidence would say that cannot happen.

[Graybeard]

I was told the first 25 747-400 were built with wire wrap interconnects, and they had to run them all back through the factory to change them.

[DozyWannabe]

@bear - you've got a funny idea of what constitutes "evidence". "Any pilot who honestly looks at the evidence would say that cannot happen" is not evidence, it's supposition.

There are plenty of alternative explanations for the THS remaining fully nose-up that don't require mechanical failure to be a part of it, the only problem is that some of them will not be especially palatable to those members of the piloting community that are hoping against hope that the PF didn't make a fatal mistake.

With all due respect, you were clinging to the idea of a V/S failure when none of the available evidence supported it, and now you're hoping for some kind of THS failure when any supporting evidence has either not been released or does not exist.

Why is the idea that pilot error played a part (note : I'm not saying that it did) so abhorrent to you, and why do so many assume that there's some kind of conspiracy going on between the BEA, AF and Airbus to blame the pilots?

[jcjeant]

Hi,

Why is the idea that pilot error played a part (note : I'm not saying that it did) so abhorrent to you, and why do so many assume that there's some kind of conspiracy going on between the BEA , AF and Airbus to blame the pilots? I'm not certain that bearfoil assume it's some kind of conspiracy between BEA AF and Airbus.
But certainly the french government is a big player in the final issue ..
http://www.pprune.org/6513772-post2025.html
BEA ... "aux ordres" ? ... why not ...

[DozyWannabe]

@jcjeant : I pretty much answered the same question here:

http://www.pprune.org/tech-log/452836-af447-thread-no-3-a-102.html#post6513849

You only need to see the continuing resentment from some quarters to see that the controversies of 1988 did them little good.

[hetfield]

But certainly the french government is a big player in the final issue ..Always remember the Air Inter Strasbourg-Entzheim accident and the influence of the prefecture...

[HazelNuts39]

HazelNuts39, is the pitch angle trace in figure 3 from the FDR released info (so far)? How confident are you of pitch angle to timeline synchronization for your graph? The traces are mutually consistent. They are derived from the assumed vertical acceleration as described. It's an iterative process: assume az/g, look what trajectory and speeds it produces, and how that fits the BEA description. Then modify the az/g to improve the 'goodness of fit' in several iterations, until you're satisfied it is good enough to give a reasonable description of the probable course of events. It's a bit like straightening a table cloth, you pull one corner to flatten a wrinkle, and get more wrinkles in other places, except this has more degrees of freedom. Confidence is a function of 'goodness of fit' in the few corners that BEA gave us. An important caveat that I should have mentioned is that the available cL-alpha data cover only the operating envelope as limited by buffet onset (alpha-max). Everything beyond that is 'educated guess' or pure conjecture.

Is it possible to deduce from your great graphs at what point the zoom climb becomes 'ballistic'? Strikes me that a some point before the apogee airspeed and pitch wouldn't be be able to generate enough lift for level flight - and the only way out would have been a carefully managed pitch down and controlled descent past the apogee. Was FL375 @ M0.68 sustainable? Guess need to make assumptions about engine thrust and response time if not already full.. At what point on your graphs was/should the stall warning be triggered?

It was never clear to me exactly how the A340 airprox zoom climb was successfully recovered - pilot or AoA protection avoiding the stall warning using full thrust and pitch down?
The zoom climb is partially ballistic whenever 'gee' is less than one. The only way out is to reduce AoA below that at which the stall begins, and that is achieved when stall warning stops (except when IAS<60 kts). Airspeed and thrust are for later. Lift-wise FL375 @ M0.68 is sustainable, but to sustain airspeed the airplane must be put on a descending path (perhaps TOGA would maintain speed in level flight - I don't know). The stall warning threshold is the red dotted line - with valid IAS. BEA has explained the reversion when UAS, but that explanation has been put into doubt because it doesn't fit the Air Caraibes and other UAS incidents.(*)

IIRC in the A340 airprox zoom climb pilot action on the sidestick took the system out of High AoA protection, and the airplane would then recover itself stick-free.

(*)Note: A poster has suggested that in case of airspeed becoming invalid, the stall warning threshold is based on the last valid airspeed, rather than a low-speed value. In my mind, that makes more sense than BEA's explanation. So perhaps you should replace the dotted line in my graph by a horizontal line from the time of A/P disconnect.

PS:: My advice is to look at those traces qualitatively, and not to expect an angle to be accurate within a tenth of a degree.

[robertbartsch]

If BEA allowed AB to say that, after reviewing the data, no modifications to the A330 are needed, that would imply to most people that the accident was caused by pilot error; right?

If BEA intended to avoid making any conclusions in the last preliminary communications, they should not have allowed AB to make any statements since people now believe that BEA must also think that no a/c modifications are necessary.

[henra]

retiredF4:

IMHO Elevator ND or trimming the THS down, after this high AOA of 60° was established, only increased the drag of the tailplane but had no permanent effect to get the nose down to the horizon and to increase the speed as necessary.


It has to be noted that at an AoA of 61° Drag along the Flight path equates to Lift by Factor ,87. So if your tail creates 1000N of Drag along the Flight Path that will eqaute to 870N of lift. Therefore increasing drag on the Tailplane is a good thing. Adding ND elevator will increase camber which again will increase drag and remaining lift.
As has been pointed out lift above the stall decreases bút it doesn't drop to zero.

Slight excursion:
I highly appreciate the fascinating insights of @gums regarding the stall behaviour of the F-16 but it has to be noted that the two designs are very different. In an F-16 much of the lift post stall comes from the strakes (they produce vortex lift which is not susceptible to stall as opposed to traditional wing lift). They are mounted forward of the wing, thereby shifting the CoL forward post stall. That was one of the main reasons for the rather agressive (meaning restrictive, being between 15 and 25°) AoA limitation on the F-16 compared to more conventional designs like F-15 and F-14. High AoA shifts CoL forward on the F-16 therby further increasing longitudinal instability.
/Excursion end

The fact that the Stall warning came back after some Nose Down Input due to the IAS exceeding 60kts again also points to the fact that the control authority was not really lost.
The problem is that reaction probably was quite slow.
So in the heat of the moment it might have appeared as ineffective even if it wasn't.

When GA planes spin it sometimes takes multiple revolutions with opposite rudder to stop it. It is surely not easy to resist the urge to try something else. However those who follow the urge often pay the ultimate price.

On the other hand we still have no clear information if the pilots in this case ever realised their real situation and Flight Path being deeply in a stall. Reading the scarce information we have I have the nagging feeling they didn't.

[RetiredF4]

HeavyMetallist
@RetiredF4: Your analysis of the likelihood of recovering from a fully developed stall is based on some flawed assumptions, principally that a stalled aerofoil has "stopped working". Lift coefficient doesn't just fall to zero after the stall AoA is reached, it falls gradually (and may even have a second peak that's not that far off the stall value). At 60 deg AoA you may still have a lift coefficient half the primary stall value, so a stalled HS can still be generating a lot of lift. In addition the wing pitching moment (nose down) typically increases significantly with AoA post-stall. In principle a conventional design is still recoverable, given reasonable lateral control and enough height.

It is based on my own expierience out of high performance aircraft with a big ugly stabilizer....... which we could stall at less angle than 45° AOA. Fortunately we had a big heavy cannon in the front with lots of weight and favorable cg.

So it would be kind if you could proof it with numbers for the AF 447, otherwise its just an oppinion of yourself.

[henra]

It is based on my own expierience out of high performance aircraft with a big ugly stabilizer....... which we could stall at less angle than 45° AOA. Fortunately we had a big heavy cannon in the front with lots of weight and favorable cg.

So it would be kind if you could proof it with numbers for the AF 447, otherwise its just an oppinion of yourself.

franzl

Although in an F-4 with its tiny VS and the Anhedral/Dihedral I would assume the main danger was not so much the stall itself but the spin.

With the CG and aerodynamic Design of it I would expect rather swift Nose Drop at the Stall. The thrill being not to spin it (unrecoverable?) and therefore not to drop a wing.
Am I right ?

[jcjeant]

Hi,

Last OT on this sensible matter :)

@jcjeant : I pretty much answered the same question here:

AF447 Thread No. 3

You only need to see the continuing resentment from some quarters to see that the controversies of 1988 did them little good. If BEA allowed AB to say that, after reviewing the data, no modifications to the A330 are needed, that would imply to most people that the accident was caused by pilot error; right?

If BEA intended to avoid making any conclusions in the last preliminary communications, they should not have allowed AB to make any statements since people now believe that BEA must also think that no a/c modifications are necessary. We must keep in mind that all press releases and BEA release are mainly read by the general public and not only by the Pprune contributors ...
General public don't go on all kind of deep analyses on such press articles or BEA note ...
With only a few words and explanations .. they form their own opinion
Public sentiment in general by reading the note and press articles can't be ambiguous .. the way is clear ...

[Lonewolf_50]

bear, please stop mixing apples and oranges in your OODA soup, rather than any useful OODA loop. PD: I learned of OODA before the Endless September arrived. It has aged well.

The OODA loop is what it is. It remains as what Boyd first laid out when he was teaching how to improve air combat maneuvering. (Dogfighting, if you prefer). Of interest is the question of whose decision cycle prevails, robot or pilot, but there is only limited evidence to support them being in that much conflict in alternate law. It remains at some points an open question.

(Or was there more conflict? (http://www.pprune.org/tech-log/453832-a320-big-flaw-unreliable-speed-adr-check-procedure-3.html#post6514394)Not sure, appears to apply to normal law, which sat on the bench_.

Taking the OODA framework and applying it to other disciplines requires adapting the basic model and flow, since the patterns and value in the iterative process, and the speed of the iterative process, lead to a form of the continuous improvement model.

That doesn't change what the Decision Cycle is.

When you fly, you are a living breathing decision cycle the whole time you are flying ... otherwise, YOU ARE CARGO. That doesn't mean you have to act or decide fast, but you do have to make good decisions, and your Observe and Orient need to be accurate. If I observe and misorient, my odds of a bad Decide or Act move goes up.

Each time you get to the Act node you can decide to do as PJ2 suggests ... wait for the next cycle is an action. (This also works when sitting in an ambush and waiting for the right time to pull a trigger, light infantry style).

Or, per the spin recovery in smaller planes, wait for the second or third turn (and sometimes tightening spin rate) to see that your opposite rudder was effective ...)

The other point in Boyd is avoiding tunnel vision, which means not just look at your target, but your whole battlefield ... enough on OODA extensions.

Chris's recent posts suggest to me that the pilots in AF 447 may have gotten behind the decision cycle somewhere in between 35000 feet level, and 35000 feet on the way down. The 11 seconds or 45 seconds PJ2 and you were discussing a while back can lead you to a root cause regarding an Observe-Orient problem which led to Decisions and Actions that may have been based on faulty input, based on faulty pitot probe input.

OODA as anti pilot?

Stuff and Nonsense.

OODA came from pilots. It is what we do. (It is what flight computers to a large extent try to mimic, in their own special way).

That Boyd took an idea and went beyond its origin speaks well of him: he was one smart guy. He did that "out of the box" and "cross discipline migration" thing very, very well.

[CogSim]

Originally Posted by bearfoil
How often did the pilots train to the discrepancy in "touch" to Roll compared with Pitch? How necessary was the Rudder, if at all, to settle the Yaw produced by Roll excursion? I believe firmly the accident began at loss of a/p, and the corrections input by PF. It is very easy to entertain getting a bit behind, which makes it not difficult to question whether they caught up, and if not, perhaps the a/c and Pilots started down different paths? In a general way, I think this will be the fulcrum of the findings.

The movement of THS; 10 degrees in 60 seconds (if continuous) strikes me as very very gradual. In my mind this fits the bill of inadvertent NU input while trying to correct for right roll (especially if being flown from right seat).

OTOH, it is indeed puzzling how the pilots seem to display good awareness of bank angle (judging by a/c not departing into a spin) but not check the pitch attitude.

[MurphyWasRight]

henra:


The fact that the Stall warning came back after some Nose Down Input due to the IAS exceeding 60kts again also points to the fact that the control authority was not really lost.
The problem is that reaction probably was quite slow.

So in the heat of the moment it might have appeared as ineffective even if it wasn't.


A prior post (cant find it at moment) comments that the "ride" coould have been relatively smooth when fully stalled and would likely get rougher as AoA decreased when ND inputs where applied.

So we have:

A: Slow response to ND input
B: Possibly rough ride as AoA starts to get better.
C: Return of stall warning as AoA is improving.
D: Highly suspect air speeds, even though at this point they were valid (when over 30kts) but likely not beliveable: 60Kts, that cant be right, we wouldnt be flying...

Given the above items it is hardly surprising that the PF might have thoughts along the lines "well that is not helping things" and return to NU inputs.
Whether persisting with ND at that point could have saved things is impossible to tell from the BEA "note".

It is usefull to remember that BEA knows a lot more that what has been released and likely had a strong working theory when deciding what information to include.

They knew that when matched with final report the note would need to be seen as:

1: Factually correct
2: Not containing true but deliberatly misleading information.

A (made up) trivial example for #2 would be "then the PF sneezed", unless that was going to be a significant factor in the final report including it in the note would be seen as deliberately misleading even though true.

If read with the above in mind the BEA note likely provides clues to the causes beyond the bare stated facts.

[infrequentflyer789]

If BEA allowed AB to say that, after reviewing the data, no modifications to the A330 are needed, that would imply to most people that the accident was caused by pilot error; right?


It might if they had done that, but they haven't, so it doesn't.

AB released a statement saying that at this point in the investigation no new modifications were advised. That does not preclude modifications coming out of the detailed investigation further down the line. It also does not preclude modifcations that have already been advised, and in fact in this case modifications already have been advised (and in fact were advised before the accident - AF had not implemented).

Even if the only thing that went wrong with the a/c was the pitots, that is still the a/c as part of the causal chain.


people now believe that BEA must also think that no a/c modifications are necessary.

Only idiots can believe that since it directly contradicts the evidence - BEA have already recommended a/c modifications as a result of this accident (and previous incidents). BEA might believe that no further modifications are necessary - but that is an entirely different proposition.

[RetiredF4]

Henra
It has to be noted that at an AoA of 61° Drag along the Flight path equates to Lift by Factor ,87. So if your tail creates 1000N of Drag along the Flight Path that will eqaute to 870N of lift. Therefore increasing drag on the Tailplane is a good thing. Adding ND elevator will increase camber which again will increase drag and remaining lift.

Are you talking about a stalled airfoil or an airfoil in unstalled condition? I´m sure there is a lot of difference in it without being an aerodynamic specialist. Otherwise any object producing drag would also produce lift. A stalled airfoil at 60° AOA is nothing more than a door in the wind producing disturbances and drag of uncalculated ammount. I bet, even EADS doesn´t know what kind of lift the THS would produce at that AOA. It has never been expected and therefore never been tested. And even if it produced some lift, in what direction in relation to the airfoil would that be? How much of it would be directed 90° to the fuselage, that´s necessary to lift the tail, isn´t it?

The fact that the Stall warning came back after some Nose Down Input due to the IAS exceeding 60kts again also points to the fact that the control authority was not really lost.

It only points to the fact, that airspeed was not stable below 60 KIAS (if that speed was the real one and not some falsified indication by pitot icing or high AOA).


When GA planes spin it sometimes takes multiple revolutions with opposite rudder to stop it. It is surely not easy to resist the urge to try something else. However those who follow the urge often pay the ultimate price.

Why do you assume that there was some spinning present? Do you have any source? Everything is pointing on a stable high AOA descent without much rotational input or even rolling input.

Although in an F-4 with its tiny VS and the Anhedral/Dihedral I would assume the main danger was not so much the stall itself but the spin.

In comparison to the overall size of the F4 it was nothing compared to tiny. We had lots of stab- authority, up to the point of stall. Then it was gone. By the way, stall was accompanied by a dominant nose rise, indicating loss of lift (stall) on the stabilator.

With the CG and aerodynamic Design of it I would expect rather swift Nose Drop at the Stall. The thrill being not to spin it (unrecoverable?) and therefore not to drop a wing.
Am I right ?

Not quite. As already mentioned, nose drop due to cg and weight on the nose (canon or camera equipment) helped to get the nose down. The aerodynamic design however (swept wing) made it prone for turning departure to stall and therefore spin. Not an issue on modern GA aircraft.

I repeat our stall recovery procedure to picture the problem (the spin recovery was different):

- Stick forward (it was not full forward, due to danger to stall the stabilator)
- ailerons and rudder neutral (not to induce any rolling moment)
- if not recovered maintain full forward stick and deploy drag chute (this assumes, you go forward with the stick until you feel the reaction of the aircraft, If you dont feel it and reach full forward stick you are out of options with the stabilator alone. The dragchute establishes itself in the relative wind and causing a violent up-movement of the tail and bringing the stabilizer back into the working regime, it unstalls the stabilator)

The F4, and the F16 are no A330, but aerodynamics are similar and fundamental. I´m wondering why this knowledge of aerodynamic behavior of airfoils got lost between bit and bytes.

On the other hand modern aircraft like A330 have a very stable aerodynamic design and are flown with a managed CG. When those aircraft are flown into very high AOA, the aft shift of CG settles the ship in a stable nose high attitude, when not immidiate corrective action is applied.

Just picture the THS in its normal working regime, on takeoff, in cruise, on approach and during landing, then it is more then logical, that it was never ever built to recover the ship out of an 60° AOA. I dont know what it´s limit would be, Imho something like 30° and then its out of authority without some other help.

[MurphyWasRight]

R F4:uote:
The fact that the Stall warning came back after some Nose Down Input due to the IAS exceeding 60kts again also points to the fact that the control authority was not really lost
.

It only points to the fact, that airspeed was not stable below 60 KIAS (if that speed was the real one and not some falsified indication by pitot icing or high AOA).


The BEA note seems to me to strongly imply that the ND input did in fact cause the AoA to improve as would be expected, hence the input did have some effect.


Around fifteen seconds later, the PF made pitch-down inputs. In
the following moments, the angle of attack decreased, the speeds became valid again and the stall warning sounded again.

[Yankee Whisky]

I am waiting for the final report and documentation of the real facts
as saved on the flight recorders and circumstantial evidence.

Any speculation makes for interesting writing, but in no way reveals what
in fact caused the apparent flight crew control anomaly and the write off of a planeload of crew and passengers !

Too many "why's" remain unanswered so as to reach any kind of conclusion.

[jcjeant]

Hi,

http://www.bea.aero/fr/enquetes/perpignan/animation.d-axla.1024x768.xvid.avi

[CONF iture]

From the above, it seems fuel transfer to the tail is automatic once in cruise. There is no mention in the BEA info that fuel was manually transferred forward, so how did the CG get to 29% in latest BEA note, from 37% in initial reporting?
37% CG would be a big difference from 29% in manual pitch control. It's real easy to over control with aft CG.
The initial estimate of 37% was probably erroneous :
http://www.pprune.org/tech-log/452836-af447-thread-no-3-a-42.html#post6484163


I oversimplified the description
You did not oversimplify anything, you make up technical stuff on your own and obviously not able to back up by the appropriate Operation Manual reference when requested.
"The only way autotrim is activated is if the stick is fully deflected and held there past the elevator travel limit"
Give me your reference that I learn something … or retract it !

If it’s getting personal or insulting to challenge what appears to be unfounded statements, so I am.

And if you want us to discuss Habsheim, start a thread on its own, or stop making redundant innuendo.

[wallybird7]

More:
1. Why did the pitots ice up?
2. Why did the A/P and A/T disconnect?
3. Why the zoom climb without clearance from ATC?
4. Why the sudden reduction of speed at top of climb?
5. Why when at the same time the plane went into the thunderstorm?

Any connection?

The unanimous comment from my group of old retired pro's: Why didn't they deviate?

Comment: When an airplane comes out at the top of an uncommanded zoom climb almost 3000 feet above their max altitude below 60 kts, it is out of control! The flight controls do not have ANY effect. As we have seen until it hit the bottom.

Thunderstorms are known to contain strong up and down drafts. For 50 years we in the pilot community have been warned to avoid, avoid, avoid them.

[wallybird7]

full text of the classic FAA guide


DO'S AND DON'TS OF THUNDERSTORM FLYING

Above all, remember this: never regard any thunderstorm as “light” even when radar observers report the echoes are of light intensity. Avoiding thunderstorms is the best policy. Following are some Do's and Don'ts of thunderstorm avoidance:

Don't land or take off in the face of an approaching thunderstorm. A sudden wind shift or low level turbulence could cause loss of control.
Don't attempt to fly under a thunderstorm even if you can see through to the other side. Turbulence under the storm could be disastrous.
Don't try to circumnavigate thunderstorms covering 6/10 of an area or more either visually or by airborne radar.
Don't fly without airborne radar into a cloud mass containing scattered embedded thunderstorms. Scattered thunderstorms not embedded usually can be visually circumnavigated.</I>
Do avoid by at least 20 miles any thunderstorm identified as severe or giving an intense radar echo. This is especially true under the anvil of a large cumulonimbus.
Do clear the top of a known or suspected severe thunderstorm by at least 1,000 feet altitude for each 10 knots of wind speed at the cloud top. This would exceed the altitude capability of most aircraft.
Do remember that vivid and frequent lightning indicates a severe thunderstorm.
Do regard as severe any thunderstorm with tops 35,000 feet or higher whether the top is visually sighted or determined by radar.
If you cannot avoid penetrating a thunderstorm, following are some Do's Before entering the storm:

Tighten your safety belt, put on your shoulder harness if you have one, and secure all loose objects.
Plan your course to take you through the storm in a minimum time and hold it.
To avoid the most critical icing, establish a penetration altitude below the freezing level or above the level of —15° C.
Turn on pitot heat and carburetor or jet inlet heat. Icing can be rapid at any altitude and cause almost instantaneous power failure or loss of airspeed indication.
Establish power settings for reduced turbulence penetration airspeed recommended in your aircraft manual. Reduced airspeed lessens the structural stresses on the aircraft.
Turn up cockpit lights to highest intensity to lessen danger of temporary blindness from lightning.
If using automatic pilot, disengage altitude hold mode and speed hold mode. The automatic altitude and speed controls will increase maneuvers of the aircraft thus increasing structural stresses.
If using airborne radar, tilt your antenna up and down occasionally. Tilting it up may detect a hail shaft that will reach a point on your course by the time you do. Tilting it down may detect a growing thunderstorm cell that may reach your altitude.
Following are some Do's and Don'ts During thunderstorm penetration:

Do keep your eyes on your instruments. Looking outside the cockpit can increase danger of temporary blindness from lightning.
Don't change power settings; maintain settings for reduced airspeed.
Do maintain a constant attitude; let the aircraft “ride the waves.” Maneuvers in trying to maintain constant altitude increase stresses on the aircraft.
Don't turn back once you are in the thunderstorm. A straight course through the storm most likely will get you out of the hazards most quickly. In addition, turning maneuvers increase stresses on the aircraft.


Table of Contents (http://www.aviationweather.ws/)
Previous Section: Thunderstorms and Radar (http://www.aviationweather.ws/064_Thunderstorms_and_Radar.php)
Next Section: Common IFR Producers (http://www.aviationweather.ws/066_Common_IFR_Producers.php)

[RR_NDB]

Hi,

1. Why did the Pitots ice up? Because the a/c flew into a region with certain characteristics, equipped with 3 identical Air Speed sensors (Pitot tubes) not adequate for this "region". These sensors (Thales) presented some anomalies before (in other flights) and were scheduled to be replaced by BF Goodrich ones. The accident occurred before the change in the F-GZCP plane.

2. Why did the A/P and A/T disconnect?
Because the Airbus SAS System requires reliable information from the sensors in order to work at full specifications. When entering (a region with certain characteristics) and the sensors start to present anomalies, the System change the a/c operation mode (LAW) disconnecting A/P and A/T, etc. We don´t yet have information on other (possibly) occurred sensor anomalies that could also lead to this "disconnection" (due mode change, LAW).

3. Why the zoom climb without clearance from ATC? In this region they were capable to use HF and Data comm. They tried to use Data comm with DKR ACC for 3 times, some minutes before. During the sudden zoom climb they certainly were not able to even think on comm for FL change.

4. Why the sudden reduction of speed at top of climb? The a/c changed speed by altitude and with THS "going" to 13° UP with/and NU commands, she eventually stalled. And they were facing non reliable air speed in their displays and also in the ISIS (3rd and redundant resource). So the real speed (during many moments) after icing up nobody will ever know. Just estimate, based at other data, (risky estimate). The info was just lost (simply not measured, due "sub heated sensors" :}:8:E).

5. Why when at the same time the plane went into the thunderstorm? The decision making to go straight (and not deviate) was made before. "Inside" the WX they started a 12° LH deviation (http://www.weathergraphics.com/tim/af447/sat-goes-0215-ann.png), then the zoom climb started. We don´t have the CVR info to understand on the "decision" to go in the straight route. Other planes deviated. On this particular aspect see BEA reports and Tim Vasquez (http://www.weathergraphics.com/tim/af447/)

below 60 kts, it is out of control! Nobody has the "real" speeds during many moments during this phase. The sensors were just not working properly. The FDR will show us the moments when the speeds were correctly measured. (Again, the a/c was equipped with air speed sensors not compatible to the region she entered)

The flight controls do not have ANY effect. The (very little) (http://www.bea.aero/fr/enquetes/vol.af.447/point.enquete.af447.27mai2011.en.pdf) information BEA released on that suggest they had "some control".

Thunderstorms are known to contain strong up and down drafts. For 50 years we in the pilot community have been warned to avoid, avoid, avoid them. Several cases of Unreliable Air Speed occurred during deviations from WX.


(*) KISS principle - Wikipedia, the free encyclopedia (http://en.wikipedia.org/wiki/KISS_principle)



Why didn't they deviate? CVR certainly will help us to understand why.

[PA 18 151]

So I asked myself: how does one recover from a stall without reliable airspeed and without AoA? This is simple stuff, you use your remaining instruments and return to basics: use pitch and power. And you certainly do not continue pulling. I would think that the proper action would be a determined nose-down push, maintained until the stall warning stops, and then trying to maintain an AoA on the edge of s/w, gently allowing the nose to raise until s/w occurs, then a small nose-down correction to silence it, etc., until back in approximately level flight at a reasonable pitch attitude.Scary stuff,. You don't use the stall warning to calibrate your response, you use your remaining instruments and your knowledge of your aircrafts pitch/power settings. Your solution puts you at an unnacceptable risk of secondary stall and that is another opportunity to enter a spin. If you do not know your aircraft's pitch/power settings for level flight then you are not competent to pilot it/manage its autopilot and take over if/when that autopilot fails.
EDIT:: Can anyone be expected to do that succesfully without being trained for that eventuality? Of course not, you are not born with the ability to recover from unusual attitudes using partial panel. You learn this when you take your instrument rating. If you have lost your scan you are not competent to fly in clouds.

It really is that simple.

I'm struck by the fact these guys managed to "fly" and reverse course at what is essentially under minimum controllable airspeed, for five minutes, in turbulence, and they hit the water belly side down and reasonably flat. That is a huge credit to the designers of the airframe, must be inherently incredibly stable.

[HeavyMetallist]

@RetiredF4: As I mentioned previously, an aerofoil doesn't cease to be an aerofoil and stop producing lift just because it's at high AoA and stalled. Yes, it produces a lot of drag (component of force parallel to the airflow) as your barndoor would, but there is still significant lift (component of force normal to the airflow). Since you don't want to just take my word for it (fair enough), here's a link to a study specifically of post-stall aerofoil behaviour (http://docs.google.com/viewer?a=v&q=cache:fbMtXUlrOzcJ:wind.nrel.gov/designcodes/papers/OstowariNaik_PostStall44XX_Part2.pdf+airfoil+pitching+moment +post+stall&hl=en&gl=uk&pid=bl&srcid=ADGEESg0Do9z3A-8JK2pE-VmQEQEHAcJ5CIyuig9l9yd8NewuhIdC78nTXOcXmlp_jPBL-MoZwO9cI1qnmJT1dvR8At-4Bf6J436YITa7-gzu4fEQHBs_hIkeMLs35uUPfOGLCde16fm&sig=AHIEtbRu8yOF6KFBt3RzZndfZtB5fnZeyQ); it has a good set of Cl/Cd/Cm-alpha graphs going up to extremely high AoA. You'll notice that the motivation for this study wasn't aircraft behaviour, but wind turbines; there's been a lot of interest in post-stall aerofoil chracteristics because they spend quite a bit of their time operating in a stalled condition!

[rudderrudderrat]

That is a huge credit to the designers of the airframe, must be inherently incredibly stable.
Attitude stable yes - pity it was not speed stable.

[RR_NDB]

Hi,

That is a huge credit to the designers of the airframe, must be inherently incredibly stable.

And Airbus SAS learned something not "testable" before. The airframe designers received (from the FDR) an important information.

And now are being able to learn other very important "things" on the a/c.:}

[Dont Hang Up]

That is a huge credit to the designers of the airframe, must be inherently incredibly stable.


Stability in a downward 'mush' is no advantage unless you have the power or pitch authority to then accelerate back into normal flight. Indeed, if the aircraft had rolled over to 90 degrees through instability, possibly aided by some agressive rudder input then may that have allowed a recovery?

[EMIT]

OUCH, previous poster.

Agressive rudder input in a transport category airplane? No thank you, has been tried before (New York, nov 2001), wasn't an overwhelming succes.

Why not just push the stick forward, instead of full aft, would most probably work like a charm. Forward trim might be used as well, to make life easier.

Unreliable airspeed, alright, but there never was any unreliability in the ATTITUDE: four independant sources available (IRS 1, 2, 3 and ISIS attitude).

When ever, in a big lumbering airliner, is 15 degrees pitch up an appropriate ATTITUDE? Only after lift-off, with take-off thrust set on all engines. NEVER around FL350.

[DozyWannabe]

@CONF - I have never "made stuff up" on this subject in my life, and I always make a point of checking what I'm saying beforehand - with others if possible, some of whom are pilots. My knowledge is strictly engineering and systems design.

So I oversimplified and rendered the information badly - that said, I've still yet to see any proof that an "inadvertent" pull back on the stick would command a nose-up so aggressive that 13 degrees of THS up-trim is required for full elevator authority, or proof that neutral stick would do the same - Chris's theory notwithstanding (for which there seems to be some debate about which speed neutral stick switches from "G-load" command to "Zero Pitch"). I think the only way we'll know for sure is when the BEA release the interim report. Only problem there is that there are those on here, yourself included, that will never believe anything the BEA says regarding an Airbus aircraft - so I fear that these arguments will go on and on and on...

[Dont Hang Up]

EMIT

My post / question was based on the theory (admittedly unproven) that there was insufficent pitch authority to escape from this situation.

I just don't buy the idea that they were scared away from their one attempt at corrective pitch action by the restart of the stall warning and consequently maintained pitch up inputs for the remaining two minutes plus. The BEA report leaves large gaps but surely they must have tried everything?

[wmelvin]

Is there a determination of how the elevator trim came to be so high, nose up? The trim has more authority than the elevator and I doubt it is possible to recover from a deep stall with the trim full, or nearly full, in the nose up condition. My reading of the Airbus control laws is that automatic pitch trim is "available" in the Alternate Law mode. Did the trim go to the nose up position from pilot input or is there some envelope protection or other reason for it?

In the jet upset cases, it is hypothesized, that an updraft or headwind increase from penetrating an upper frontal boundary caused the aircraft to pitch up; the pilot pushed over but also trimmed nose down then the aircraft developed a steep descent with the pilot pulling hard to resist; the elevator force was sufficient to jam the elevator trim jackscrew so nose up trim could not be accomplished; the aircraft developed a steep dive which could not be corrected. Is it possible in this case, the pilots were attempting to cause a nose down condition and the elevator force was sufficient to jam the elevator trim?

There is a dynamic stall condition, discussed by S.S. Horner in his book on Fluid Dynamics where the AOA at which airflow re-attaches to the upper surface of the wing is significantly less than the AOA at initial stall.

[GarageYears]

wmelvin and several recent posters: Please read the thread, particularly Part 3 (found here> http://www.pprune.org/tech-log/452836-af447-thread-no-3-a.html). I believe every aspect of your question(s) have been examined, discussed, re-examined and turned inside out through the course of those posts. Re-hashing those same discussions will not further our understanding unless new information comes along with such posts. I don't mean to be dismissive, but many here have been with this thread since the initiating event and I am sure you will find an enormous amount of info (along with a lot of dross, it is true) in the tens or hundreds of pages that make up the body of work here.

[DozyWannabe]

@wmelvin

Time-consuming as it may be, I recommend you read the earlier thread(s) - a lot of your questions will be if not answered, at least debated there.

[RetiredF4]

HeavyMetallist

@RetiredF4: As I mentioned previously, an aerofoil doesn't cease to be an aerofoil and stop producing lift just because it's at high AoA and stalled. Yes, it produces a lot of drag (component of force parallel to the airflow) as your barndoor would, but there is still significant lift (component of force normal to the airflow). Since you don't want to just take my word for it (fair enough), here's a link to a study specifically of post-stall aerofoil behaviour; it has a good set of Cl/Cd/Cm-alpha graphs going up to extremely high AoA. You'll notice that the motivation for this study wasn't aircraft behaviour, but wind turbines; there's been a lot of interest in post-stall aerofoil chracteristics because they spend quite a bit of their time operating in a stalled condition!

Good reference, with witch i was familiar already. But i´m unable to see how you would prove your point of enough THS authority being available at 60° AOA.

Lets start on page 153 of your reference (http://docs.google.com/viewer?a=v&q=cache:fbMtXUlrOzcJ:wind.nrel.gov/designcodes/papers/OstowariNaik_PostStall44XX_Part2.pdf+airfoil+pitching+moment +post+stall&hl=en&gl=uk&pid=bl&srcid=ADGEESg0Do9z3A-8JK2pE-VmQEQEHAcJ5CIyuig9l9yd8NewuhIdC78nTXOcXmlp_jPBL-MoZwO9cI1qnmJT1dvR8At-4Bf6J436YITa7-gzu4fEQHBs_hIkeMLs35uUPfOGLCde16fm&sig=AHIEtbRu8yOF6KFBt3RzZndfZtB5fnZeyQ) with the lower graph, which shows how drag is a function of AOA. The more AOA, the higher the drag. To reduce drag, you have to reduce the AOA.

Now lets look at the upper table, Coefficient of lift versus AOA. Lift increases with AOA, until stall AOA is reached. In that graph that is at about 20°AOA. This drop in Lift continues till 30° AOA. From a max CL of 1.4 at 20°AOA CL has dropped to CL 0.9 at 30° AOA. If we continue to increase the AOA to 45°AOA, CL would increase to 1.2 again. But remember and check the lower graph, we just have increased the CD by the factor of six from stallspeed until 45° AOA. ! If we increase the AOA above 45° to 60° AOA, CL decreases again and CD continues to increase.

What does it tell for the A330?
First of all i dont know what kind of profile the THS represents, but your reference shows, that most mentioned airfoils have a similar behavior, just the numbers change. So lets use what we got until someone comes up with the correct profile.

Lets start with the 46.5°AOA of the THS (60° Aircraft AOA minus 13.5° NU trim) , which gives us a CL of 1.2 and a CD of 1.2. Lets do some ND input by trimming the THS to 0° and let´s see what it does to our numbers. We end up at 60° AOA at the THS, a CL of 1.0 and a CD of 1.6! Our upforce on the THS needed for lowering the nose has decreased and the drag has increased, our ND trim has caused the opposite of our desired outcome.

But that´s not all, lets start again at 47.5° AOA with CL1.2 and CD 1.2 and let us pull full aft stick. That decreases the AOA of the THS for example to 30°AOA, but at the same time we are loosing CL to 0.9, (loosing upforce on the THS again) but we improve CD to 0.5.

Conclusion in this simulated case with this airfoil: Somehow you are stuck between max CL pre-stall and max CL post stall and the airfoil itself has no authority to change that.

To make matters worse, the same applies to the other airfoil, the wings. There´s probably a similar change in CL and CD and any positive change on the THS (somehow generating ND force for the aircraft) might bring the AOA of the wing into the region, where the decrease of AOA initially causes an increase of CL thus rendering your success on the tail usesless for the complete fuselage.

As mentioned before, i´m no aerodynamics guy and might be completely wrong, but i like to make the point that the crew most probably had no conventional way to get AF447 out of a 60°AOA stall and that their behavior described by the BEA in the last 3 minutes of the fall might reflect the helpless situation they found themselves in due to normal aerodynamic layout and behavior of the aircraft, which has never been tested in that flight region.

Another point to remember: That THS has never been designed and tested to generate negative g or great nose down pitching momentum (like in a fighter aircraft), but to keep the nose up in cruise and especially during approach to land.

Another point: CG might have travelled aft (some fuel in the tanks not equally distributed, but now in the aft parts of the tanks, Some cargo in the belly falling to the rear part of the hold, and shift of the center of lift due to high AOA, a mixture which ruins your day.

[Chris Scott]

Quote (http://www.pprune.org/tech-log/452836-af447-thread-no-3-a-103.html#post6515113) from HazelNuts39 (Jun15/1433z):
So I asked myself: how does one recover from a stall without reliable airspeed and without AoA? I would think that the proper action would be a determined nose-down push, maintained until the stall warning stops, and then trying to maintain an AoA on the edge of s/w, gently allowing the nose to raise until s/w occurs, then a small nose-down correction to silence it, etc., until back in approximately level flight at a reasonable pitch attitude.
EDIT:: Can anyone be expected to do that succesfully without being trained for that eventuality?

Quote (http://www.pprune.org/tech-log/454653-af-447-thread-no-4-a-2.html#post6516611) from PA_18_151 (Jun16/0733z):
Scary stuff,. You don't use the stall warning to calibrate your response, you use your remaining instruments and your knowledge of your aircrafts pitch/power settings. Your solution puts you at an unnacceptable risk of secondary stall and that is another opportunity to enter a spin. If you do not know your aircraft's pitch/power settings for level flight then you are not competent to pilot it/manage its autopilot and take over if/when that autopilot fails.

Think you are missing HN39's point.

This aeroplane should not have stalled. It should not have zoom-climbed 3000ft, trading kinetic energy for potential energy ("speed for height"). But, for whatever reason, it did. Your advice – like that buried in a badly-written QRH procedure, and discussed on these threads for over two years – is spot-on prior to the climb from FL350, and might even have sufficed passing FL375 (AoA +4). Very soon after that, however, the wing stalled, and the AoA started rising rapidly. From then, the PF (perhaps unknowingly, although he did select TOGA thrust) was faced with a very different problem, which is the one that HN39 is addressing. (By the way, TOGA is little more than CLB thrust at high altitude.)

So the aeroplane is already in the stall (becoming a deep one). FPA display has been lost, and with it any chance of obtaining a rough idea of AoA. The VSI is soon off-scale (down), and it's night IMC. I assume you agree with HN39 that "the proper action would be a determined nose-down push", but that this should be sustained until well after the stall warning stops. Yes, there's a great risk of a secondary stall during the flare. But, with no reliable source of IAS, do you risk over-speeding the aircraft or and/or hitting the sea? If you end up very fast, how much height would be lost from an FPA of, say, -30deg**, at the 2.5G limit? With the altimeter racing downwards in a way you've never experienced except during a simulator "reset", at what point do you start pulling?

** (assuming stall recovery was started at, or soon after, the apogee)

[syseng68k]

Retired F4, 35

As mentioned before, i´m no aerodynamics guy and might be completely
wrong, but i like to make the point that the crew most probably had no
conventional way to get AF447 out of a 60°AOA stall and that their
behavior described by the BEA in the last 3 minutes of the fall might
reflect the helpless situation they found themselves in due to normal
aerodynamic layout and behavior of the aircraft, which has never been
tested in that flight region.
This is way I see it as well and mentioned in a previous post. If the
a/c has never been tested in such a situation, how can the crew be
expected to recover ?.

Testing the water here, but is that an argument for a tail chute to be
fitted to civil transports, or is that too far off the wall and would
it add to problems in that it might cause structural damage to the
airframe ?.

Another thing that's been nagging me about the last bea report is that
when the ap and ath disengaged, there was a "roll to the right". This
strikes me as odd as one would expect that at the point of disengagement,
they should hand back without change of ap or ath settings. Thus,
no immediate change. Why the uncommanded roll to the right, systems
fault, or what ?...

[RR_NDB]

Hi,

Testing the water here, but is that an argument for a tail chute to be
fitted to civil transports, or is that too far off the wall?
Considering the fact "prevention" can fail what kind of resource could be used to "remedy"?

Unfortunately the "bean counters", ROI considerations are against us, technicians.

and would it add to problems in that it might cause structural damage to the airframe ?.

It would be better "written off" a/c´s than other similar tragedies.

If even for "sub heated Pitot´s" we have long delays for "fixing" one could easily imagine what will happen for airliners "stall recovery resource(s)". Probably no chances and probably there are better approaches to the issue. Like the ASAP introduction of "real redundant" air speed sensors. So, Airbus SAS seems to be working on that important issue, vital (AF447 is showing this) for their "advanced planes" reliable operation.

Probably the simplest fix would be just "hotter Pitot´s. And this can be "tested" by the "players". With benefits to A, B and other companies (and a/c operators, etc.)

[Chris Scott]

syseng68k,

Presumably a tail-chute could be fitted, perhaps to a strengthened APU bulkhead. How its getting tangled-up with the APU (which might be running) could be prevented, I'm not sure.
Other problems
(1) It would have to be certificated for public transport, and presumably deployed on a test-flight.
(2) Would it be deployed automatically or manually (by crew action), and what criteria would be used?
(3) Would it subsequently be jettisoned automatically or manually (by crew action), and what criteria would be used?
(4) In the event of jettison failure, which might have followed unwanted deployment, would the aircraft be capable of maintaining level flight? (The second prototype BAC 1-11 had to make a very sudden forced landing on Salisbury Plain when it suffered this problem during deep-stall testing,)
Hard cases make bad law?

Re the uncommanded roll to the right as the AP disconnected, the BEA does not quantify it. Turbulence?

Like some others, I have a feeling that the PF may have experienced some difficulty in achieving left-stick without up-stick when unexpectedly handed the baby at 4am CET. I also hang on to the faint hope that the BEA may acquire ASI 2 data from the QAR. Unless it was over-reading, I cannot understand why he perpetuated the mistake as the aircraft climbed.

[jcjeant]

Hi,

Like some others, I have a feeling that the PF may have experienced some difficulty in achieving left-stick without up-stick when unexpectedly handed the baby at 4am CET. I also hang on to the faint hope that the BEA may acquire ASI 2 data from the QAR. Unless it was over-reading, I cannot understand why he perpetuated the mistake as the aircraft climbed. The BEA note:

From 2 h 10 min 05 , the autopilot then auto-thrust disengaged and the PF said "I have the
controls". The airplane began to roll to the right and the PF made a left nose-up input. The stall
warning sounded twice in a row. The recorded parameters show a sharp fall from about 275 kt
to 60 kt in the speed displayed on the left primary flight display (PFD), then a few moments
later in the speed displayed on the integrated standby instrument system (ISIS).
Note 1: Only the speeds displayed on the left PFD and the ISIS are recorded on the FDR; the speed
displayed on the right side is not recorded.
Note 2: Autopilot and auto-thrust remained disengaged for the rest of the flight.
At 2 h 10 min 16, the PNF said "so, we’ve lost the speeds" then "alternate law […]".
Note 1: The angle of attack is the angle between the airflow and longitudinal axis of the airplane.
This information is not presented to pilots.
Note 2 : In alternate or direct law, the angle-of-attack protections are no longer available but a
stall warning is triggered when the greatest of the valid angle-of-attack values exceeds a certain
threshold.
The airplane’s angle of attack increased progressively beyond 10 degrees and the plane started
to climb. The PF made nose-down control inputs and alternately left and right roll inputs. The
vertical speed, which had reached 7,000 ft/min, dropped to 700 ft/min and the roll varied I don't know (or on what they base their feeling) why many people think the pilot perpetuated command to climb ... as the BEA note is clear about ...

The airplane’s angle of attack increased progressively beyond 10 degrees and the plane started
to climb. The PF made nose-down control inputs and alternately left and right roll inputs
BTW .. exactly like the Perpignan pilote made.

[GarageYears]

Tail chute:

Probably a very bad idea. Aside from the complexity of certifying such a thing (presumably a different design/size would be needed for each aircraft type), the risk of inadvertent deployment causing it's own disaster would seem much more likely than the incredibly rare situation where it might be of positive benefit.

As far as I can tell from what we know to date, the crew were not aware they were in a stall (you'd have assumed someone would have used the words "we're in a stall" and that the BEA would have reported such in the note) and as such it seems unlikely they would have "punched the button" to deploy such a thing.

[FE Hoppy]

A sustained nose down input would have done the trick. No need to reinvent the wheel.

[DozyWannabe]

The airplane’s angle of attack increased progressively beyond 10 degrees and the plane started
to climb. The PF made nose-down control inputs and alternately left and right roll inputs

The note also says:

After the autopilot disengagement:

the airplane climbed to 38,000 ft,
the stall warning was triggered and the airplane stalled,
the inputs made by the PF were mainly nose-up,
the descent lasted 3 min 30, during which the airplane remained stalled. The angle of attack increased and remained above 35 degrees,
the engines were operating and always responded to crew commands.

which *implies* (but does not explicitly state) that the nose-down inputs were outweighed by the size and/or frequency of the nose-up inputs made. The implication to pilots who find themselves in that unenviable situation would therefore be to hold nose-down input long enough for the autotrim to neutralise the nose-up setting that it had previously commanded or to use manual pitch trim if you want to be certain of where the THS is positioned by directly controlliing it mechanically.

[syseng68k]

RR NDB, 38

It would be better "written off" a/c´s than other similar tragedies.

It it saved one life, it would be worth it, imho, though you would never get
the beancounters or actuarial types to agree, who just factor in the occasional
hull loss into the risk assessment. Cruel old world, but that's business.

Chris Scott, 39

Presumably a tail-chute could be fitted, perhaps to a strengthened APU bulkhead.
How its getting tangled-up with the APU (which might be running) could be
prevented, I'm not sure.

Other problems
(1) It would have to be certificated for public transport, and presumably deployed on a test-flight.
(2) Would it be deployed automatically or manually (by crew action), and what criteria would be used?
(3) Would it subsequently be jettisoned automatically or manually (by crew action), and what criteria would be used?
(4) In the event of jettison failure, which might have followed unwanted deployment, would the aircraft be capable of maintaining level flight? (The second prototype BAC 1-11 had to make a very sudden forced landing on Salisbury Plain when it suffered this problem during deep-stall testing,)

Hard cases make bad law?
There would be difficulties, but apart from the politics, the rest is just an
engineering problem and it is proven technology elsewhere. Not to make light
of it all, but if there was the will, the job would get done. Hard cases can
make bad law, but that's not an excuse for doing nothing.

A serious problem for airbus and others is that the fitting of a tail
chute or any radical stall recovery solution would be an indirect admission
that their planes can stall, when they have spent so many years carefully
crafting and disseminating the myth that they cannot. They don't even train
for it, it can't happen, so don't worry. Seems to me that they have been
drinking far too much of their own koolaid, but perhaps i'm completely wide
of the mark with that view. Just don't mention the war, right ?.

If you consider that all the fbw technology filtered down initially from
military requirements, where it was tested to the limit in every way, you
could argue that civil aviation currently has only half the deal. Fbw
technology allowing relaxed stability requirements as per mil, but in
reality no way to get it all back if the envelope is pushed just that little
bit too far.

Someone posted a link to the 320-232, D-AXLA video earlier and watching it,
was surprised by the speed and quantity of input from the pilot on the stick.
With a weight of (?) 200 tons, it would take seconds to get any significant
response, or am I way off beam ?...

[GarageYears]

jcjeant:

I don't know (or on what they base their feeling) why many people think the pilot perpetuated command to climb ... as the BEA note is clear about ...

The airplane’s angle of attack increased progressively beyond 10 degrees and the plane started
to climb. The PF made nose-down control inputs and alternately left and right roll inputs
BTW .. exactly like the Perpignan pilote made.

What does this mean then from the very first paragraph at the START of the incident:

The airplane began to roll to the right and the PF made a left nose-up input

What do you imagine this did if it didn't precipitate a climb?

Perhaps I'm way off here, but clearly this ISN'T the first time an AP and AT has dropped out while in cruise at FLxxx (please, correct me if I'm wrong), but given there are no other reports of those aircraft not making it to their intended destinations, my assumption is this is not considered a particularly problematical situation in the normal run of things.

Agreed this situation was precipitated by a loss of pitot data, resulting in a loss of speed inputs, but the point is at moment (x) the airplane is cruising fairly normally under full auto control and some short time later (y) things have gone rather wrong under PF command. The BEA have provided what they consider is key relevant data/information in the form of the published note, with the full understand that this document was likely to be hyper-analyzed by some small segment of the aviation community in particular, knowing that eventually the full details will be released. So it seems very implausible that the information provided is not the fundamental framework of the whole event.

AP/AT off, roll to the right, PF left NU input, climb....

I think trying to paint a mysterious picture on top of this is erroneous and misleading. My presumption from what you have written, your are implying the climb was NOT of the pilots command? I can see nothing to support that assertion.

[alph2z]

Re.: Parachute problems, flight test challenges, and AOA sensors problems (Emphasis mine):

...According to statements from the surviving co-pilot and flight test engineer, the flight crew was troubleshooting a noise associated with stalls conducted during previous flight test activities. A non-scheduled stall was conducted after the scheduled testing and the plane's angle of attack, the difference between the plane's flight path and its wing's chord line, increased past the maximum allowance of 34 degrees.

All is Lost...

The flight crew lost control of the Challenger. Following the recovery procedures set in place, they were unable to regain control, going so far as to, in a last ditch effort, deploy the plane's emergency spin recovery parachute. Control was temporarily regained, but now the deployed chute fouled the flight characteristics. Unable to release the chute, with with the plane's starboard engine failing, the flight crew was forced to bail out of the doomed aircraft.

The plane impacted the desert floor near the village of Cantil at 9:10 in the morning. Unfortunately, the pilot, Eric Norman Ronaasen, was killed when his chute failed to deploy, and the copilot, Dave Gollings, received minor injuries. The flight test engineer, Bill Scott, was not injured

Months after the crash, a Canadair flight engineer examined computer information that had been available before the crash and discovered the banging was caused by an engine problem.

The NTSB, after its investigation, concluded the probable cause of the accident was a problem with the angle-of-attack indicator binding, due to a failure in the seals of the hydraulic system. It also concluded that four separate systems failures had caused the crash - three of which had plagued the twin-engined jet in earlier tests.....

Challenger #1001 (http://www.check-six.com/Crash_Sites/Canadair-CGCGRX.htm)
.

[bearfoil]

Garage Years

"The airplane began to roll to the right and the PF made a left nose-up input"

How much climb? Was he reacting to FD or VSI? BEA themselves say that the a/c didn't "start to climb" (Why do you assume it "started to climb" from level?)until the NU was passing through ten degrees NU. The pilot made one NU (one) input. Was he (attempting) correcting for a natural ND tendency (or, stopping a descent?) that the a/p had been solving, but at drop, the a/p input was stopped? (Yes?, since the a/p stops at quit; there are no more "trend" solutions). Since he then is described as inputting several ND inputs, was he attempting to control an opposite, and again, (perhaps) "natural tendency" for NU? Under what circumstances can the PF be overridden by Trim? He had attitude (supposedly) so what was the disagree between PF and THS? (If any). If at any time PF was aware of an NU of ten+ degrees, he tries to climb exactly Why? BEA have left the barn door open, and the ensuing nonsense on either side allows Airbus to relax at PAS?

It is just as easy for me to believe the THS went into banana mode as it is for PF (F/O1) to be thought to be wired backwards.


The airplane’s angle of attack increased progressively beyond 10 degrees and the plane started to climb. The PF made nose-down control inputs and alternately left and right roll inputs..Trying to see where in the report the PF can be faulted for a THS at 13.2 degrees.

[FA10]

...And you certainly do not continue pulling...


Lots of theories here about the chances to recover from a deep stall.

However it took quite some time to get the Bus INTO that condition, with obviously a lot of confusion going on (climb 3000ft without noticing?).

If the PF on a machine with conventional control column starts to pull when the A/C is already in a stall condition, the second set of eyes on the flight deck will most likely immediately realise the mistake, he will shout, take control, assist,... at least do something useful.

When the PF of an AB type fiddles around with his sidestick, the PNF obviously has only very limited information about what his colleague is doing.

Especially in a high-tension situation, where you find yourself in the middle of a CB, a moving column in front of myself pleases me much more than a frozen sidestick, that does not reflect the inputs of the PF.

I suppose that endless believe in the ability of the flying computer to cope with all posible scenarios and elliminating one of the principals: Keeping the crew in the loop proved to be one of the main killers...

I guess before Airbus there were not many stick-piloted aircraft with 2 crew (from PZL´s Blanik till Rockwell´s famous glider) that did not give the second crew member a feedback on his stick from the inputs of the other pilot!

[DozyWannabe]

I guess before Airbus there were not many stick-piloted aircraft with 2 crew (from PZL´s Blanik till Rockwell´s famous glider) that did not give the second crew member a feedback on his stick from the inputs of the other pilot!

Lordy, lordy... If the PNF was unhappy with the developing situation - indeed if any PNF is unhappy, on a FBW Airbus all they have to do is state "I have control" and push the button on their own sidestick, which will immediately transfer full control to the other side of the flight deck.

In an aircraft that has manual reversion, there are many circumstances when being able to monitor the other pilot's inputs (and the aircraft's responses) is helpful. In a fully-hydraulic aircraft those circumstances are fewer. While the Airbus philosophy restricts more information to the visual channel, the presence of 3 ADIs should be more than enough to determine that things are not going well in terms of the aircraft's attitude.

[infrequentflyer789]

Hi,

The BEA note:

I don't know (or on what they base their feeling) why many people think the pilot perpetuated command to climb ... as the BEA note is clear about ...

The airplane’s angle of attack increased progressively beyond 10 degrees and the plane started
to climb. The PF made nose-down control inputs and alternately left and right roll inputs
BTW .. exactly like the Perpignan pilote made.

You need to read the english more carefully (and that might not be easy for non-natvie speakers, not sure I could read the French exactly right).
BEA haven't helped by releasing this as a narrative rather than a simple time-line - but then it isn't an investigative report, it's a press release designed to shut up the distressing (for some, I'm sure) media speculation. To be fair, it has largely succeeded in that.

Leaving out the other stuff, the sequence of pilot input and plane response reads like this:


PF left nose-up input
ptich increases, plane starts to climb
vertical speed reaches 7000fpm
PF makes nose-down inputs
vertical speed reduces

The 7000fpm figure appears after the nose-down input in the text, but is qualified with "which had reached", meaning that figure being already reached at the time of the nose-down input.

This timeline is why "people think the pilot perpetuated command to climb" - because there is no indication of anything else, and when the pilot input went nose down the plane responded.

[Chris Scott]

Hi jcjeant,

The BEA Update is written in plain language, and time-markers are very limited. There is much room for interpretation (that is, ambiguity). That is why many of our best contributors here are either silent, awaiting the full data, or contributing but rarely.

You have chosen to interpret it in a way which supports your pre-conception that the upset was likely to have been caused by a fault in the Airbus FBW system. Good luck to you. As you would expect of an ex Airbus skipper, I look at it from a different perspective: was it weather, system fault, or pilot mishandling in difficult circumstances; or some combination of these factors?

Remember, the BEA states that, after the AP disengaged,
"the airplane began to roll to the right and the pilot made a left nose-up input",
but it doesn't say how long the nose-up input lasted.

I think the clue lies in the next paragraph, dealing with the initial climb, of which you have only quoted the first part. Here is all of it:

"The airplane's pitch attitude increased progressively beyond 10 degrees and the plane started to climb. The PF made nose-down control inputs and alternately left and right roll inputs. The vertical speed, which had reached 7000ft/min, dropped to 700ft/min and the roll varied between 12 degrees right and 10 degrees left. The speed displayed on the left side increased sharply to 215 kt (Mach 0.68). The airplane was then at an altitude of about 37,500 ft and the recorded angle of attack was around 4 degrees."
[my highlighting]

So, a likely interpretation is that the PF's nose-up input initiated the climb by increasing pitch to above +10deg. (You may disagree.) That is an extraordinary attitude at high altitude. He may have ceased nose-up inputs at that point, but the new pitch attitude produced a climb VS of 7000ft/min. At that rate, it would only take about 15 seconds to get to the next point, where the PF made his short-term nose-down input, causing the reduction in VS to only 700ft/min with an AoA of +4, which was just below the stall AoA.

At this stage, the situation was already very serious and very tricky to stabilise, and it is easy to criticise with armchair hindsight. However, had he maintained some forward stick for a bit longer, a descent would have started and the aeroplane would not have stalled. His altimeter should have been not more than a few hundred feet in error, so he would have known he was too high. A gentle descent with the nose somewhere near the horizon would have been a step in the right direction. The pitch-up tendency caused by the selection of TOGA thrust would have been countered by the FBW system applying down-elevator, backed up by the THS.

Unfortunately, according to the BEA, the PF's next pitch commands were nose-up. The reason for that is unclear, but that is what led to the stall.

[FA10]

Trust me, DozyWannabe, I know how to take controls - have to do it occasionally.


...the presence of 3 ADIs should be more than enough to determine that things are not going well...


...maybe there is a generic problem in that assumpion?

[infrequentflyer789]

A serious problem for airbus and others is that the fitting of a tail
chute or any radical stall recovery solution would be an indirect admission
that their planes can stall, when they have spent so many years carefully
crafting and disseminating the myth that they cannot.

As opposed to the direct admission of designing and building in a full stall warning system ? Did they do that just for fun on their day off, being as it wasn't in-spec ?

Those that designed the a/c clearly knew it could stall. Whether that message got messed up between there and the people that train the pilots that fly it, is another more interesting question.

[Chris Scott]

Quote from syseng68k:
Someone posted a link to the 320-232, D-AXLA video earlier and watching it, was surprised by the speed and quantity of input from the pilot on the stick. With a weight of (?) 200 tons, it would take seconds to get any significant response, or am I way off beam ?...

No, you are right. In case you haven't seen it before, here is my opinion:

http://www.pprune.org/tech-log/316096-lh-a320-rough-landing-hamburg-24.html#post3979423

[DozyWannabe]

Trust me, DozyWannabe, I know how to take controls - have to do it occasionally.

I wasn't implying you didn't, I was just throwing that information out there, apropos of nothing...

...maybe there is a generic problem in that assumpion?

I'd say that the very respectable safety record of the FBW Airbus series suggests otherwise, but that's just my opinion.

[bearfoil]

Chris Scott

Ten degrees nose UP? That tells us nothing about the AoA, or the position of the THS. Was it already at 13.2? Did pilot cease his initial NU and begin his repeated ND inputs? The 7000fpm suggests the climb was (initially) very rapid, and there is no reason to think the THS was not "UP" having corrected for a chronic descent? Turbulence? There is nothing in this sequence (sic) to base a conclusion of PF chronic NU. BEA say ".......A INPUT....."

How outlandish was the descent that THS needed ten degrees to arrest it, and start to climb? Or more to the point, what was the THS' deflection at a/p drop? BEA don't say. A/P will hang in until 13 degrees Pitch UP before it quits. If the descent was transient, and chronic, why shouldn't the THS react with 13 degrees nose up? FOR TRIM.... Initially, PF may have seen the nose dropping as the problem needing nose up. When he realised his TRIM state, he started his repeated ND?

you say,

So, a likely interpretation is that the PF's nose-up input initiated the climb by increasing pitch to above +10deg. (You may disagree.) That is an extraordinary attitude at high altitude. He may have ceased nose-up inputs at that point, but the new pitch attitude produced a climb VS of 7000ft/min. At that rate, it would only take about 15 seconds to get to the next point, where the PF made his short-term nose-down input, causing the reduction in VS to only 700ft/min with an AoA of +4, which was just below the stall AoA.
What About.....

IOW. If the THS was at 13.2 degrees at a/p handover, and the a/c was sinking, wouldn't the PF input NU to stop descent? When he realizes his THS (can he?), he stops, but the a/c "has started to climb" (Bow Howdy, 7000fpm?) His elevator pushed the a/c past what it needed? Once climbing, furtrher ND input stalled the TAIL AS A UNIT. It could not bite to drop the nose, It was committed to Ballistic trajectory to actual aerodynamic Stall, that dropped the Nose but no energy available to keep the nose dropping so it settled back on its tail., but PF input ND, which restalled the TAIL. Now PF is in a corner. He CANNOT recover, he is bracketed between no energy and a Tail that cannot unstall. The tail is a huge Spoiler, now, and it is stuck deployed full. When the a/c goes NU , it is limited by the drag of the tail from going inverted, but since the Tail cannot unstall the a/c cannot drop. This I believe was the result of what is seen as the initial "recovery" that initiated the STALLSTALL. The AoA was never going to be less than 4 degrees, insufficient to Break the STALL.

This means that the THS may have been responsible for the A/P quitting. 13.2 degrees exceeds its Limit for TRIMMING. Does Airbus want such a huge lifting surface as a TRIM DEVICE? Pending further data, it could well be.

Why did the Pilots not MAN TRIM ND? First of all, they were trained not to, and secondly, How were they to know the THS was at 13.2 degrees NU or at the limit. Actually, BEA have not said precisiely. With elevators, they had horrible nose up, or "almost" a recovery. The Rolling on descent may have been there solution to gain STALL recovery.

There is just such a disconnect between a/c and pilot, is it impossible to entertain "holding back stick all the way down"? Every time they Push, the a/c goes berserk, and refuses to unSTALL anyway. Holding back stick gives a stable, "quiet" ride. Deadly, but quiet.


The LEFT ROLL command? Let's say the Autopilot had been trimming for NU and Right ROLL. Just before a/p quits, the trend intensifies, perhaps with gusto. Serious ND trend and right ROLL. These aspects can be as much as 9 degrees ND and 45 degrees ROLL. The a/p quits with max NU and max left ROLL. PF is not ready for this amount of Trim; his initial (and likely ham handed, as discussed) inputs ask for maximum Left Roll and Maximum NU.

Hell of a way to start flying. In an astonishing climb, where ND inputs stall the Tail, and the trajectory is ballistic. Ballistic because no aerodynamic controls were available to descend, and the climb will only halt when the energy is too small to fly, only ballistic flight,

I am not sure the PF was not trying to keep the nose dropping when the AoA was 4 degrees, but the Plane couldn't get over the last little bump.

[MurphyWasRight]

FA10 Trust me, DozyWannabe, I know how to take controls - have to do it occasionally.



Quote:
...the presence of 3 ADIs should be more than enough to determine that things are not going well...


...maybe there is a generic problem in that assumpion?

Totally agree, while 3 ADIs might give a strong "hint" that things are not going well to the PNF they do not provide all of the data needed to understand the situation.

Specifically while it might be very obviouse that something was wrong the PNF would have no way of knowing that the PF might be making inputs contrary to what the situation seemed to warrant.

This could well delay the PNFs response since he might hesitate to ask a seemingly insulting question such as "we appear to be stalled, which way are you pointing the nose?" and instead assume the PF was dealing with a more arcane situation.

With direct knowledge of the PF inputs the PNF would be better able to recognize human as well as system failures.

BTW: I -do not- think the cause is anywhere close to pure pilot error, the above is meant as an example of need for awareness of -all- inputs affecting the aircraft.

---

The statement that "XXX cases of UAS at altitude did not have bad outcomes" is depressingly reminscent of NASAs attitude towards O ring erosion on the booster rockets.

In both cases a primary safety device was comprimised but disaster was prevented (for a while) by what was supposed to be a backup.
(O rings on shuttle, pilots and pitch/power on airplanes).

[bearfoil]

THIOKOL refused to OK the Launch due Weather. "Light this Candle" (Alan Shephard). They launched, and Thiokol, though grim, dodged a bullet.

The Launch was manslaughter. (Challenger Hearings, Feynman).

[MurphyWasRight]

THIOKOL refused to OK the Launch due Weather. "Light this Candle" (Alan Shephard). They launched, and Thiokol, though grim, dodged a bullet.

The Launch was manslaughter. (Challenger Hearings, Feynman).

bear

Actually Thiokol -engineers- said no but were overuled by Thiokol -managers- and NASA in the infamous conference call.

The point though is that like the UAS the O ring erosion was a known issue that was swept under the rug with "nothing bad has happened" etc.

BTWW Same thing for tile damage from foam shedding from the external tank, that is the really tragic one since NASA reverted to their old behavior of justifying exceptions since they had not caused problems in the past.

[Mr Optimistic]

Until it is shown that the pilots knew they were in a stall, the limitations of the aircraft to generate the appropriate moments are irrelevant surely ? If they applied the appropriate inputs but the system, or some stability trap, frustrated them well fair enough but the evidence doesn't seem to currently point that way. First thing is to figure out what they knew and if they didn't get the correct appreciation, why not and how can it be avoided in future ?

And what was the late heading change about, can you do that in a stall ?

[Chris Scott]

Bear,

You can read the BEA Update as well as I.

There's no mention of any descent prior to the initial climb, and the THS seems to have been no more than 3degNU until 02:10:51 (at about FL375).

Don't think the rest of your hypothesis merits response, except the last paragraph. As far as that's concerned, I expressed a similar sentiment on May 27th.

[syseng68k]

Infrequentflyer, 789


As opposed to the direct admission of designing and building in a full stall
warning system ? Did they do that just for fun on their day off, being as it
wasn't in-spec ?
Iirc, that's a regulatory requirement, whereas stall recovery solutions are
not :ooh:

I'm not being flip here and from fwics, whenever there is a known problem,
airbus do all the right things, issue advisories or press for regulatory
involvement. Development is a continuous process and no product is ever perfect.
The concern is, as you suggest, a mismatch between the goals of the manufacturers,
who have a passion, like most engineers, to build the best product, pursuit
of excellence etc and the end users, who are driven more by efficiency and the
need to minimise cost. In such a situation, there must be plenty of scope for
conflicts of interest at various points in the curve, hence the disjoint.

As you say, an interesting question. No doubt taxing the minds of many people
at present...

[bearfoil]

"There's no mention of any descent prior to the initial climb, and the THS seems to have been no more than 3degNU until 02:10:51 (at about FL375)."

I did not mention one either. I said the a/p had been trimming a descent OUT, With THS NU. And prior to a/p disconnect. At disconnect, if pilot became aware of the descent, and it was increasing, because the THS was no longer tracking it, his NU would make perfect sense. In additon, if robust, and additive to THS (~10degrees), the a/c would begin a breathtaking climb. No mention is made of the THS position of 3 at any particular time. Only that at some point it was the start of NU. During autopilot? After? We don't know

If PF also input right ROLL he may have over input, and began chasing his initial left roll.


..Until it is shown that the pilots knew they were in a stall..
Why would they know? They had two chirps that quit, and then a climb. At the Top of Climb, the STALLSTALL re entered, and quit.

THERE WAS NO FURTHER STALL WARNING.

WHY? BECAUSE SHE DID NOT STALL.

There was AoA of 4 degrees, and low speed. Who said it had to be 60 knots or less?

THERE WAS NO BREAK. Intead she slid down and back, LEAVING THE NOSE TO INCREASE ITS NOSE UP.

There was no nose drop to announce a STALL, merely (sic) a mush that had no end.

What about the PF's continuous stick back on the way down? Was he convinced he could CAUSE the a/c to STALL, and thence recover?

NOSE DOWN by now merely STALLED HS further, increasing the MUSH.

Stability Trap, Mr. O?

[Mr Optimistic]

We will soon be at FL100. Where was the desperation ?

[bearfoil]

The French are not Italian. They may have waited to a lower.

[syseng68k]

GarageYears, 41

Tail chute:

Probably a very bad idea. Aside from the complexity of certifying such a thing
(presumably a different design/size would be needed for each aircraft type),
the risk of inadvertent deployment causing it's own disaster would seem much
more likely than the incredibly rare situation where it might be of positive benefit.
It's way beyond my knowledge level, but it might be possible to find a solution
that just provided enough tail lift, while not presenting a danger in normal flight,
perhaps a release after use switch or time delay. The inadvertent deployment
issue could be solved via a two key, two man input with an on time requirement
of several seconds. Probably never happen though, but seems to me that something
needs to be done.


As far as I can tell from what we know to date, the crew were not aware they
were in a stall (you'd have assumed someone would have used the words "we're
in a stall" and that the BEA would have reported such in the note) and as such
it seems unlikely they would have "punched the button" to deploy such a thing.
I don't think they did. The pf saw the alt unwinding towards zero, didn't
believe the air speed and kept pulling up to get out of what he thought was a
dive. The fact that UAS is a known problem coupled with the fact that the pf
didn't believe that the a/c could stall, all his training telling him this,
made him disbelieve what were probably quite accurate asi readings.

The question is: How does engineering solve this problem ?...

[Smilin_Ed]

There was no nose drop to announce a STALL, merely (sic) a mush that had no end. Bear, is that "mush" not a stall? :confused:

[bearfoil]

Until it is shown that the pilots knew they were in a stall. Mr.O

Bear, is that "mush" not a stall? ...Of course, an unannounced one. What does she do at 60 knots and 48 AoA?

How do the pilots know? BEA report one STALLSTALL, then PF NU, and no STALLSTALL? They were not properly introduced I'm saying. By now, without a NOSE DROP, THS has no authority, it is Stalled further if ND is input, and if effective, the Nose goes UP, NOT DOWN.

Was the stubborn NU on the way down an attempt to get her to BREAK?

Or was he getting most of his "Nose Down" from pulling back? IOW, full back dropped the nose (it would), and pushing caused the Nose to rise. (also true). The Horizontal Stabilizer in this extreme aspect would behave like a spoileron, or spoilerator, getting authority from drag only, not aero.


wadda ya think

[Mr Optimistic]

But didn't the speeds become valid again, implying the nose came down as commanded ? Then renewed stall warning and maybe an instictive undoing of the last input as it 'caused' the warning ? I know the US commuter aircraft accident had the pilot pulling back but he could have been confused with a super stall, a known possibility in that case. In anything but an experimental regime, would anyone pull back when in a known stall ?

[mm43]

Bearfoil;
From BEA update ...
The airplane began to roll to the right and the PF made a left nose-up input. The stall warning sounded twice in a row.That statement appears to be quite straight forward, and gives a clear impression that at the time of the AP/ATHR disconnect the aircraft was in level flight at FL350. The status of the airspeed data is then described, and absolutely no mention (other than the UAS reference) of other untoward environmental conditions is made. Likewise, if the THS actuator was proceeding on direction of the PRIMs to apply another +10° to the CRZ +3° trim, don't you think they would have mentioned it?

Don't you think it would be overly disingenuous of the BEA then to leave out of their update those obvious factors that you so keenly believe they have?

Of course my naive assumptions may be proved wrong, but until they are, I will continue to believe in the integrity of the investigation.

[Smilin_Ed]

Many aircraft, B-47, B-52, and F-4 come to mind, had no extensive problems with uncommanded deployment but they had specific reasons for having them installed. For the F-4, with the introduction of fatter tires than required for carrier ops, the reason diminished, but the chute was still there and I was glad to have it on a wet, mid-length runway one night.

Tactical aircraft, where their normal employment requires flight near the boundaries of control effectiveness, are fitted with spin-chutes for tests near those boundaries, but there is no justification for that kind of testing in transport category aircraft.

Spin-chutes didn't always work as desired. A colleague of mine took an F-4 out for stall investigations and got into a spin. He deployed the spin chute but it came off completely and did no good. They had to punch out.

As someone here, or in another thread, pointed out, you don't fly transport aircraft near the edges of the maneuverability envelope. Teaching the pilots how to recognize how far the airplane can be flown before they get into some kind of controlability problem should be enough. This seems to be an area that may not have been given enough emphasis in training, given the circumstances of AF447.

[bearfoil]

mm43

No. it doesn't. The opposite is true.

Bearfoil;
From BEA update ...
The airplane began to roll to the right and the PF made a left nose-up input. The stall warning sounded twice in a row.
That statement appears to be quite straight forward, and gives a clear impression that at the time of the AP/ATHR disconnect the aircraft was in level flight at FL350.It tells me that the a/c was descending, and rolling off to the right. The PF tells me that, BEA needn't repeat it. If the a/c was not in s/l, but was being kept trimmed by an a/p that released the controls, this means that the PF couldn't just "sit there", he needed to continue to trim. My assumption would be that if continuous, this turbulence may have ended up being "too much" for the a/p parameters, and its loss might be due to inability to maintain trim. This inability, whether from uas and commanded, or turbulence and self commanded might just be the wheels coming off re: UPSET.

[mm43]

Bear;

Strange that that propensity to develop the starboard roll was evident all the way to the deck. Was it not fuel imbalance that the A/P was keeping trimmed, or do you think that it had another aerodynamic cause? Rudder, elevator trim or something else?

[syseng68k]

Chris Scott, 54

No, you are right. In case you haven't seen it before, here is my opinion:

LH A320 Rough Landing @ Hamburg

Have limited flying experience, but remember the words of the instructor:
"Be gentle with the aircraft, small inputs have a big effect". Or a little
more crudely: Make love to the aircraft, don't rape it. Strange, but you
get the best out of a well sorted and tuned car using the same technique.

Easy to say sitting at home in an armchair, but who can say what the
reaction would be on a cold dark night, with no visual cues and lost faith
in the machine's integrity...

[infrequentflyer789]

Sometimes I wonder if we're all reading the same report...


Ten degrees nose UP? That tells us nothing about the AoA, or the position of the THS. Was it already at 13.2? Did pilot cease his initial NU and begin his repeated ND inputs? The 7000fpm suggests the climb was (initially) very rapid, and there is no reason to think the THS was not "UP" having corrected for a chronic descent?


I doubt THS could get to 13 nose up in normal law without triggering protections. Unless AoA was busted (as at Perpignan) - but there is no indication given of that.

Even if it did, how didi it then get back down to 3deg (during climb and nose-up inputs) in order to rise from 3 to 13deg later in the time line ?

BEA clearly state pitch increased progressively at start of climb, 7000fpm was the rate reached at the end of the first climb, at the point PF started nose-down inputs.


Turbulence? There is nothing in this sequence (sic) to base a conclusion of PF
chronic NU. BEA say ".......A INPUT....."


and indeed they don't specify a duration, however the next mention of inputs is nose down after the (first) climb.

During that first climb PF gets stall warning. At 37500ft he gets stall warning again and "maintains" NU input in response. If the response to the SW is consistent, then the conclusion is chronic PF NU.

If the descent was transient, and chronic, why shouldn't the THS react with 13 degrees nose up?

What descent ? You either mean transient or chronic, or, with respect, you don't know what you mean. It can't be both. If it was transient, there would be insufficient time for THS to move, if it was chronic (and significant rate), there would be a noticeable altitude loss before a/p drops out.


Why did the Pilots not MAN TRIM ND? First of all, they were trained not to, and secondly, How were they to know the THS was at 13.2 degrees NU or at the limit.


So they were trained not only not to touch the trim wheel, but also that if they concentrated hard enough, it would actually disappear ?

To me, some of the scariest comments on these threads have been on training practices. Was PFs only training for stall warning "pull up" (to minimize alt loss) ? That is all it would take.

[DozyWannabe]

The point though is that like the UAS the O ring erosion was a known issue that was swept under the rug with "nothing bad has happened" etc.

I wouldn't call a worldwide Service Bulletin mandating the replacement of all Thales AA pitot tubes on Airbus A330/340 aircraft a case of sweeping the issue under the rug, would you?

As far as the Space Shuttle Challenger incident goes, the reason the Thiokol managers overrode their engineers (under intense pressure from NASA, don't forget) was because they were competing for the renewal of their contract to supply the SRB assemblies, and they feared that exposing a weakness in their product to the customer during the competitive phase would work against them. A similar thing happened with the DC-10 cargo door debacle - Convair (the supplier of the fuselage assemblies, including the cargo doors) knew of the problems and had encountered them in testing, but upper management at Convair downplayed the risk to McDonnell-Douglas, fearing that their supply contract might be endangered.

In this case, Airbus, Air France and Thales all have some stake in them owned by the French State, so it's unlikely that such contractual shenanigans would be a problem. The one time that politics may have been played (Habsheim), it backfired on all parties in a bad way.

Despite the Challenger disaster, Morton Thiokol retained their supply contract to build STS SRB units, presumably in part because when one makes such a major mistake in public, efforts to avoid any such thing happening again will be redoubled. I'd be very surprised if the same isn't true of the BEA (as I said earlier, note their bringing in of the NTSB as an independent observer in the Air Inter accident). The Airbus FBW technology is proven and flies people around the world safely every day - there's nothing to hide anymore.

[GarageYears]

The question is: How does engineering solve this problem ?...

Build BETTER training devices that provide extended flight and upset envelope training. I have posted several time about the RAeS ICATEE working group. Real aerodynamicists can model beyond flight data realistically (plus support such s/w models with wind-tunnel data). Please take a look RAeS Flight Simulation Group - Introduction to the ICATEE (http://www.raes-fsg.org.uk/cgi-bin/yabb1f/YaBB.pl?num=1252658572) This group has been studying this issue for 2 years and can make a significant contribution to all extended envelope training situations.

[Hyperveloce]

To me, some of the scariest comments on these threads have been on training practices. Was PFs only training for stall warning "pull up" (to minimize alt loss) ? That is all it would take.

The training practices may need some new scenarii but none of the past Pitot incidents (with the same pattern) analyzed in the BEA interim report n°2 showed such a crew reaction (maintained NU inputs), some of them did not identify the "unreliable IAS" situation, some of them chose not to react to the stall alarms, some of them put their A/C in descent, etc... the persistent NU inputs of the PF are one thing, but there are two pilots (and 3 a bit later) in the cockpit: the PNF did not stop this critical high altitude sequence (pitch & AoA skyrocket)(the BEA says he has been trying several times to get the CPT back to the cockpit), and the CPT did not recover it (even if there was ND inputs which reduced the AoA).
The recent BEA note is so coarse, there are huge gaps in the chronology, we have been given a limited set of new facts, and it raises so many questions. Hopefully in one month, the BEA will have fully released its findings, and we will have all the time plots of all the flight parameters (and automation states), crew inputs, a more detailed review of the CVR content, of the procedures implemented (or not), maybe a numerical simulation of the stall,... many new facts, and maybe unanswered questions. Note that AF and Airbus have not been waiting for the BEA findings to alter the stall procedure or prohibit the emergency maneuver (memory items CLB/5°) in cruise phase.

[Zorin_75]

what was the THS' deflection at a/p drop? BEA don't say. A/P will hang in until 13 degrees Pitch UP before it quits.
According to BEA, the THS passed from 3 to 13 degrees nose-up in about 1 minute at 2:10:51, 46 seconds after a/p disconnect. Even applying the least plausible interpretation to the report, that it was already arriving at 13 deg at 2:10:51, THS would just have started moving when the a/p disengaged.

When there are such large holes in what we know at this point, I don't really understand why it is necessary to make up stuff that even contradicts the few facts we do have?

[bearfoil]

IF789

If the descent was transient, and chronic, why shouldn't the THS react with 13 degrees nose up?

Transient as in intermittent, and Chronic as in trend. Both. Pilots do this often, tweak, wait, tweak, wait ad ab. Another term could be: Hunting.

BEA report the THS moved from 3 - 13 degrees. Three degrees could mean sometime before the a/p is lost. This fits in with one excursion only, and could include both a/p and PF inputs, separated by the drop.

"...So they were trained not only not to touch the trim wheel, but also that if they concentrated hard enough, it would actually disappear ?

To me, some of the scariest comments on these threads have been on training practices. Was PFs only training for stall warning "pull up" (to minimize alt loss) ? That is all it would take..."

The pullup at STALL may mimic an artifact in training (not changed til after 447 went in) where the "Recovery" is w/o excessive altitude loss.Where was the added Thrust?

Turbulence? There is nothing in this sequence (sic) to base a conclusion of PF chronic NU. BEA say ".......A INPUT....." you say

and indeed they don't specify a duration, however the next mention of inputs is nose down after the (first) climb.

During that first climb PF gets stall warning. At 37500ft he gets stall warning again and "maintains" NU input in response. If the response to the SW is consistent, then the conclusion is chronic PF NU.

If the THS had some excess NU, and PF added more NU at handover, what is to say he was not attempting his trained recovery. At STALLSTALL he would NOT immediately input ND but per SOP he would maintain "pull" to minimize altitude loss, not knowing his meager ND, (relaxed back pressure) was way insufficient to keep from Stalling.

He may have actually input Nose down, but it was not beyond neutral and was "READ" as NU by the FMC? He may have been trying to fly with elevators "Only", (not knowing NU/ND would react completely different in response) that would be his set from Stall Training? He certainly didn't train approach to STALL with 13 degrees of THS NU.

At this point, would he become a little unnerved at the reaction to seemingly "appropriate inputs. He would be getting large (and variable and delayed) actions from his hybrid elevator, Tail Plane mongrel. Once again, was his continuous back pressure on the way down his stubborn insistence on his reaction to the a/c at this point (Pull Back, Nose drops, Push, Nose raise?)

If so, at what point was the "feel" of the a/c so foreign to him that he acted in "unpredictable" ways?

thanks for the response :ok:

[DozyWannabe]

Or more to the point, what was the THS' deflection at a/p drop? BEA don't say. A/P will hang in until 13 degrees Pitch UP before it quits.

You're getting things a little muddled, bear. A/P disconnect appears to be caused by the loss of air data, not the position of the THS.

Read the article that PJ2 linked to, it contains some very useful information - including the fact that the ADR disagree lasted for 45 seconds, which means that return to Normal Law is unlikely to have happened (Svarin should find that interesting).

[bearfoil]

Doze

You're getting things a little muddled, bear. A/P disconnect appears to be caused by the loss of air data, not the position of the THS.

No. A/P disconnected via unreliable air data OR it limited out. Not loss. It is the cause of the discrepancies that is important, and unknown. The a/c was being trimmed at handover. It was descending and rolling OUT of A/P. This suggests to me that the cause may have been severe turbulence AND/OR an autopilot that could not keep up.

Either way, these are two distinct possibilities. With Ice, we are stuck with only one. The timing of the faults suggests a/p dropped before the reads became unreliable. This means it is possible the a/c remained in Normal Law for a short term, until the discrepant reads instigated the Law change. I have not seen a proof against this. Only some "assurances".

[GarageYears]

Zorin_75:

At 2 h 10 min 51 , the stall warning was triggered again. The thrust levers were positioned in the TO/GA detent and the PF maintained nose-up inputs. The recorded angle of attack, of around 6 degrees at the triggering of the stall warning, continued to increase. The trimmable horizontal stabilizer (THS) passed from 3 to 13 degrees nose-up in about 1 minute and remained in the latter position until the end of the flight.

Full text of the paragraph from the BEA.

Your version:

According to BEA, the THS passed from 3 to 13 degrees nose-up in about 1 minute at 2:10:51, 46 seconds after a/p disconnect. Even applying the least plausible interpretation to the report, that it was already arriving at 13 deg at 2:10:51, THS would just have started moving when the a/p disengaged.

Not saying the same thing are they. You are interpreting a version that assumes completion of THS movement at 2:10:51 (which is only 46 seconds and quite short of "about 1 minute" and since they have exact times, why not state 46 seconds?), my read is that starting around 2:10:51 the THS moved from 3 degrees NU to achieve 13 degrees about minute later.

You pick? Either is possible from the given text. Note the PF NU inputs (maintained) at and beyond 2:10:51... what would they do to the THS?

[DozyWannabe]

@bear - The article also states that the turbulence was moderate, not severe. The A/P is likely to have disengaged at the first sign of pitot blockage, which was some seconds before the computers confirmed ADR DISAGREE.

[gonebutnotforgotten]

I am still not convinced by most theories posted here on why there was a nose-up input as soon as the autopilot dropped out due to unreliable speed data. The most rational so far:
it was an involuntary action, an unwanted side effect of putting in a left roll demand to counteract a right wing low tendency, or
It was the response to a false indicated speed increase,

I find (1) unconvincing because it was one hell of an input, sufficient to cause a 7000 fpm climb and, apparently, two stall warnings, sounds sort of deliberate to me.

similarly (2) sounds wrong because, while we don't know for sure what the indications were on the PF's panel, due to the lack of recorded data, we do know that the other two systems showed abrupt speed decreases, and the previous high altitude pitot icing incidents, including the Air Caraibes (http://http://www.eurocockpit.com/docs/ACA.pdf)precursor, also only registered speed decreases. So it would be extraordinary for the unforunate AF co-pilot to be the only one to have been presented with the opposite.

That same Air Caraibes report also said that at the time of the beginning of the anomalous air data, the indicated altitude rapidly dropped 300 ft (presumably due the lack of a vaild Mach number to correct the static pressure (? ) the BEA has already told us that with no vaild M, the stall warning system uses a value near zero...) Maybe the PF was reacting to that, admittedly 300 ft isn't much, but I suspect it looks a lot on the tape, and it would have been rapid. Perhaps this is the trigger we've been looking for.

[bearfoil]

At 2 h 10 min 51 , the stall warning was triggered again. The thrust levers were positioned in the TO/GA detent and the PF maintained nose-up inputs. The recorded angle of attack, of around 6 degrees at the triggering of the stall warning, continued to increase. The trimmable horizontal stabilizer (THS) passed from 3 to 13 degrees nose-up in about 1 minute and remained in the latter position until the end of the flight. BEA. The first two statements, by BEA, describe a recovery from APPROACH TO STALL.

The next Statement describes a good picture of the THS after its position at a/p loss plus PF's additive, created the climb. If PF was unaware of THS postion at handover, he is figuring, hmmm. three degrees. His attempted recovery from APPROACH TO STALL, instead of being from three degrees, was from SIX, which he appears to have held.

If he had captured TOGA, (such as it was), the absence of the STALLWARNING means he is a success. Unfortunately, the TRIM is helping him maintain back pressure up to thirteen degrees.

SO. Two recoveries from APPROACH TO STALL. Holding back pressure all the way. With the eager help of THS.

[Zorin_75]

garageyears:
my read is that starting around 2:10:51 the THS moved from 3 degrees NU to achieve 13 degrees about minute later.Actually I absolutely agree with you. Please read again, what I was saying is that even when read that unreasonable way, you can't get the report to align with bear's theory of the a/p disconnecting as the THS reached its end stop (which is also completely ignoring that we have already quite a good reason for the a/p disconnecting in the UAS).

[bearfoil]

Zorin

Your turn to read again. I say no such thing. THS at 3 degrees, at drop. With the a/c descending and rolling right. AT STALLSTALL, PF selects back pressure to escape the Warning this is not the same as NOSE UP. It is agee thing, to which the THS responds. The only training he had re: APPROACH TO STALL was certainly without the THS involved? PITCH is protected in A1 and NL.

yep.

[Zorin_75]

Your turn to read again. I say no such thing. THS at 3 degrees, at dropUhm, I did:

Ten degrees nose UP? That tells us nothing about the AoA, or the position of the THS. Was it already at 13.2? Did pilot cease his initial NU and begin his repeated ND inputs? The 7000fpm suggests the climb was (initially) very rapid, and there is no reason to think the THS was not "UP" having corrected for a chronic descent?
A/P will hang in until 13 degrees Pitch UP before it quits. If the descent was transient, and chronic, why shouldn't the THS react with 13 degrees nose up?
This means that the THS may have been responsible for the A/P quitting. 13.2 degrees exceeds its Limit for TRIMMING.

[DozyWannabe]

@bear - in anything other than Normal Law, protections can be overridden with sufficient control input. If you're now suggesting that he pulled hard backstick expecting the protections to have his back, then he either failed to hear the call "Alternate Law" from the PNF or there was something wrong with his training. All this supposes that your theory is correct - for which, of course, we have no evidence.

[PJ2]

The THS "reaching its stops or limits at "2deg (ND) or -14deg (NU)" will not cause the a/p to disengage.

[mm43]

I doubt if there is overly much difference between the initial FDR printouts (http://oi55.tinypic.com/23lfk8k.jpg) of the Jetstar VH-EBA UAS incident and those recorded by AF447. Have a close look at what happened to the TAT, SAT and CAS traces and similarly to the altitude when the Mach correction wasn't available.

Probably worthwhile bearing those traces in mind when discussing the initial upset. How LOC came about is another matter.

[mm43]

PJ2;

A nagging thought has been, why wouldn't/didn't the THS move the additional degree to 14° NU with what appears to have been fairly constant NU inputs?

[GarageYears]

bear:

Where do YOU get this stuff from? Nowhere does anything state the aircraft was DESCENDING at the start of event, except in your posts - I honestly think the BEA might have mentioned such a condition. And now you are imagining some kind of AP drop off at THS authority limit.... This is bunkum and the worst kind of dramatic imaginings, solely it seems an attempt to fit some other version of reality to the BEA framework in a twisted fashion.

[GarageYears]

Zorin_75: :ok:

(re-read your post... got it third time round :ugh:)

[gums]

Salute!

- First of all, there are "stalls". Then "deep stalls". Then "deep" stalls. So despite my posts describing a true "deep stall" in another jet, I can't say for sure if the Bus will do the same thing. I brought it up because the other jet was FBW and it was not a T-tail ( the config normally associated with a "deep stall"). My only valid comparison is the FBW system tried to do its best to no avail. In our case we had a large aft c.g. and a demonstrated pitch moment at a certain AoA that allowed the jet to settle into a true "deep stall".

Then there are "deep" stalls. Way beyond C-sub-ell max, but basic static stability and nose down stab/elevator usually allows recovery.

- I can't think of any pilot that does not employ the OODA principle, especially when making an instrument approach. Further, if it had not been for Boyd, Moody Suter, Sprey, et al, I would never had the chance to fly the first operational 100% FBW jet ( and no trim wheel for the THS, just our last-ditch manual pitch override doofer).

- PJ and another one or two agreed with my view of the THS trim law, but another pilot has implied that under certain conditions or reversions that the pitch control trim becomes more of an "attitude" command versus a gee command, and then an AoA command when certain limits are reached.

My point is that from the manuals' description, the HS is trimmed so that a stick command is always a gee when the gear is up or some AoA limits are reached or .... Gets confusing, don't it? So pull back a bit, then relax and the jet climbs at 1 gee corrected for pitch. So at a 30 deg attitude, it would trim the jet hands-off for a 0.87 gee normal force on your body, at 45 degrees 0.707, and the beat goes on. This makes sense for a heavy, as a continuos 1 gee would result in ever-increasing pitch, wouldn't it?

So the jet might appear to be in an attitude control mode, but damned if I can find that in the descriptions.

The potential problem that may deserve attention in this accident and with respect to the Airbus control laws is what does the THS do if the pilot has a small, but constant back stick when the jet is already at 1 gee , or even a bit less ???

more later.....

[PJ2]

mm43;

I believe the slightly higher limits are the mechanical trim stop limits.

gums, when I first checked out on the A320, a really sharp friend (also doing the same) observed that the one-gee law would actually cause a slight increase in the rate of climb as the aircraft got further away from earth... ;-)

[gums]

Salute!

Gee, PJ, I thought I was losing it....

the one-gee law would actually cause a slight increase in the rate of climb as the aircraft got further away from earth... ;-)

Unlike our little jet with different operational requirements, that's what we saw when we relaxed pressure on the stick in a slight climb. Hmmmmm......

So I was pleasantly surprised when I read the Airbus pitch correction for the "gee input" by the stick. Made sense to me for a heavy.

One of my dreams is to have all the heavy pilots fly a system that commands gee until the jet reaches a certain AoA, then follows the AoA limit ( I don't like the term "protection"). Upon releasing the stick the system tries to establish the last gee command, whether a trim gee by the confusers or a manually set gee. I know for a fact what my jet's laws did.

I have a bad feeling that the Bus may not re-trim the stab real quick after a prolonged maneuver at the AoA limit, then just relaxing on the stick.

respectfully,

[PJ2]

Funny what goes around when no one knows...never saw the "slight increase"! Funny what still goes around...

I think the term "protections" conveys a prophylactic sense which speaks to some airline folk who want to reduce training, hire cheap and claim that commonality means one-for-all training. You can always tell a salesman but not very much.

The A332 trim rate is 0.2deg/sec on AP and I believe the rate is the same for manual flight using the sidestick. The same 3 electric motors, (only one operating at a time, IIRC) drive the THS. Manual/mechanical trim is meant to be used and employs two hydraulic motors. The mechanical stops are part of both systems. A pawl and ratchet system prevents back-movement under air loads. From what I have been able to learn, I don't believe that the THS would return to a previous setting after a prolonged manoeuvre if the stick is merely relaxed, (neutral, no orders to the FMGEC > FCPCs). I believe the THS would remain "as is" in manual flight and the airplane would "settle" into the setting just as any airplane would. In autoflight it would of course, trim.

[PickyPerkins]

Not disagreeing at all with RetiredF4's detailed post, and in fact building on it, I would like to ask a question concerning the situation once the a/c had established an AoA of about 61°.

http://pickyperkins.home.infionline.net/A330DiagramGroundSpeed.jpg

Using the graphs in HeavyMetall's reference for the lowest aspect ratio (6), isn't it the case that stick forward, elevator down, would lead to a decrease in the lift on the THS, and hence a more nose-up attitude of the a/c?

http://pickyperkins.home.infionline.net/PostStallCL.bmp

So for a pilot getting to the bottom of his barrel of options, might that inhibit further stick forward experiments, and strongly encourage a stick back tendency?

http://pickyperkins.home.infionline.net/PostStallCD.bmp

It is true that drag on the THS would also increase and that would tend to act in opposition, but by how much?

[CONF iture]

I've searched PPRuNe to see if this has been posted already but apparently not. I believe the AW&ST article is the first actual commentary on who was flying.
PJ2, Actually I did put a link for this one (http://www.pprune.org/tech-log/452836-af447-thread-no-3-a-80.html#post6500012), but I certainly wouldn’t blame you for not noticing …
I don’t really want to comment further about it, but I take that opportunity to admit one more time, if needed, how little respect I have for my beloved BEA :
Faced with the regular disclosure of partial and often approximate information since 16 May, the BEA wished to publish this note
Their note is nothing else than that : partial and approximate

I know you don't agree, I know how people want to trust their Institutions, but all I need is the full data, nothing more, but nothing less.

Goodnight to all !

[RetiredF4]

@ PickyPerkins

Thank you for doing the handywork posting the graphs i referenced to. I´m too dumb on computers to do it:).
You visualized my point of view.

[DozyWannabe]

I know you don't agree, I know how people want to trust their Institutions, but all I need is the full data, nothing more, but nothing less.

I ask again - why is it you're asking this of the BEA and no other accident investigation bureau/institution? Is one incident forever going to preclude acceptance of any finding of theirs that includes "pilot error" as a factor?

Also, you're a pilot, I'm an engineer - but I wouldn't trust myself to derive a correct conclusion from the raw data. What makes you so sure you could?

[A33Zab]

PJ2:

The A332 trim rate is 0.2deg/sec on AP and I believe the rate is the same for manual flight using the sidestick. The same 3 electric motors, (only one operating at a time, IIRC) drive the THS. Manual/mechanical trim is meant to be used and employs two hydraulic motors. The mechanical stops are part of both systems. A pawl and ratchet system prevents back-movement under air loads. From what I have been able to learn, I don't believe that the THS would return to a previous setting after a prolonged manoeuvre if the stick is merely relaxed, (neutral, no orders to the FMGEC > FCPCs). I believe the THS would remain "as is" in manual flight and the airplane would "settle" into the setting just as any airplane would. In autoflight it would of course, trim.


I’ve got these figures of THS rate.

If they stated: "there's only 1 motor in control" they mean the ELECTRICAL motors (P1,P2 & P3) whichs drives the control sleeve. Control sleeve position will be overriden by mechanical input (from handwheels)


The THS itself (the ballscrew) is driven by 2 hydraulic (B & Y) motors.

Max. Operating Load and Max. Speed (2 motors)

18940 daN (42578.81 lbf) ----- 0.4°/s
16950 daN (38105.11 lbf) ----- 1.0°/s
6770 daN (15219.56 lbf) ---- 1.2°/s

Half speed for 1 motor (1 hydr. B or Y system failure) operation.

Since the hinges are on the trailing edge and therefore the aerodynamic load on the drive spindle is in the AND (stab. Nose Up) direction (see diagram by Tailspin Turtle) it would take 12 sec. to travel from 13° back to 3° ANU. Or is that a too simple conclusion?

The rate for the BEA mentioned 1 minute to go from 3 to 13 ANU was 10°/60 = 0.16°/s.

Limit load (both compression as tension) is 32500 daN (73062.9 lbf).
I assume this limit load is mentioned as being the limit for the Hydr. motors to drive the THS.

[Hyperveloce]

Hi there,
Just wondering about the opportunity to provide to the pilots additional flight parameters such as the AoA and beyond: wouldn't it be possible to measure the efforts generated by all the control surface actuators (in response to the flight control loop orders), to estimate directly/specifically their aerodynamic authorities and display them ? (roughtly, the harder to actuate the control surface, the more aerodynamic authority it has ? or an estimation performed through more sophisticated aerodynamical models for each control surface and the airframe ?).

[Lonewolf_50]

gonebutnotforgotten
That same Air Caraibes report also said that at the time of the beginning of the anomalous air data, the indicated altitude rapidly dropped 300 ft (presumably due the lack of a vaild Mach number to correct the static pressure (? ) the BEA has already told us that with no vaild M, the stall warning system uses a value near zero...)Maybe the PF was reacting to that, admittedly 300 ft isn't much, but I suspect it looks a lot on the tape, and it would have been rapid. Perhaps this is the trigger we've been looking for.
Man-machine interface. Is the unreliable M related only to airspeed/pressure sensing, or is it also related to a TAT sensing anomaly?
mm43
I doubt if there is overly much difference between the initial FDR printouts of the Jetstar VH-EBA UAS incident and those recorded by AF447. Have a close look at what happened to the TAT, SAT and CAS traces and similarly to the altitude when the Mach correction wasn't available.
Probably worthwhile bearing those traces in mind when discussing the initial upset.
How can BEA confirm or determine if the ice and pitot tube interaction was accompanied by an ice TAT probe interaction?

I will guess that a trend line from "X" minutes prior to the event could show where Temp anomalies are or aren't likely. The recorded cockpit briefing about temperature not developing as planned/forecasts may be of no moment ... or maybe a clue pointing to TAT readings sensing beginning to degrade as Airspeed sensing began to go wrong?

[JenCluse]

Given the intense interest from the piloting community in this baffling event, and my previous decision to refrain from posting in case this ancient showed how out of touch she was, is there any way to bulk download this series (1-4) of threads, so that one might search previous posts on a specific topic?

(This after skimming over a post which I can't now find, which seemed to suggest that the standby instruments (ICUS, I believe they/it are called) are not fundamental, steam-driven devices, but take their data input from some branch of the computed data stream.)

I can find the printable area, but 40 posts only. I'd like to be able search any discussion point. Anyone?

[HazelNuts39]

I find (1) unconvincing because it was one hell of an input, sufficient to cause a 7000 fpm climb and, apparently, two stall warnings, sounds sort of deliberate to me.Was it 'one hell of an input'? 7000 fpm is achieved after 18 seconds of 0.2 g; 12 seconds of 0.3 g, or 9 seconds of 0.4 g.

the BEA has already told us that with no vaild M, the stall warning system uses a value near zero...)BEA Interim Report #2:
In alternate or direct law, the angle-of-attack protections are no longer available but a stall warning is triggered when the greatest of the valid angle-of-attack values exceeds a certain threshold. In clean configuration, this threshold depends, in particular, on the Mach value in such a way that it decreases when the Mach increases. It is the highest of the valid Mach values that is used to determine the threshold. If none of the three Mach values is valid, a Mach value close to zero is used. For example, it is of the order of 10° at Mach 0.3 and of 4° at Mach 0.8. This explanation is either in error, or we are misinterpreting it (The BEA Update adds a new meaning to 'invalid airspeed'). In the case of the UAS incident documented in the Air Caraibes Memo, stall warning occurred at 4.2 degrees, with ADR DISAGREE. In the case of AF447, the stall warning must have been at a similar value, because the low-speed value of 10 degrees is well beyond stall AoA.

[DozyWannabe]

@JenCluse:

I think you're talking about ISIS (shown here)

http://www.samtel-hal.com/cache/isis.pdf

It's a digital display, and yes air data will still come from the pitot tubes and static vents (as would be the case with "steam gauge" hardware too), but the data is taken from standby units as I understand it - it is intended to be independent of the main systems. At present there's no evidence that any data on either the main PFDs or the ISIS display was incorrect - other than airspeed data, which was anomalous for a duration of approximately 45 seconds.

[takata]

Hi JenCluse,
I'm also feeling your pain as I'm trying to catch-up with this thread after a while.

(This after skimming over a post which I can't now find, which seemed to suggest that the standby instruments (ICUS, I believe they/it are called) are not fundamental, steam-driven devices, but take their data input from some branch of the computed data stream.)
I can find the printable area, but 40 posts only. I'd like to be able search any discussion point. Anyone?
I understand that you are refering to ISIS and what you think is, in some way, both right and wrong. Stand-by instruments are supplied by analogic pnematic sources: probe pitot 3 and static probe 3. However, both probes are used for direct sensing and/or connected to numerical Air Data Module (ADM). In this case, those ADMs are feeding ADR3 which may replace either ADR1 (Captain probes) and ADR2 (FO probes) while the stand-by gauges (ISIS) are displaying the direct measures from the same probes.

You can also refer to the BEA first report which have a detailed diagram with usefull explanation about Airdata, page 22:
http://takata1940.free.fr/Airdata.jpg

[MurphyWasRight]

LW50:

The recorded cockpit briefing about temperature not developing as planned/forecasts may be of no moment ... or maybe a clue pointing to TAT readings sensing beginning to degrade as Airspeed sensing began to go wrong?


I missed the implications of the briefing on temps not developing as expected before now but this fits well with my (just a)theory that almost every item included in the BEA note will in hindsight appear to be relevant to the final determination of the event chain.

As I have said before they (BEA) have a much better picture of what happened than outsiders so the note would have been written with the likely ultimatly signinificant factors in mind.

Also possibly significant is that the final item in the note is the mention of dual control inputs (no hint if matching) and the transfer of control.

My take on this is that beyond this point there was still (at the time the note was created) discussion/uncertainty within BEA as to whether the situation was (even theoretically) recoverable and the level of understanding of the situation in the cockpit.

[jpete]

JenCluse

is there any way to bulk download this series (1-4) of threads, so that one might search previous posts on a specific topic?

Although it isn't a download, you can achieve a search of only the 4 threads by using the following search string in Google
ths af447 site:http://www.pprune.org/tech-log/
this will search for mentions of THS in the AF447 threads of tech log only. Just change the THS in the string to whatever you want to look for. Adding the site:URL end part is the magic that restricts Google to only searching in the tech log on here

Copied to post #1 for general dissemination

[Chris Scott]

A33Zab,
That explanation is invaluable, thanks. So it's only too easy to get confused between the electric (control) motors and the hydraulic (actuation) motors...

From the auto-trim point of view, the movement from 3NU to 13NU (after FL375) was very nearly continuous? I guess it only uses the one speed of 0.2deg/sec, as quoted by PJ2. Although you state a 2-motor speed of 0.4deg/sec, perhaps limiting the speed to 0.2 would avoid any slow-down if either the blue or yellow hydraulic motor failed?

On the face of it, those who have advocated use of manual trim-wheel in stall recovery seem to be vindicated, as its 1deg/sec might be needed. The argument probably only applies in Pitch-Alternate Law. We still have no idea what the effect of full down-elevator is when the THS is at 13NU. (That is: what would be the equivalent THS setting with neutral elevators?) Sadly, the answer does not seem to be pertinent to this accident.

Quote:
Since the hinges are on the trailing edge and therefore the aerodynamic load on the drive spindle is in the AND (stab. Nose Up) direction...

Please forgive me for offering readers a warning on terminology here, from my experience of various jet transports with THS under one name or another. "Nose-up" is a term normally used to indicate the trim-effect on the aircraft, not the physical position of the THS. Because the THS is hinged at its aft spar, it's easy to think of "nose-up" as meaning that the leading-edge of the THS is up. Not so.

For maximum nose-up trim, a THS is at its fullest NEGATIVE incidence, with its leading-edge fully down. The A330's mechanical limit seems to be -14 (see PJ2/mm43 discussion). We call that "14 degrees nose-up" (14NU). For a typical take-off, the THS setting calculated on the trim-sheet by load-control might be 3NU. This requires the incidence of the THS to be -3 deg. However, load control and most pilots refer to it as "+3". Confusing, perhaps, but that's the way it is.

[HazelNuts39]

wouldn't it be possible to measure the efforts generated by all the control surface actuators (in response to the flight control loop orders), to estimate directly/specifically their aerodynamic authorities and display them ? In fact the A320 system does something very similar when it compares the gross weight calculated from air data and inertial data to the gross weight calculated in the FMS from loadsheet and fuel used. In case of a discrepancy, a CHECK GW message is generated. (See Perpignan Final report). I don't know if that function is available in the A330. Using that same logic in reverse the airspeed could be calculated from the FMS weight. The BUSS presumably does something similar.

[Lonewolf_50]

Murphy, about misunderstanding in the cockpit, it might get curiouser and curiouser ...

IF
you lose airspeed, or it becomes unreliable
AND
You are an A330 Pilot who is already reasonably familiar with previous incidents in your model aircraft ... since the manufacturer and your company put out bulletins and findings of previous UAS incidents
AND
You are in a UAS incident yourself

You do X, Y, and Z in response. (Comments in the cockpit on loss or lack of indications, which I presume it means airspeed, is what I am working with here ...)

At what point can you determine that the airspeed has stopped acting up and become reliable again?

The forensic analysis points toward about 45 seconds worth of bad airspeed, and then a return to reliable state ... how does the crew know that it has returned to reliable? Did they ever have a cue (problems with stall AoA's on pitot tubes considered) that AS was telling them good information again?

What is dwell time (human observation and perception not being a step function) between "it's bad" and "it looks good again" in a

Benign cockpit environment?
High task load cockpit environment?

Do the training pamphlets or the sim sessions devoted to UAS training address what symptoms tell you that your airspeed tapes are once again reilable?

You aren't using the old style circle gages, where the needle might stop agitating and then return to a more normal looking state. You are looking at the vertical strip. What discrete cues do you have? From the air data diagram that takata so kindly posted, it seems that the ISIS display might be the first cue, depending upon which pitot tube got itself back in order first.

I apologize for asking what might be an idiotic question, but I've not flown an A330, nor been exposed to A330 UAS malfunction/emergency training.

Hazelnuts39:
Was it 'one hell of an input'? 7000 fpm is achieved after 18 seconds of 0.2 g; 12 seconds of 0.3 g, or 9 seconds of 0.4 g.

Do you mean 1.2, 1.3 and 1.4g? (I am guessing that you meant a delta from a "stable" 1 g reference.)

[RR_NDB]

Hi,

Hyperveloce @ Post #105 (http://www.pprune.org/tech-log/454653-af-447-thread-no-4-a-6.html#post6519147)

opportunity to provide to the pilots additional flight parameters (http://www.pprune.org/tech-log/454653-af-447-thread-no-4-a-6.html#post6519147)

Sounds a very good idea. And with a good approach: To provide extra and (redundant) important info

Considering the computational capabilities in these "advanced planes" it seems it would be quite easy to do this "pattern recognition" in real time.

Question:

What kind of sensors you imagine in order to provide the data to the System? Some are yet available?

[syseng68k]

Hyperveloce, 105


Hi there, Just wondering about the opportunity to provide to the pilots
additional flight parameters such as the AoA and beyond: wouldn't it be
possible to measure the efforts generated by all the control surface
actuators (in response to the flight control loop orders), to estimate
directly/specifically their aerodynamic authorities and display them ?
(roughtly, the harder to actuate the control surface, the more
aerodynamic authority it has ? or an estimation performed through more
sophisticated aerodynamical models for each control surface and the
airframe ?).

Have been thinking about something similar, but in relation to wing loading
as an extra parameter input for the calculation of lift. For the control
surface case, this could be instrumented via strain gauge type transducers,
one would think, near the control surface root.

In the wing loading case, deriving lift force is complicated by acceleration
forces, but since other areas of the system (adirs) already measure this, the
acceleration could be subtracted out to derive actual lift on the wing. How
usefull any of this might be, I don't know, but there are times when the more
cross check type data that there is available, makes it easier for dumb logic
to determine what is really happening....

[HazelNuts39]

Lonewolf50;

Yes, your interpretation is correct. Added to the acceleration due to gravity, the 'sensed' acceleration values are 1.2, 1.3 and 1.4 g. The values of rate of change of velocity (dV/dt) are 0.2, 0.3 and 0.4 g. Sorry for being somewhat pedantic, couldn't help it.

[Smilin_Ed]

LW50: At what point can you determine that the airspeed has stopped acting up and become reliable again?

When all of your speed indications are again the same and are consistent with the pitch and power that you have been flying to prevent upset.

[henra]

Using the graphs in HeavyMetall's reference for the lowest aspect ratio (6), isn't it the case that stick forward, elevator down, would lead to a decrease in the lift on the THS, and hence a more nose-up attitude of the a/c?



This is maybe to some extent true if you change the angle of the THS.

But commanding ND Elevator does not change so much AoA but it changes the camber and thus the effective profile of the tail surface, thereby completely changing the lift curve.
You would need to compare the Cl of the cambered tailplane. I'm pretty sure the cambered tail will have more lift even at increased AoA.

Moreover as you already describe ND elevator would increase drag, thus increase ND forces. At these angles you can exopect the drag to be of similar magnitude if not higher that the Lift, especially for a heavily cambered profile.
what is also missing and what will also be influenced by ND elevator is the moment of the airfoil. Strong camber will add a ND moment.

Taken these aspects altogether, even if I have no specific data for the given case I would strongly tend to believe that ND elevator would have lead to Nose Down attitude change even at these AoA.

[PJ2]

A33Zab;

Yours is a valuable contribution to this aspect of the discussion. My thanks to you for clarifying the THS system and correcting, for other readers, my mistaken impression of the THS operation. I had read the electric-motor component as "the slower" of the two driving systems, ostensibly due to the lower demands of the AFS and, in manual flight, the sidestick. I've gone back to what I have and re-read with your contribution in mind.

Chris Scott;
From the auto-trim point of view, the movement from 3NU to 13NU (after FL375) was very nearly continuous?
I've watched the A340s and A330s THS wheel on the pedestal move slowly in response to small stick movements and changes in airspeed. One can see the movement out of the corner of one's eye because of the paint but I always valued the aural confirmation of movement on those airplanes that had manual trim.

The THS response to sidestick orders is immediate, in other words. I believe then, that as the sidestick was held in the NU command position, (orders still gee-referenced but with rate/gain limits), the THS would follow through, likely at a varying rate in relation to slightly changing sidestick orders. I think once the AoA was below that which the autotrim ceases to function, it would respond to sidestick ND orders. As stated, it would certainly have responded to manual trim wheel input.

A good, (and needed) discussion on "NU" and "ND", thanks. That's exactly the way the "minus 4" symbol for the THS position on the FLT CTRL ECAM page was explained to us when learning the A320 in 1992. Minus meant "tail down, nose up", and I think made good sense once one understood.

On the "roll to the right" upon disconnection, I have to re-iterate that the fuel system for the A340 and A330 were next to flawless in maintaining cross-wing balance of fuel. And there is nothing in the BEA Note to point to cause but any fuel imbalance on this airplane is real cause for close observation. I can't see rudder trim being an issue because it is automatic as well. One doesn't "disconnect the airplane, trim, then reconnect" as one did with the DC8/B727, etc and at that point certainly, it isn't reasonable to consider that asymmetric thrust was a factor.

[henra]

Hyperveloce, 105


Have been thinking about something similar, but in relation to wing loading
as an extra parameter input for the calculation of lift. For the control
surface case, this could be instrumented via strain gauge type transducers,
one would think, near the control surface root.


syseng68k
If I misnderstood what you wrote I apologize upfront, but the wing loading in a stable stall will be roughly the same as in level flight, i.e. 1g.

The main difference would be the exact point where the lift is applied. In a fully stalled wing this will move from the Quarter Chord to the middle of the chord, so the moment on the wing might change a bit but the vertical force won't change much.
In a really deep stall on the other hand it might indeed work as the force on the wing it will decrease a significant bit as the fuselage will take over some part of the lift. In the attitude of AF it might contribute ~30%, so it could work to some extent but only at a point where it might be too late.

[Lonewolf_50]

henra:

With a little "back of the napkin" sketching, I arrive at this from your post number 120 (http://www.pprune.org/tech-log/454653-af-447-thread-no-4-a-6.html#post6519980) of this thread.

With the THS not stalled, though at an angle of attack significantly displaced from the usual "nearer to horizontal" (horizontal reference being THS airfoil chord line), the effective lift force, and thus corrective pitching moment, (roughly perpendicular to longitudinal axis of the aircraft, and then acting along the fulcrum to tip the aircraft's nose back down) is reduced roughly as the sine of the angle (AoA). The larger the AoA, from a nominal airflow parallel to chord line, or AoA absolute value of zero, the greater the lost force (vector subtraction). I din't factor in camber, as I both don't know it, and expect that it would wash out. This presumes no stall. Once a THS is stalled, it's a different problem.

I sketched it out this way based on expected control forces, controllability, and control response from an input to an airfoil movement. Each knot of reduced airspeed robs the lifting force of magnitude proportional to the 1/2*k*v^2 relationships in the lift equation.

What does this do to the aircraft's response ...

even with the THS not being stalled at that unusual angle of attack, it would take longer for both the elevator to influence THS, and for the lift acting on the THS to move the tail up (and thus the nose down) since the lift force isn't pushing as hard for a given flight control pitch command.

At some point at very low airspeed, depending upon CG, the airspeed ( airflow? ) may be so low that there isn't sufficient pitching moment (due to the combination of low speed influencing lift reduction, low air density, and sine of the angle vector subtraction) to move the nose at all.

Why sine of the angle?

As I sketched out the FBD on the THS, my vector component NOT adding lift increased with increased AoA, lift vector component decreasing.

Does that make sense to you?

EDITED since it was really confusing to me, and I wrote it. :p

Second EDIT: the more I think of this, the more it seems to me that an AoA indicator might be handy at high altitude. The AoA margins are reduced up there, a point which strikes home as I review HazelNuts39's graph on Mach Number and AoA and stall warning. If you lose speed at high altitude, that thin air really puts you at a disadvantage in terms of what your flight controls can do for you. :eek:

[Chris Scott]

PJ2,
Thanks for reminding me that the THS display on the ECAM F/CTL page (the lower display unit, not normally displayed in the cruise), would have eventually shown either
"13° UP" or "-13° UP" for what I refer to below as "13NU". (Perhaps you can tell us which?)

As far as I know, the crew were not presented with any ECAM drills on that lower DU, despite the various failures timed at 0210 and 0211. Presumably, the CRUISE page would still have been displayed, which does not include THS? If so, the best way of reading the THS position would have been the scale on the manual pitch-trim hand-wheel, but that involves turning the head. As I think you said, there is no "whooler", or equivalent, to announce THS movement.

Quote:
The THS response to sidestick orders is immediate, in other words. I believe then, that as the sidestick was held in the NU command position, (orders still gee-referenced but with rate/gain limits), the THS would follow through, likely at a varying rate in relation to slightly changing sidestick orders. I think once the AoA was below that which the autotrim ceases to function, it would respond to sidestick ND orders.
[my highlighting]

I'm not sure you needed to include the clause highlighted, so will try to explain why.

From 02:10:51
Would the auto-trim look at AoA when and while the system considered the AoA data invalid? Was the "Vc Prot Law" or "limited-authority stability order" of Andy Tracey's paper disabled some of the time – or constantly? When the stall warning was operating, the AoA was considered valid, so the THS would stop provided the stick was neutral.
When the stall warning was not operating, the THS could auto-trim if the AoA was below the stall-warning threshold. And if the AoA data was invalidated, my understanding is that it could auto-trim regardless of how high the AoA was.
But would the back-stick not have overridden the "Vc Prot Law" anyway? If that was not the case, would the THS have ended up at 13NU?

Hope this makes some sense.

[henra]

What does this do to the aircraft's response ...

even with the THS not being stalled at that unusual angle of attack, it would take longer for both the elevator to influence THS, and for the lift acting on the THS to move the tail up (and thus the nose down) since the lift force isn't pushing as hard for a given flight control pitch command.

At some point at very low airspeed, depending upon CG, the airspeed ( airflow? ) may be so low that there isn't sufficient pitching moment (due to the combination of low speed influencing lift reduction, low air density, and sine of the angle vector subtraction) to move the nose at all.


I agree with the much slowed response at the low speeds and AoA. It is quite safe to assume that any recovery at these speeds and angles would have been relatively slow.
Where I'm not so sure if I understand it correctly is the thing with the Sine.
I will give it a try based on how I understood your approach:
As long as the flow does not separate it is not the speed along the chord but the speed along the Flight Path which is relevant for the lift. Therfore the lift component which is trying to tilt the aircraft Nose Down around an unknown center of rotation will be simply a function of the square of the speed along the trajectory.
The sine will apply with regard to the center of the Earth (i.e. the external coordinate system) but not for the coordinate system of the aircraft itself. Therefore it should be irrelevant for Nose Down with regard to the aircraft coordinate system. This assumes that the sum of external forces acts at one point on the aircraft and acts in the direction of the flight path.
This is a little oversimplification because the external forces consist of gravity, momentum, aerodynamic effects and thrust of the engines. It has to be noted that the latter two do not act exactly on the same point on the aircraft therefore making my assumption not 100% correct. Therefore the aircraft coordinate system is not completely independent.

Was a bit lengthy and probably even more difficult to understand what I mean but I hope I understood your description correctly if not you can ignore my comment.

[PJ2]

Hi Chris - I'm struggling a little to understand what you mean. I'm not referencing anything too complicated, just the statement regarding the Abnormal Law, below.

Now I know the BEA has stated that the airplane never reverted to "Abnormal Law" as a result of the high AoA but they dont' explain why and that is really confusing. Another piece missing, which will show up sooner or later.

Here's my reference - not sure it answers your question:

Abnormal attitude law
This law from the FCPCs is engaged when certain aircraft parameters exceed predetermined values.

The laws in place would be:
- in roll: - yaw alternate law
- in pitch: - a modified Nz law, without autotrim.

After aircraft recovery, and until the aircraft is on the ground, the available laws are:
- in roll: - yaw alternate law, (in this case Alternate Law 2)
- in pitch: - Nz law, (with recovered autotrim).

"In haste", as you say Chris!...gotta do some work. Hor's de combat so to speak, over the next while.

[syseng68k]

henra, 122


If I misunderstood what you wrote I apologize upfront, but the
wing loading in a stable stall will be roughly the same as in level
flight, i.e. 1g.
Not at all and thanks for the patience. Shows just how much I (don't)
know about aerodynamics.


The main difference would be the exact point where the lift is applied.
In a fully stalled wing this will move from the Quarter Chord to the
middle of the chord, so the moment on the wing might change a bit but
the vertical force won't change much. In a really deep stall on the
other hand it might indeed work as the force on the wing it will
decrease a significant bit as the fuselage will take over some part of
the lift. In the attitude of AF it might contribute ~30%, so it could
work to some extent but only at a point where it might be too late.
Ok, if lift doesn't change significantly until beyond stall, but
the position does, other methods could be used, such as >1 pressure
transducers at critical points on the surface. If these were summed
differentially. the position of lift could be calculated, thus approach to
stall.

Going back to the strain gauge solution, assume that the lift position
produces a variable twisting moment to the wing ?. If so, this could
be measured to determine lift position.

Probably completely off the wall and i'm sure there's no shortage of
ideas, but it is Friday :)...

[Chris Scott]

Quote from Lonewolf_50:
The recorded cockpit briefing about temperature not developing as planned/forecasts may be of no moment ... or maybe a clue pointing to TAT readings sensing beginning to degrade as Airspeed sensing began to go wrong?

I'm inclined to think the former. Looking again at that Jetstar A330-200 incident FDR-trace (http://oi55.tinypic.com/23lfk8k.jpg) that mm43 kindly provided us last night (GMT), the TAT down-spike when Mach and CAS collapse is quite small (~5C), whereas the coincident up-spike of SAT is large (nearly 25C), in order nearly to meet the TAT – as you would expect at Mach 0.2.
When looking at the possibility of step-climbs, we look at the decreasing weight and the SAT (and its deviation from ISA), not TAT. Some large jets are limited mainly by buffet margins, others by thrust (the classic examples being respectively the VC10 and B707-320). With its high-bypass turbofans the A330-200 is thrust-limited, so the higher the temperature, the later the step-climb. The crew would have been comparing the SAT readings with the forecast OATs on their computer flight-plan: hence the remark.

Considering your second option, on the other hand: if the CAS/Mach readings had already started insidiously under-reading, causing an over-reading SAT, the A/C would have been actually speeding-up without their knowledge. Not sure if it's possible to prove or disprove that possibility. GS doesn't help because we don't know the wind-component.

[A33Zab]

PJ2:
I didn't go into 'Abnormal Law" as stated by BEA and 'confirmed' by this data:

If CAS < 60 Kts AOAi & AOAc are set to 0° and SSM (System Status Matrix) is set to NCD (No Computed Data),
this is also valid for TAS 0 Kts if CAS < 60

If CAS < 30 Kts it declares itself invalid and outputs 0 Kts and NCD.
........
(Stall warning is generated by highest AOA and not the median value)
This means actual AOA could be indeed 40° while not triggering ABNORMAL ATTITUDE LAW due to a median *AOA* (ouch, made a mistake here, should be CAS) value below <60° (resulting in a AOAi of 0°) as already mentioned in the BEA ‘leak’.

http://www.pprune.org/tech-log/452836-af447-thread-no-3-a-79.html#post6499464 for whole post.

@PJ2:

would this be a correct understanding of what you meant?
and

@Chris:

Please forgive me for offering readers a warning on terminology here.


Exact! Thx to you both for making this clear and with apologies, english is not the native language.

[mm43]

A33Zab;
(Stall warning is generated by highest AOA and not the median value)
This means actual AOA could be indeed 40° while not triggering ABNORMAL ATTITUDE LAW due to a median *AOA* (ouch, made a mistake here, should be CAS) value below <60° (resulting in a AOAi of 0°) as already mentioned in the BEA ‘leak’.Did you actually mean:-

Actual AOA could be indeed 40° while not triggering ABNORMAL ATTITUDE LAW due to a median CAS value < 60 kts (resulting in a AOAi of 0°) as already mentioned in the BEA ‘leak’.

[A33Zab]

@mm43:

Yes, indeed in this context from the BEA May, 27 update:



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.
Note: When the measured speeds are below 60 kt, the measured angle of attack values are considered
invalid and are not taken into account by the systems. When they are below 30 kt, the speed values
themselves are considered invalid.


and the 'leak':


After stalling, the A330's angle of attack stayed above 35°. But while this exceeded the threshold for the abnormal attitude law, the flight control computers had already rejected all three air data reference units and all air data parameters owing to discrepancy in the airspeed measurements.
Abnormal law could only have been triggered by an inertial upset, such as a 50° pitch-up or bank angle of more than 125°. "That never occurred," says French accident investigation agency Bureau d'Enquetes et d'Analyses.


Stalled AF447 did not switch to abnormal attitude law (http://www.flightglobal.com/articles/2011/06/01/357394/stalled-af447-did-not-switch-to-abnormal-attitude-law.html)

[wallybird7]

Nobody knows what is going to happen when an a/c enters a thunderstorm. No one.

If a strong enough updraft occurs it could push the plane up at 7000 fpm at the same time sending the nose down thus possibly causing nose up trim.

In addition no one knows for sure what the airplane will do when it is in a deep stall.

One thing i know, it is difficult to fly an out of control plane with trim alone.

[wallybird7]

GarageYears

"your are implying the climb was NOT of the pilots command? I can see nothing to support that assertion. "

Theres nothing to support it either.

[Chris Scott]

Quote:
Hi Chris - I'm struggling a little to understand what you mean. I'm not referencing anything too complicated, just the statement regarding the Abnormal Law, below. Now I know the BEA has stated that the airplane never reverted to "Abnormal Law" as a result of the high AoA but they dont' explain why and that is really confusing.

Struggling? That makes two of us! Sorry it didn't make sense. By the way, I had not considered the possibility of Abnormal Attitude Law. Perhaps I should.

If the initial drop of IAS on ASI 1 and ASI 3 was sufficient to trigger Abnormal Attitude Law, what would have happened when the ASI 1 recovered to 215kt at FL375? The IAS/Mach, AoA, Pitch, and Roll no longer justified it. In these circumstances, you note:
After aircraft recovery, and until the aircraft is on the ground, the available laws are:
- in roll: - yaw alternate law, (in this case Alternate Law 2)
- in pitch: - Nz law, (with recovered autotrim).

So autotrim would have been recovered? In which case, it would have resumed trimming: partly, perhaps, to neutralise the increasing up-elevator that had been selected by the system – to maintain the G demanded by the sidestick as the airspeed fell – while the THS had been frozen. It would certainly have needed to trim nose-up for the continuing decay of actual CAS as the aircraft approached FL380.

It is unclear precisely when the AoA reached +30, but at that point the THS would presumably be disabled again. My assumption is that the 1 minute taken for the THS to move from 3NU to 13NU would have been completed before that point.

In the post you are discussing, I proposed that the PF's "nose-up inputs" may have overridden any stall protection in Pitch-Alternate Law. If Abnormal Attitude Law had been triggered, it seems from Andy Tracey's paper that stall protection is not provided.

Reading the above, it all sounds exceedingly speculative in the absence of data. I only post it to show that, after some thought, I share your present confusion.

PS
Thanks, A33Zab, I had no knowledge of the BEA's "Leak". Whoops... :{

[PJ2]

A33Zab;

Thanks for your note re Abnormal Law...yes, I knew from the BEA 'leak' that the aircraft had not entered "Abnormal Law". In my note to Chris I was describing how the autotrim function would return once the aircraft was within the AoA limits for it's operation. Why it didn't go into Abnormal Law as a result of the high AoA does make sense to me, thanks.

To be sure, in all cases of all Laws, the THS is trimmable through the mechanical system unless jammed.

Please don't worry about your English...you're doing just fine.

[barit1]

Forgive please, if this is already addressed...

During the descent - in a nearly level attitude - the aircraft turned either 90 to the right, or 270 to the left. Now then -

If the crew actually felt they were overspeed, and noted the fact they were changing direction, did they not feel the centrifugal force that would be created by that change in velocity vector?

With benefit of hindsight, this would seem to be a crude but effective means of differentiating between overspeed vs underspeed, when Pitot instruments are unreliable. Turning at anything approaching a standard rate, at or above cruise Mach, would glue one to his seat. :8

[Shadoko]

Hi,
A very humbly question: is the speed from Pitots corrected with AoA data?

If not, or if a 61° AoA is not take in compute (as "invalid"), the IAS from Pitots has to be: cos61 x square root (107² +107²) = 0.48 x 151 = 72kts IMHO (61°, 107kts refering to post#100, first figure).

[Hyperveloce]

Hi,
What kind of sensors you imagine in order to provide the data to the System? Some are yet available?

As syseng68k suggests, probably that simple strain gauge tranducers would do the job (on the mechanical link between the actuator and the control surface, and even on the control surface itself). All the strain signals (along with IMU signals, GPS signals, AoA signals, airspeed signals,...) could feed in real time an aerodymical model of the A/C (control surface+airframe) to derive these aerodynamical authorities and much more: in particular, it could also help to cross-check the integrity of all the incoming signals (navigation and attitude, AoA and airspeeds,...) in a more efficient/flexible way than it is currently done, and generate a reliable stall warning.

[CONF iture]

Please forgive me for offering readers a warning on terminology here, from my experience of various jet transports with THS under one name or another. "Nose-up" is a term normally used to indicate the trim-effect on the aircraft, not the physical position of the THS. Because the THS is hinged at its aft spar, it's easy to think of "nose-up" as meaning that the leading-edge of the THS is up. Not so.

For maximum nose-up trim, a THS is at its fullest NEGATIVE incidence, with its leading-edge fully down. The A330's mechanical limit seems to be -14 (see PJ2/mm43 discussion). We call that "14 degrees nose-up" (14NU). For a typical take-off, the THS setting calculated on the trim-sheet by load-control might be 3NU. This requires the incidence of the THS to be -3 deg. However, load control and most pilots refer to it as "+3". Confusing, perhaps, but that's the way it is.
To further illustrate that good reminder from Chris Scott, here are two views from the left hand side of the THS :

http://i25.servimg.com/u/f25/11/75/17/84/af447_19.png (http://www.servimg.com/image_preview.php?i=93&u=11751784)

http://i25.servimg.com/u/f25/11/75/17/84/af447_20.png (http://www.servimg.com/image_preview.php?i=94&u=11751784)

Good eyes are needed, but a scale is visible just forward of the THS.
Both THS are at the GND SET position, which is also 4 deg UP, or minus 4. That’s the position where the THS is automatically reset after every landing.
The 0 (zero) position is also visible, with 2 units above and 14 below.

As the CG was pretty much forward for the AF447 takeoff, at 23%, the THS had probably been set initially at a position around 6 deg UP.


I ask again - why is it you're asking this of the BEA and no other accident investigation bureau/institution? Is one incident forever going to preclude acceptance of any finding of theirs that includes "pilot error" as a factor?
Also, you're a pilot, I'm an engineer - but I wouldn't trust myself to derive a correct conclusion from the raw data. What makes you so sure you could?
DozyWannabe, if tomorrow I don’t make it, I want my wife, children, best friends, to have a look TOO at MY data when my authorities will set their dirty move (http://www.lexpress.fr/actualite/indiscrets/rio-paris-a-qui-la-faute_1000516.html).

PS : Thanks for your friendly pm - I will answer or comment when time permits.


EDIT : Much better view on the scale on this image proposed by aguadalte
Punta Cana, Dominican Republic: Orbest Airbus A330-200 CS-TRA - detail of horizontal stabilizer - Punta Cana International Airport - PUJ / MDPC - photo by M.Torres - Travel-Images.com (http://www.travel-images.com/photo-dominican214.html)

http://i25.servimg.com/u/f25/11/75/17/84/af447_13.jpg (http://www.servimg.com/image_preview.php?i=98&u=11751784)

[JenCluse]

I move that jpete, Probationary PPRuNer, be promoted immediately to Newby 1st class, with Elephant stamp and Star. Seconders? Ah. I see Thread Boss has already done that in effect.

Thank you jpete, for a wonderful tool. I hadn’t grasped the full power of Googles’ ‘site:’ prefix, and was only using it simplisticly to search my counties’ sites.
____

Thanks also all you others, who responded to my ISUS search query (ICUS sic). I drove the first steam age A320s for a bare 100 hrs or so, and I see the pneumatics are as always. However, I was mostly interested to discover the degree of autonomy that the new digital S/B 'A/H' has.

I twice suffered loss of all driven flight instruments, (not on the 'bus!) once a total shut-down for 8-10 mins, and the other more insidiously with both the ADI & HSI freezing in place with no warnings at all, both in dark-night. I therefore used to be somewhat paranoid about finding out how long each type that I flew could last on battery power, after loss of all electrics (had once), and would develop an Apollo 13 like power-down procedure, until all that was left was the 3A (IIRC) S/B AH, a torch, and the SB pneumatic gauges, radio sched call each 30 mins, from the one (motorbike rated) 20 A.H. NiCad battery in the belly, just in case I lost all elects again. With 3 hours max to the nearest available, highly desirable. I also frequently practiced hand flying on cruise to 'freshen up,’ and was always very manually current.

This preamble is to say that, given the 'bus’s zero stick-feedback, zero throttle-feedback, and non-amplified vertical attitude tape response (as, say, the Concord had in cruise), the thought of hand flying the A320 on cruise was a bridge too far for me. To return to the thread: Even given that hand-flying is actually only providing joystick like CWS inputs in Airbus's FBW logic, and far too coarse for the delicate inputs needed at high altitude, I cannot see how any line pilot, and possible even any test-rated pilot could hand fly this extreme initial jet upset, the cause of which remains a mystery I understand.

My interest in the ISIS was actually related to the soundness of the attitude display, whether it’s attitude reference is integrally sourced as the original A320s were, or whether the ISIS is dependent on external reference in any way. The manufacturers specifications make no mention of this. If the ISIS *is totally independent in attitude, it does seem odd that there appears to have been no flight guidance reference made to it, at least discussed. In such an upset, I would have been glued to it, regardless of all other indications.

But then, who can discuss realistically the forces in the interior of a Cb, except those who have done so and survived, usually with hairy stories?

In regard to PJ2’s comment: “..never saw the "slight increase"!” (a pitch-up) PJ2, it certainly existed in the climb. In every previous type I flew when possible I used to hand-fly to TOC. Constant refresh of cp movement/mach tuck, etc. That went out on the 320, with the pitch up. It was approx 0.2 deg/sec, which required approx 10 g nose down force applied to the top curve of the side stick to neutralize. (A gentle little-finger pressure.) Somewhere on file I have copies of the two (typewritten) letters I wrote to Airbus directly, requesting an explanation for the apparent external force being applied in CWS, which could compound a situation if a pilot chose to rely on CWS in some situation. Seems it may have been the 1g at increasing altitude effect. I never did heard back from Airbus.

In the meantime I go back to exploring the implications of RetiredF4's eye opening post (his~her reference to Ostawiri & Naik’s post stall study of the NACA 44x airfoil lift curves specifically) which has chilling implications which I find quite disturbing.

In memory.

[henra]

Lets start with the 46.5°AOA of the THS (60° Aircraft AOA minus 13.5° NU trim) , which gives us a CL of 1.2 and a CD of 1.2. Lets do some ND input by trimming the THS to 0° and let´s see what it does to our numbers. We end up at 60° AOA at the THS, a CL of 1.0 and a CD of 1.6! Our upforce on the THS needed for lowering the nose has decreased and the drag has increased, our ND trim has caused the opposite of our desired outcome.

First of all: Thanks a lot for the interesting reference.

However there is one thing where I have to admit I disagree a little bit with your conclusion.
The drag along the flightpath would apply a Nose down moment on the Aircraft of (Rho/2 Cd A V^2 * sin (Alpha)* distance from center of rotation (sin Alpha ~0,87 @60°) while lift would act with the cosine, being 0,5 @60°.
Overall moment at 60° would be 0,5 * 1,0 + 0,87 *1,6 = 1,885 times the constant rest.
at 46;5° it would be 0,72 * 1,2 + 0,69 *1,2 = 1,698 times the rest.

Therefore the increase in drag would have more impact than the reduction in Lift in the given example.
The overall ND moment would increase in your given example.
As I already mentioned: at these AoA Drag acting on the tail is a good thing. It supports Nose down moment.

[RetiredF4]

Re Henra

Lets face the fact, that finally it´is the sum of all moments, which lead to the desired outcome. Wether the CD produces some ND input and CL produces some NU input or vice versa, the effectiveness of it depends on the sum of the individual forces. And as stated before, the CD and CL change of the wing has its share in the equation as well.

If we agree, that it never was the designed specification of this THS to get the AC out of a 60° AOA but to keep the aircraft in balance under the normal flight conditions by creating less or more ANU forces (curvature to produce downward lift on the THS), then the regime we are talking about here is as distant as the moon from the sun.

To get the desired AND vector, the THS airfoil has to produce lift opposite to its original design ( like if the main wing would have to produce not upward lift, but downward lift like during extreme pushover). I´m not saying that that is not possible, there are airfoils with symetrical curvature working as well by just changing the AOA of the airfoil (like the F-104 Starfighter main wing), but those designs are specifically crafted for that aircraft and its desired performance, and despite that, it had been highly succeptable to stalls and creating high amount of drag once outside desired flight envelope. I´ve lost lots of friends in accidents with this otherwise wonderful aircraft.

Another point i´m still curious, but have no expierience in it, so just correct me. I get the impression, that some posters simplify the function between lift and drag to the region of the pre stall flight envelope, and thus setting a constant between drag and lift, meaning where there is drag on an airfoil, there also has to be equivalent lift. 35 years ago i learnt in my academics classes, that if the airfoil produces lift, it also causes drag called induced drag (but not vice versa) up to max CL is reached (nearly linear function of drag and lift), but after that point drag continuous to increase while lift decreases. In our post stall discussion there are no formulas available to compute lift from drag, because there is no constant behaviour anymore outside the wind tunnel.

It is also way off reality to disregard the huge amount of drag produced by that stalled airfoil in the context of recovery efforts. To get any usefull output of the ND movement, we have to get the wind flowing over the lifting devices by increasing speed (which is the ultimate reason to decrease the AOA). When we increase the drag and thus reduce the forward velocity vector even more, we increase the descent rate by a comparable amount, thus increasing the total AOA again.

So it is no sense in discussing single points of some more or some less lift if it is not changing the overall problem to get the aircraft out of the extreme AOA condition within the available altitude.

One thing i´m sure : The simple stick forward and wait, like some are saying here, would only have worked up to a defined value of AOA and speed, and i see that value of AOA well below 45°. The discussion what the crew did do wrong in recovering the aircraft once established in the descent (estimnated FL 300 downward ) is IMHO useless, because there was no "conventional method" available which would have brought AF447 out of this position.

The problem lies in the beginning with the extreme climb rate, the chance to recover the following upset ended somewhere between FL 370 and FL 300.

[HazelNuts39]

@RetiredF4 and Henra;

Perhaps you should have another look at Figure 6 in the document linked by
zumBeispiel (http://www.pprune.org/tech-log/449639-af-447-search-resume-part2-50.html#post6440142) six weeks ago. Admittedly that is not an A330, but what makes it different?

we have to get the wind flowing over the lifting devices by increasing speed (which is the ultimate reason to decrease the AOA).First you have to reduce AoA to unstall the wing, then you need to increase speed to not stall again.

[henra]

In our post stall discussion there are no formulas available to compute lift from drag, because there is no constant behaviour anymore outside the wind tunnel.



indeed there is no way to calculate Lift from drag as Lift is not a function of drag.
What we have are merely datapoints from experiments on given airfoils.

That is everything we have. That said I hope I wasn't unclear when referring to drag acting as lift. That does not mean a mixing of Cl and Cd. They are orthogonal to each other. The trick is the coordinate system. When the aircraft has a high AoA there is an angle between both Cl an Cd and the Airframe itself as Cl and Cd are measured along the stream of air, i.e.the flight path. Wwhereas the moment and its related arm act on the airframe which is angled against these forces.
Lets assume the plane is falling at 90°. In that case it would only be the drag component of the Tail that would lead to ND forces. The Cl compnent would work orthoginally to the flight path, i.e. forward, which would be Nulll in that case.
I hope it gets clearer now what I want to explain.

[Graybeard]

In all these pages, I have seen no mention of the basic airfoil of the A330. It's probably some version of supercritical, correct? What is its stall behavior?

[TyroPicard]

RetiredF4
One thing i´m sure : The simple stick forward and wait, like some are saying here, would only have worked up to a defined value of AOA and speed, and i see that value of AOA well below 45°. The discussion what the crew did do wrong in recovering the aircraft once established in the descent (estimnated FL 300 downward ) is IMHO useless, because there was no "conventional method" available which would have brought AF447 out of this position.
I don't think this is correct. Sustained full forward sidestick is effectively a demand for minus 1g. If the initial elevator movement did not achieve this the THS would immediately move in a nose-down trim sense to assist - if necessary as far as the stop. The difficulty lies in recognising the requirement for the forward sidestick.

[rudderrudderrat]

Hi Tyro,

"The maximum elevator deflection is 30° nose up, and 17° nose down. The maximum THS deflection is 13.5° nose up, and 4° nose down."

I agree with RetiredF4. If the side stick had been held fully forward, the stab trim could have changed the THS's angle of attack by about 17.5 degs. Unfortunately, all that would have achieved would have been to INCREASE its angle of attack. The Stab is permanently stalled in their situation.

[HazelNuts39]

Greybeard;

On these airplanes the airfoil section (thickness, camber and 'incidence') varies considerably from wing root to tip. The stall behaviour is not simply a function of the airfoils, but involves all of the wing and even all of the airplane including the fuselage.

[Machinbird]

by Rudderrat

"The maximum elevator deflection is 30° nose up, and 17° nose down. The maximum THS deflection is 13.5° nose up, and 4° nose down."

I agree with RetiredF4. If the side stick had been held fully forward, the stab trim could have changed the THS's angle of attack by about 17.5 degs. Unfortunately, all that would have achieved would have been to INCREASE its angle of attack. The Stab is permanently stalled in their situation.
As others have mentioned, what matters is the NET pitching moment. If there is enough pitching moment nose down, rotation nose down will occur.
There is Lift and Drag after stall. Lift does not suddenly dissapear post stall but follows a variable path toward zero lift at ~ 90 degrees angle of attack (and becomes 100% drag there)
The path depends on:

Angle of attack-The Primary variable
Aspect Ratio,
Airfoil thickness ratio,
Reynolds Number that the wing is operating at.

Take some time to understand the following document regarding post stall aerodynamics.
http://home.comcast.net/~shademaker/pdf/Ostowari&Naik_PostStall44XX.pdf

[RR_NDB]

Hi,

Probationary PPRuNer, be promoted immediately (http://www.pprune.org/tech-log/454653-af-447-thread-no-4-a-7.html#post6520852) to Newby 1st class

(*) jpete 1st post (http://www.pprune.org/tech-log/454653-af-447-thread-no-4-a-6.html#post6519852)

Perfect! And with the Simple Characteristics defended by Kelly Johnson, Einstein, St. Exupéry, Occam, Leonardo Da Vinci and (some) other designers / engineers.
(http://en.wikipedia.org/wiki/KISS_principle)

Machaca told me (and he was right) to look all posts before the ones i made and was making on a/c recorders. And now i will be able to post, checking not only here (AF447) but in many other PPRuNe threads.

Albert Einstein's: "everything should be made as simple as possible, but no simpler."

Leonardo Da Vinci's: "Simplicity is the ultimate sophistication."

Antoine de Saint Exupéry's: "It seems that perfection is reached not when there is nothing left to add, but when there is nothing left to take away."

Harry Hawker's: "Simplify, and add lightness."

William of Ockham: "the simplest explanation is most likely the correct one."

Bertrand Russell: Whenever possible, substitute constructions out of known entities for inferences to unknown entities."

K.I.S.S. The acronym was first coined by Kelly Johnson, lead engineer at the Lockheed Skunk Works (creators of the Lockheed U-2 and SR-71 Blackbird spy planes, among many others).

When possible we must avoid K.I.C.S. designs. There are (many) examples...

[rudderrudderrat]

Hi Machinbird,

Thanks for the link - but I think you mean "Lift does not suddenly dissapear post stall but follows a variable path toward zero lift at ~ 90 degrees angle of attack."

Previous posters have suggested that if the stab angle of attack is about 45 degs (around the second lower peak of CL), then changing the angle would merely reduces the lift on the tail (hence aircraft nose would rise).

[RetiredF4]

@Machinebird
Take some time to understand the following document regarding post stall aerodynamics.
Comcast (http://home.comcast.net/~shademaker/...tStall44XX.pdf)

Machinebird, the link is known and my statement concerning the effectiveness of the THS for ND is based on it. .

See my post No. 35 and PickyPerkins post No. 100 in this thread.

Another thought to relation of THS and wing:
In the post-stall the main wing produce some ANU lift, the tail produces some AND lift. Due to the more sophisticated design of the wing it could be possible that the wing produces more ANU lift than the THS could counter with AND lift. Remember, in normal operation the THS only produces ANU lift, therfore the camber is on the downside of the THS.

And lets look at the drag, which is working as a turning force around the lateral axis: The drag of the forward part of the fuselage and of the forward parts of the wings woking ANU night well be greater than the AND force by the fuselage aft of the wings and the THS.

[RR_NDB]

Hi,

even all of the airplane including the fuselage (http://www.pprune.org/tech-log/454653-af-447-thread-no-4-a-8.html#post6521591)

The "rectangular belly" is perhaps an "important part" of this?

[henra]

The drag of the forward part of the fuselage and of the forward parts of the wings woking ANU night well be greater than the AND force by the fuselage aft of the wings and the THS.

Fully agreed. That is the real unknown when it comes to the net effect at 60°.
It might indeed be the case that the forward fuselage creates equal or higher lift than the tail, although at least for the A330-200 with the relatively short forward fuselage I doubt it.
edit:
Where I do not fully agree is the hypothesis taht the main wing will create a net NU moment. In non stalled condition the Center of lift is close to the quarter chord line. At fully stalled condition the Center of lift will move to the mid chord line. (It is only in a transient phase when the tips start to stall and the root is not yet stalled where you might encounter a NU effect. But at 50°+ AoA that is defintiely not the case any more)
/edit

This assumption is to some extent supported by the cm curve in the link as it suggests a net reduciton of cm with increasing AoA, i.e. a ND tendency.

If this was the case with AF447 we will probably never know as it appears that the Pilots unfortunately never really tried it.

[henra]

Hi Tyro,
I agree with RetiredF4. If the side stick had been held fully forward, the stab trim could have changed the THS's angle of attack by about 17.5 degs. Unfortunately, all that would have achieved would have been to INCREASE its angle of attack.

hmm, somehow I'm completely at a loss how you know that for sure !?

Would you mind sharing your source ?

[rudderrudderrat]

Hi henra

post 147 first line is copied from FCOM.

[FE Hoppy]

At 2 h 12 min 02, the PF said "I don’t have any more indications", and the PNF said "we have no valid indications". At that moment, the thrust levers were in the IDLE detent and the engines’ N1’s were at 55%. Around fifteen seconds later, the PF made pitch-down inputs. In the following moments, the angle of attack decreased, the speeds became valid again and the stall warning sounded again.

The elevator was effective in the correct sense as reported by BEA. It wasn't used enough to recover.

[DozyWannabe]

DozyWannabe, if tomorrow I don’t make it, I want my wife, children, best friends, to have a look TOO at MY data when my authorities will set their dirty move (http://www.lexpress.fr/actualite/indiscrets/rio-paris-a-qui-la-faute_1000516.html).

Hi CONF,

If the information is accurate, I wouldn't call it a "dirty move". It may not be pleasant to hear, but according to what the minister has been told, the pilots were supposedly trained to handle pitot icing/blockage incidents, mode reversion and all that comes inbetween in this incident.

Note that he's not saying the aircraft is blameless - he can't because the information is out that the pitots were probably the initiating factor. The BEA were the ones who released that information publicly when their initial interim report was released a couple of years back - if they really wanted to "stitch up" the pilots then they would never have released that information.

As I said earlier - the BEA can be said to have handled a single AF/Airbus incident "badly", though personally speaking I wouldn't have wanted any part in that political minefield. That incident was 23 years ago and they're still dealing with the damage that did their reputation today. It would be utter madness for them to try to do so again, so I say again, give them a chance.

[wallybird7]

DozyWannabee

“I ask again - why is it you're asking this of the BEA and no other accident investigation bureau/institution? Is one incident forever going to preclude acceptance of any finding of theirs that includes "pilot error" as a factor?

Also, you're a pilot, I'm an engineer - but I wouldn't trust myself to derive a correct conclusion from the raw data. What makes you so sure you could? “

Just for the record, the mission of EVERY investigating body, is to gather the facts and make a determination of Probable Cause and to recommend a fix, but NOT to affix blame, fault, or liability, which is left to legal venues.

As a pilot all I want is all the given facts and then I can make my own conclusion regarding what to do about it. In this case avoid thunderstorms at all costs. Period.

It is way too early to start figuring out how to recover from an out of control deep stall, when in fact they won’t even allow pilots to practise “hand flying” in the simulator at altitude, let alone in the airplane.

No one knows what the plane is going to do.

[jcjeant]

Hi,

This forum is titled "Tech Log" as it is normal that all the technical assumptions are discussed here.
Many hypothesis have been advanced .. so that even someone qualified can be confused by their number and diversity.
There's just a fact that caused all these technical discussions .. it's the pilot performance in the first seconds after the autopilot go off
This in my opinion is the most important fact
Does this fact is explained by all the technical assumptions in posts?
I doubt it strongly.

As I said earlier - the BEA can be said to have handled a single AF/Airbus incident "badly", though personally speaking I wouldn't have wanted any part in that political minefield. That incident was 23 years ago and they're still dealing with the damage that did their reputation today. It would be utter madness for them to try to do so again, so I say again, give them a chance.Just a tought
They have this chance .. but they already begin to ruin it by the release of their last note.
And this chance was again badly wounded by the statement of a member of the french political body (Mr Mariani directly involved in the following of AF447 cause his position in the government)

[DozyWannabe]

Just for the record, the mission of EVERY investigating body, is to gather the facts and make a determination of Probable Cause and to recommend a fix, but NOT to affix blame, fault, or liability, which is left to legal venues.

Wallybird, you're right!

However, I don't see anywhere in the post you quoted where I referred to "blame". I was very careful in the word I chose, and said "finding", using the language common to the AAIB, NTSB et al.

As a pilot all I want is all the given facts and then I can make my own conclusion regarding what to do about it. In this case avoid thunderstorms at all costs. Period.

Well, glad that's all sorted then - back home, lads!

Seriously though, when they've got facts worthy of release - I suspect they'll release them. It has long been common practice in aviation accident investigation to cover as much ground as they can so that as much can be learned as possible - that's the reason it tends to take a while.

For example, the AAIB's final report on the BA038 incident took 2 years and one month to produce. In that case, they had the entire airframe more-or-less intact, both flight recorders and the QAR to hand, as well as the entire crew and passenger list available for interview - but the lack of evidence regarding icing on the heat exchangers meant that there was a significant hole in the available data and thus there was a seemingly "inexplicable" element to the incident. The BEA produced their initial interim report on AF447 with small fragments of wreckage, a set of ACARS messages and no flight recorder data. Now that they have more wreckage and the flight recorders they are probably going to have to do a lot of rewriting of their assumptions, as well as a whole battery of human factors testing in relation to what they've learned from the CVR.

As an aside, the discussion on BA038 on here went back and forth for those entire two years, and speaking for myself I learned a lot of new and surprising things from that discussion - but at no point did any contributor demand that the raw data traces were released or chide the AAIB for not doing so. Now was that because it was the AAIB doing the investigation and not the BEA? Was it because the aircraft involved was a Boeing 777 and not a member of the Airbus FBW family? Or was it because there's an understanding that these things take time and should not be rushed. Being on here for as long as I have, I suspect it was a combination of all three.

Finally, there were theories doing the rounds (that ended up in the BBC/NOVA "Lost" documentary on AF447 back in 2009) that the thunderstorm cells were obscured on radar by a smaller cell in front of the cell they entered - what do you do in that situation if avoidance beomes impossible?

when in fact they won’t even allow pilots to practise “hand flying” in the simulator at altitude, let alone in the airplane.

I haven't seen any evidence of this regarding Air France, could you point me in the direction of some?

Just a tought
They have this chance .. but they already begin to ruined it by the release of their last note.
And this chance was again badly wounded by the statement of a member of the french political body (Mr Mariani directly involved in the following of AF447 cause his position in the government)

"Badly wounded" how? They released all the information they could determine after just a few days looking at the data - just as the AAIB did with BA038. M. Mariani may or may not have been speaking with the blessing of the BEA (it wouldn't be the first time politicians have gone over the heads of their accident investigators - and it's certainly not something peculiar to France).

The problem is that there are some on here that remain wedded to the idea that the BEA will always try to find against the pilot before anything else - something that is not supported by any evidence gathered over the last 20 years at least. Yes, their paycheques are ultimately signed by an entity that has a controlling stake in the manufacturer and airline involved, but the same was true to some degree of the old UK AIB (in the days when the UK made airframes and BA was state-owned), as well as the NTSB prior to the mid '70s (Boeing/McD-Douglas and Lockheed were privately owned, but heavily dependent on government/military contracts).

[hetfield]

Finally, there were theories doing the rounds (that ended up in the BBC/NOVA "Lost" documentary on AF447 back in 2009) that the thunderstorm cells were obscured on radar by a smaller cell in front of the cell they entered - what do you do in that situation if avoidance beomes impossible?Dozy, that's a very good question! Like most of us know this can be a real trap!

Regards

[Graybeard]

Dozy:Finally, there were theories doing the rounds (that ended up in the BBC/NOVA "Lost" documentary on AF447 back in 2009) that the thunderstorm cells were obscured on radar by a smaller cell in front of the cell they entered - what do you do in that situation if avoidance beomes impossible?

I just want to repeat that documentary was rubbish. Just reading the pilot's guide for the Wx radar on 447 would have given them real info about the radar capabilities. Instead, they went off on an ill-founded theory about the "storm behind the storm."

[jcjeant]

Hi,

"Badly wounded" how? They released all the information they could determine after just a few days looking at the dataI disagree about this
The note released it's just some fragments of the CVR record.
Unless the CVR had recording problems ... we can't tell the BEA released all the information provided by the CVR
This is why all those posts trying to fill the gaps open by this incomplete release.
If I quote one of your message incompletely .. you will complain of context .. ♫♪la la la♪♫
Same for a incomplete BEA release ....

[RR_NDB]

Hi,

They released all the information they could determine after just a few days looking at the data (http://www.pprune.org/tech-log/454653-af-447-thread-no-4-a-9.html#post6521881)

They had time to carefully select and release the "right information" taking into account all all possible implications. Trying to secure a big and comfortable "room to maneuver".

From 2 h 10 min 05 , the autopilot then auto-thrust disengaged and the PF said "I have the controls". The airplane began to roll to the right and the PF made a left nose-up input....The airplane’s pitch attitude increased progressively beyond 10 degrees and the plane started to climb....The PF made nose-down control inputs and alternately left and right roll inputs.The vertical speed, which had reached 7,000 ft/min, dropped to 700 ft/min and the roll varied between 12 degrees right and 10 degrees left. At 2 h 10 min 51, the stall warning was triggered again. The thrust levers were positioned in the TO/GA detent and the PF maintained nose-up inputs...The PF continued to make nose-up inputs. The airplane’s altitude reached its maximum of about 38,000 ft, its pitch attitude and angle of attack being 16 degrees...The PF made an input on the sidestick to the left and nose-up stops, which lasted about 30 seconds..."

Their text, lacking important details, seems typical of PR instructed to deliver "the right information" for that moment making easier for BEA to "navigate" the "turbulent environment" (http://www.pprune.org/tech-log/449639-af-447-search-resume-part2-45.html#post6435767). The lack of important details does not allow a more precise analysis. And this also may prove later to be against BEA mission because seems showing a preliminary bias with the emphasis on "NU from PF".

And in emphasizing persistent NU commands seems as to considering as the "major cause", coherent to LF leak (http://www.lefigaro.fr/actualite-france/2011/05/17/01016-20110517ARTFIG00426-af-447-la-piste-d-une-erreur-de-l-equipage-air-france.php), deviating the focus from plane malfunction(s). This was IMO a tentative to gradually prepare public opinion and timely to the big "interests involved in Paris Air Show". The seriousness because the LF leak was too early (IIRC in the Monday subsequent the first analysis during the weekend). Probably an opportunistic attitude from one(s) representing the huge interests" behind. With risk (and probably, intentions) in creating a "frame" for BEA "output".

Human machine interface issues, "software issues", etc. would be considered "machine faults"?

System processing of "ridiculous Pitot redundancy" has chances to be considered "machine fault"?

All this are "against crew" and makes easier the work of the "high rocks" at FR government and it´s personnel (including BEA top managers).

Most of us yet now could imagine the Report and it´s recommendations. Certainly BEA is "ahead of the schedule" on that.

They seems to be navigating very well in this turbulent weeks since wreckage location. (http://www.pprune.org/tech-log/449639-af-447-search-resume-part2-76.html#post6454750)

[DozyWannabe]

I just want to repeat that documentary was rubbish. Just reading the pilot's guide for the Wx radar on 447 would have given them real info about the radar capabilities. Instead, they went off on an ill-founded theory about the "storm behind the storm."

Read what I said carefully. I'm not saying that's what happened to AF447, but if that situation is a possibility then "Avoid all thunderstorms - period" becomes insufficient advice. What do you do about the thunderstorm you can't see?

I disagree about this
The note released it's just some fragments of the CVR record.
Unless the CVR had recording problems ... we can't tell the BEA released all the information provided by the CVR
This is why all those posts trying to fill the gaps open by this incomplete release.
If I quote one of your message incompletely .. you will complain of context .. ♫♪la la la♪♫
Same for a incomplete BEA release ....

Not really - a full CVR transcript is also out of context without an FDR record to explain what was happening in conjunction with the statements recorded by the CVR. What they've released is what they could determine by matching up portions of the CVR data with what they have extracted (and processed) from the FDR so far.

Case in point - I remember an accident described in the "Black Box" book* that accompanied the C4 series of the same name in 1996. The investigators in that case released CVR data where the pilots were complaining of being tired - this got picked up by the press and the accident was written up as fatigue-induced pilot error. Another angle they were working at the time was that the altimeter fitted to the aircraft in question was known to occasionally misread by up to 1,000 ft and there was plenty of evidence that this had indeed happened. As such, they recommended replacing that model of altimeter, but because public perception had already accepted the "fatigue" explanation, the regulator wasn't interested.

This is about the fifth time I've asked this question and I've yet to get an answer - why are there demands for raw data being made of the BEA in this instance from some quarters on this forum, when to my memory no such demands have ever been made of the NTSB or AAIB?

@RR_NDB : I note that when journalistic "tittle-tattle", as it applies to piloting and aircraft operations, is brought up on here it tends to be treated with the scorn it deserves, except in very exceptional circumstances. Why should press hearsay about the motivations of the BEA and the French government be any different? [EDIT : Just had a quick squiz on Google, and as far as I can see the "Paris Air Show" connection was made at the end of a paragraph in a Flight International article, and all that said was that the timing of the discovery "could have come out of a film script" - no connection to the possible release of information to protect Airbus or Air France has been made outside of the speculation on this forum.]

* - The book was heavily oversimplified and wasn't really all that - but my local Waterstone's was getting rid of it for 2 quid...

[Turbine D]

RR_NDB

Perhaps a more appropriate saying:

Do not be desirous of having things done quickly. Do not look at small advantages. Desire to have things done quickly prevents their being done thoroughly. Looking at small advantages prevents great affairs from being accomplished. - Confucius

[OK465]

This was actually predated by the noted Far Eastern scholar, Wo Phuc, who noted:

"If you want it badly, you'll get it badly."

[Graybeard]

Dozy: What do you do about the thunderstorm you can't see?

What is the context of your question?

Was it that the AF pilots didn't receive adequate training in use of their Wx radar?

Was it that AF didn't upgrade to the latest radar that has special processing for tropical storms?

Do you believe there are thunderstorms with no rainfall or ice pellets?

[DozyWannabe]

@Graybeard : The context of my question is that if enough people thought it was a possibility, even if it doesn't turn out to have much of a bearing on the case, would it not be a good idea to look into remedying the situation - whether that be via training or technology?

I've got no answers to your other questions, other than a nagging suspicion that for all our knowledge, nature will always find a way of surprising us - sometimes not in a good way.

[Turbine D]

Thanks OK465,

That is the message, very concise. :ok:

[CONF iture]

The BEA were the ones who released that information publicly when their initial interim report was released a couple of years back - if they really wanted to "stitch up" the pilots then they would never have released that information.
Did they have any choice really when ACARS + Air Caraibes report were all over the Net … FIRST.

Probably an opportunistic attitude from one(s) representing the huge interests" behind. With risk (and probably, intentions) in creating a "frame" for BEA "output".
Framing the mind, correct, but better known as 'communication' or even better, 'information' …

http://i25.servimg.com/u/f25/11/75/17/84/af447_10.jpg (http://www.servimg.com/image_preview.php?i=95&u=11751784)

Three lies in one single image :

People will think that what the pilot was seeing on his weather radar screen was exactly that big red area.
People will think that the pilot has intentionally penetrated that same big red area.
Why are not depicted both the aircrafts that have flown the same route as AF447, in the same big red area ?

[DozyWannabe]

Did they have any choice really when ACARS + Air Caraibes report were all over the Net … FIRST.

Would they have done so if they weren't? That's a question neither you, I, nor anyone else can answer. I believe they would have (after all, it was pretty much the only information they had to go on and they had to base the first interim report on *something*). You may think differently.

Three lies in one single image :

People will think that what the pilot was seeing on his weather radar screen was exactly that big red area.
People will think that the pilot has intentionally penetrated that same big red area.
Why are not depicted both the aircrafts that have flown the same route as AF447, in the same big red area ?


That image looks a little... how should I say... "dramatic" for an accident report. It looks more like a sensationalistic press cutting to me - how are the BEA responsible for that (any more so than the investigators in other countries responsible for their press getting the wrong end of the stick)?

Let me know what you think of the stuff I PM'ed you!

[JenCluse]

To RR_NDB's (post 4/150), should be added Harry Hawker's suggestion to his designers:

"Simplify, and add lightness."

( I'll try to remember that, if I post again :~? )

[hetfield]

Three lies in one single image :

People will think that what the pilot was seeing on his weather radar screen was exactly that big red area.
People will think that the pilot has intentionally penetrated that same big red area.
Why are not depicted both the aircrafts that have flown the same route as AF447, in the same big red area ?

Wasn't there a LUFTI on same track as AF447 30 min ahead?

[BOAC]

I will repeat my opinion that the release by BEA was 'mischievous'. As I have said, they have the CVR and FDR. By limiting the CVR info they neatly 'imply' that PF pitched to 10 deg, and climbed 3000' (above optimum and possibly above max) without a single word being said by either pilot about it. Either that is the case, which I positively doubt, or there is some reason why this part of the conversation has not been released. It would have saved pages of PPRune. Without it we are all guessing in the dark

[rudderrudderrat]

Hi BOAC,

I will repeat my opinion that the release by BEA was 'mischievous'.

I don't think the BEA are "mischievous". The full official report usually takes months, sometimes years to produce. There was a thirst for a briefing as to what may have gone wrong, and the BEA released some information.

Something similar happened with the B777 at LHR 2008 pprune discussions. It's not the BEA who are mischievous - it's us - the pprune posters.

http://www.pprune.org/rumours-news/309337-aaib-initial-report-out-ba-b777-crash-lhr.html

[HazelNuts39]

I will repeat my opinion that the release by BEA was 'mischievous' ... based on the entirely arbitrary assumption that the pilots observed the departure in pitch, v/s and altitude and discussed it, and BEA withheld that crucial part of their conversation, in the knowledge that they will have to justify that decision when they publish the transcript.

As a minor detail, BEA said : "The airplane's pitch attitude increased progressively beyond 10 degrees ..." (to 13 degrees according to my 'simulation').

[Hyperveloce]

I will repeat my opinion that the release by BEA was 'mischievous'

The context of this note should be taken into account: the BEA was under pressure to anticipate their analysis of the newly recovered DFR/CVR, some results were leaked and Airbus was already claiming that the analysis of the recorded data had shown that the plane was not the problem. The BEA reacted strongly to this claim, making clear it is the only authorized entity to establish new facts about the AF 447 accident, that those leaked/unauthorized informations were not helping the families who had lost a relative, that it will soon release a first set of new facts derived from the recorders and an interim report by the end of july. A bone to gnaw, waiting for the main course.

[HazelNuts39]

That message, time stamped 2:12, was received at 2:12:51. At that point in time the airplane was already in a full stall, with all airspeeds probably misreading due to pressure errors at the pitots and at the static sources due to the high AoA. I wonder if that could tell us anything about the speed displayed on PFD2 prior to stall, but can't figure it out. Anyone?

[BOAC]

based on the entirely arbitrary assumption that the pilots observed the departure in pitch, v/s and altitude and discussed it- not 'arbitrary, HN, logical? Many people on this forum need to separate themselves from the 'journo crap baby pilot' stuff - we have here two adult, qualified, intelligent pilots trained and checked by a major airline and I find it incomprehensible that such an excursion would not have been noticed and commented on. Certainly in my experience anyway PNF would have challenged what was happening - perhaps not in yours?

[Graybeard]

Was the PNF possibly too busy/engrossed/distracted with pages of ECAM warnings and alerts to notice what the plane was doing?

[Elledan]

Was the PNF possibly too busy/engrossed/distracted with pages of ECAM warnings and alerts to notice what the plane was doing?

GB
Wouldn't be the first time that an entire crew got totally focused on a single issue, real or imaginary, and lost track of the rest of the data being reported.

Not noticing is something humans are very good at, unlike computers :)

[jcjeant]

Hi,

The context of this note should be taken into account: the BEA was under pressure to anticipate their analysis of the newly recovered DFR/CVR, some results were leaked and Airbus was already claiming that the analysis of the recorded data had shown that the plane was not the problem. The BEA reacted strongly to this claim, making clear it is the only authorized entity to establish new facts about the AF 447 accident, that those leaked/unauthorized informations were not helping the families who had lost a relative, that it will soon release a first set of new facts derived from the recorders and an interim report by the end of july. A bone to gnaw, waiting for the main course. Indeed you right.
BEA reacted strongly for support the Airbus claim
Indeed for "le citoyen" lambda people the note of the BEA (and emphasized in newspapers) the pilot error is what made fall the plane.
Even the minister Mariani confirmed this after the BEA release.
The seed is planted

[Machinbird]

This whole "pilot error" discussion puts me in mind of the deceptive term "pilot induced oscillation" which is actually an airframe problem. PIO is often known as Pilot in the Loop Oscillation or also Aircraft Pilot Coupling, but the key point is that it is more an aircraft design problem-not so much a pilot technique issue.

I can already see that the AF447 cockpit was not well designed for pilot situational awareness in the type of flight conditions encountered by the AF447 crew: No direct AOA indications, no old fashioned steam gage altimeter unwinding, no bell or clacker on the trim wheel to draw attention to its motion, too many competing alarm sounds in the cockpit.

Then there are training issues surfacing and also fundamental how do you fly the 'Bus issues that are surfacing.


We are left with three core issues relating to piloting and aircraft systems.
What did they see on their radar and how did they interpret it?
Why did they allow the aircraft to climb from Fl 350 to 380?
Why did they not succeed in recognizing/recovering from the stall?
It is up to BEA to thoroughly examine these issues. These are all very complex issues and it will not be trivial to sort out. The crew was handed a "Pop Quiz" and failed. The reasons for this failure will extend far past the crew.

[SaturnV]

Dozy, the Lufthansa preceding, and the Iberia and Air France following, on the same airway as AF447 that night 'saw' the Cbs and deviated, AF459 rather dramatically with its zig-zagging total of about 100 NM to the left and right of the track. Dispatch had sent AF447 a message regarding satellite observation of Cbs vicinity of ORARO and TASIL well before AF447 reached ORARO.

Until the full CVR is released, we won't know what the crew of AF447 saw (or failed to see) on their radar after ORARO.

[DozyWannabe]

SaturnV (liking the handle!):

Absolutely - I wasn't trying to imply one way or the other - I was just expressing the opinion that the investigation should be thorough.

jcjeant:
BEA reacted strongly for support the Airbus claim

The way I read it the BEA were actually a bit annoyed at Airbus for jumping the gun, so they released the "note" to calm the press, which naturally went into something of a feeding frenzy as soon as the recorders made it back to Paris. You *could* read it as the BEA supporting Airbus's statement, but alternatively it could simply mean that while Airbus jumped the gun, their information was solid.

[HazelNuts39]

Machinbird;

Apart from the ice particles that blocked the pitots, do you think that the weather played an important role in this accident?

[Machinbird]

One thing that has puzzled me, in the interaction between the two copilots. If the Junior one was flying and the senior one thought he was screwing up in a big way, I can see an advisory call of the problem followed by "I have the aircraft" if the senior guy still was not happy with the other man's performance. (And proper CRM says the reverse can happen too). But this apparently did not happen. Was the senior copilot not concerned with the junior one's performance? Or was he too busy with other things?

Is there something exclusively in the interaction between copilots that would act to inhibit this type of takeover of control. Perhaps Captains orders who was to fly a segment?

[JD-EE]

Was the PNF possibly too busy/engrossed/distracted with pages of ECAM warnings and alerts to notice what the plane was doing?

Or consider something simple like Pavlovian conditioning. Mr. Optimistic gets this point in message 69 thread 4 (http://www.pprune.org/tech-log/454653-af-447-thread-no-4-a-4.html#post6518361).

From 2 h 10 min 05, the autopilot then auto-thrust disengaged and the PF said "I have the controls". The airplane began to roll to the right and the PF made a left nose-up input. The stall warning sounded twice in a row.

The order is not clear. But the pilot did give a NU command. And as the plane answered the command the stall warning STOPPED. The pilot was rewarded for the NU command.

At 2 h 10 min 16 The airplane’s pitch attitude increased progressively beyond 10 degrees and the plane started to climb. The PF made nose-down control inputs and alternately left and right roll inputs. The vertical speed, which had reached 7,000 ft/min, dropped to 700 ft/min and the roll varied between 12 degrees right and 10 degrees left. The speed displayed on the left side increased sharply to 215 kt (Mach 0.68). The airplane was then at an altitude of about 37,500 ft and the recorded angle of attack was around 4 degrees

The pilot made a nose down input.

At 2 h 10 min 51 , the stall warning was triggered again.

The pilot was spanked for that nose down input. After the first "reward" the pilot repeated the "rewarded" move, NU.


At around 2 h 11 min 40 ... During the following seconds,
all of the recorded speeds became invalid and the stall warning stopped.

The pilot was again "rewarded" for NU commands.

...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 PF made an input on the sidestick to the left and nose-up stops, which lasted about 30 seconds.

By now the PF was "conditioned". He was rewarded for NU and punished for ND.

The plane turned the stall warning OFF when it had no good reason to do so. So what's this nonsense BEA has been spouting that the plane is not at fault? The plane has a critical design flaw in its software that conditions pilots to do the wrong thing even when they know better.

In the past four days worth of posts on catch-up reading I've seen nothing to change my view in this regard.

[Machinbird]

HN-39

Apart from the ice particles that blocked the pitots, do you think that the weather played an important role in this accident?


At this point, without clear cut evidence on the accelerometers of a monster updraft, I don't think so. (But it was night IFR conditions)

Those that wish to add a large turbulence/updraft element to this loss of control seem to be putting unnecesary sauce on the stew. Occam's Razor applies.

There appears to be sufficient explanation in the pitch angle achieved to account for the climb to FL380.

[DozyWannabe]

JD-EE:

You're not taking the training around loss of air data into account. One of the things quickly learned from the pitot/static failure accidents in the 1990s was that aural warnings generated were confusing, thus the correct thing to do is to disregard them and fly pitch-and-power until the speed readings become valid again. This is not something peculiar to the A330, it is common to all advanced jetliners.

If the pilot were simply reacting to those warnings, when he had visual cues in the form of the ADI in front of him and the power settings - neither of which requires air data to function, and are therefore likely to be correct - telling him that he was too nose-high for the power he had selected at that altitude, then he was either not adequately trained for this scenario, or so thrown by the sudden onset of the situation that the training went out the window.

[Zorin_75]

If the Junior one was flying and the senior one thought he was screwing up in a big way, I can see an advisory call of the problem followed by "I have the aircraft" if the senior guy still was not happy with the other man's performance. (And proper CRM says the reverse can happen too). But this apparently did not happen. Was the senior copilot not concerned with the junior one's performance? Or was he too busy with other things?
I think there's still no evidence against the more plausible explanation that the senior FO was PF in RHS?

[HazelNuts39]

The pilot made a nose down input.That nose-down command between 2:10:16 and 2:10:50 reduced the vertical speed from 7000 fpm to 700 fpm and AoA to (less than - in my simulation) 4 degrees (reducing v/s implies less than 1g, hence reduced AoA) - no stall warning at that point. Then a nose-up command must have followed calling for about 1,1 g and an AoA that exceeded the stall warning threshold, and ultimately the stall AoA. The 1,1 g increased the vertical speed from 700 fpm (to about 2500 fpm in my simulation). Look at BEA's "3D view" for the partial level-off followed by a steepened final climb just before the apogee. The airplane stalled before the apogee.

[infrequentflyer789]

jcjeant:


The way I read it the BEA were actually a bit annoyed at Airbus for jumping the gun, so they released the "note" to calm the press, which naturally went into something of a feeding frenzy as soon as the recorders made it back to Paris. You *could* read it as the BEA supporting Airbus's statement, but alternatively it could simply mean that while Airbus jumped the gun, their information was solid.

Can someone provide reference where Airbus actually said the plane was not at fault ?

All I have / recall is Airbus release (I believe approved by BEA) that no action was required for 330 operators yet - on the basis of analysis of the data at that point. That got turned by the media into "plane not at fault", along with other leaks leading to headlines like "no pilot in the cockpit" and other idiocy. BEA release was to stamp on the media leak/bulls**t fest - which it laregly did.

AFAICS neither Airbus or BEA can possibly come out and say the plane wasn't at fault - because the pitots (brand / model) were (and still are) held to be at fault.

Airbus might not have built the pitots, but they approved them for the a/c (and maybe because they are French...), so they are on the hook. The fact that they advised AF to replace them 2yrs before the crash, and AF blocked it, may get Airbus back off the hook (certainly doesn't make AF look good) but the plane is clearly still part of the cause. And that is before we even get into any possible human factors issues in the control system design, warnings systems, THS etc.

[infrequentflyer789]

That nose-down command between 2:10:16 and 2:10:50 reduced the vertical speed from 7000 fpm to 700 fpm and AoA to (less than - in my simulation) 4 degrees (reducing v/s implies less than 1g, hence reduced AoA) - no stall warning at that point. Then a nose-up command must have followed calling for about 1,1 g and an AoA that exceeded the stall warning threshold, and ultimately the stall AoA. The 1,1 g increased the vertical speed from 700 fpm (to about 2500 fpm in my simulation). Look at BEA's "3D view" for the partial level-off followed by a steepened final climb just before the apogee. The airplane stalled before the apogee.



The BEA text, and the "3D view" graphic, seems to me to indicate that stall warning was before the second nose-up input and ascent. It is almost as if PF reacts with back stick every time stall warning goes off (even before it starts doing strange things below 60kts). We'll find out, hopefully, if that is really the case when we get more data in the next report.

[GarageYears]

By now the PF was "conditioned". He was rewarded for NU and punished for ND.

The frigging plane turned the stall warning OFF when it had no good reason to do so. So what's this nonsense BEA has been spouting that the plane is not at fault? The plane has a critical design flaw in its software that conditions pilots to do the wrong thing even when they know better.If we were discussing training my dog to roll-over, then perhaps I'd agree, but we're not. All the occupants of the cockpit that night had hundreds of hours of experience and many 10's, if not hundreds of sim hours, so trying this to simplify this to a game of action and reward really seems crass.

UAS is known to be a situation where alarms, alerts and warnings can be misleading - training dictates fly pitch and power until speed indications recover. Which in flight, unless I'm seriously off-target, they have proven to always do (excepting situations were pitots have been blocked due to human error or bees - never did like bees).

I'm sure that the BEA and Airbus will be looking at the stall warning behavior very closely, but let's not loose sight of a couple of key issues that are VERY UNLIKELY to be the aircraft's fault:

1) Flying THROUGH the CB area was the initiating and fundamental cause. All other flights on the same airway track deviated. Had they done the same we would not be here.

2) The reaction to the stall warning was NOT the initiating problem. The problem was how they arrived at the stall condition in the first place, i.e. pitch and power at the loss of airspeed indication would not lead to this.

Therefore I would suggest that the real work needs to be focused on why they elected to NOT deviate. Everything else falls out in turn.

I think it fair to say that 99.9% of pilots would claim to have NEVER flown through a CB - because it is known to be a very, very bad thing to try. There are a few stories around from those few that made that mistake and sadly the majority end badly.

The fact that the pitots iced in this situation was a prior issue, one that Airbus had already issued a warning, that for whatever reason AF was slow to action. Whether the appropriate regulatory authorities should set in when there is a potential safety of flight concern, is another topic worth discussion.

[DozyWannabe]

I think it fair to say that 99.9% of pilots would claim to have NEVER flown through a CB - because it is known to be a very, very bad thing to try. There are a few stories around from those few that made that mistake and sadly the majority end badly.

I think this was the one that got a lot of people's attention:

Southern Airways Flight 242 - Wikipedia, the free encyclopedia (http://en.wikipedia.org/wiki/Southern_Airways_Flight_242)

[OK465]

" Before the battle of Verdun, our activity was disturbed by frequent thunder-storms.
Nothing is more disagreeable for flying men than to have to fly through a thunderstorm. During the Battle at Somme, a whole English Flying Squadron was forced down behind our lines and became our prisoners of war, because they had all been surprised by a thunderstorm.
I had never yet made an attempt to get through thunder clouds, however I could not suppress my desire ' to make the experiment.'
During one whole day, thunder was in the air. But in order to look after various things, I had flown over to the nearby fortress of Metz.
I had an adventure during my return.
After visiting the aerodrome of Metz, I had intended to return to my own base, when an approaching thunderstorm became noticeable. As it approached from the north, its vertical clouds looked like a gigantic . . pitch-black wall.
Old, experienced pilots, there, urged me not to fly. However, I had promised to return to my base. I should have considered myself a coward if I had failed to come back because of a silly thunderstorm.
Therefore, I meant to try.
I was in the air when the rain began falling. I had to throw away my goggles, otherwise I should not have seen anything. The trouble was that I had to travel over the mountains of the Moselle, where the thunderstorm was now raging.
And as I rapidly approached the black cloud which reached down to the earth, I said to myself that probably I should be lucky to get through it.
As I flew at the lowest possible altitude, I was compelled absolutely to leap over
houses and trees with my machine.
Very soon . . I no longer knew where I was. The gale seized my machine as if it were a piece of paper and drove it along. My heart sank within me. I could not land among those hills.
I was compelled to go on.
I was surrounded by an inky blackness. Beneath me the trees bent down in the gale. Suddenly, I saw right in front of me a wooded hill. I could not avoid it. I was able to fly only in a straight line. My Albatross managed to avoid its trees. And now I had to quickly avoid every obstacle that I encountered.
My flight became a jumping competition.
Purely and simply. I had to jump over trees, villages, spires and steeples, for I had to keep within a few yards of the ground . . otherwise I should have seen nothing at all [ in the dark, obscuring rain and violence. ]
The lightning was playing around me. At that time I did not yet know that lightning cannot touch flying machines. I felt certain of my death for it seemed to me inevitable that the gale would throw me at any moment into a village or a forest. Had the motor stopped working I should have been done for.
Suddenly, I saw that on the horizon the darkness had become less thick. The thunderstorm had passed, over there. If I were able to get that far, I would be saved.
Concentrating all my energy, I steered towards the light. Suddenly I got out of the thundercloud. The rain was still falling in torrents, but still I felt . . saved.
In pouring rain, I landed at my aerodrome. Everyone had been waiting for me. Metz had reported my start and had told them that I had been swallowed up by a thunder cloud. Withstanding the dangers during my flight, I had experienced glorious moments . . and I now realize that it was all very beautiful. But I shall never again fly through a thunderstorm . .
. . unless the Fatherland should demand it. "
Source : Manfred von Richthofen's letters to his family

I think he hand flew the whole way.

[HazelNuts39]

The BEA text, and the "3D view" graphic, seems to me to indicate that stall warning was before the second nose-up input and ascent.The occurrence of stall warning indicates that AoA was increasing through the threshold. The BEA text says it was six degree at threshold, and "continued to increase". What caused that increase, if not a nose-up input? BEA again: "About 15 seconds later, (...) the PF continued to make nose-up inputs".

Incidentally, you must have unbelievably sharp eyes if you can discern s/w in "3D view".

At M.68/215kCAS, FL375, AoA=4 degrees, the load factor is 0.8. It must be greater than one to increase vertical speed.

[DJ77]

Those that wish to add a large turbulence/updraft element to this loss of control seem to be putting unnecesary sauce on the stew. Occam's Razor applies.


Occam's Razor, yes and also simple physics.
The potential/kinetic energy transfert calculation shows that at 37500 ft TAS should be 413 kt @ Std + 10 assuming constant thrust and drag. The BEA reports CAS 215 kt (M 0.68) at this altitude which gives 399 kt TAS @Std + 10.
Factoring decreasing thrust with increasing altitude and a bit of extra drag due to manoeuvring should easily account for the 14 kt discrepancy.

In no way the a/c gained energy as would have been the case in an updraft.

[infrequentflyer789]

The occurrence of stall warning indicates that AoA was increasing through the threshold. The BEA text says it was six degree at threshold, and "continued to increase". What caused that increase, if not a nose-up input? BEA again: "About 15 seconds later, (...) the PF continued to make nose-up inputs".


I was assuming that the nose-down inputs didn't completely stop the climb, therefore the a/c is still trading speed for height (albeit at a slower rate) and thus increasing AoA.


Incidentally, you must have unbelievably sharp eyes if you can discern s/w in "3D view".


Was going by the "5" marker which the text states as s/w point. Looks to be between the climbs - but I admit there's not many pixels in it.

I should have been clearer that my post was merely conjecture, and the info we have at this point form BEA doesn't give sufficient detail or clarity to confirm it.

[HazelNuts39]

Was going by the "5" marker which the text states as s/w point. Agreed, thanks for clarifying that. The next sentence says that at that point - "the PF maintained nose-up input", which I read as saying that the input commenced earlier.

Otherwise, I don't see how the airplane arrives 15 seconds later at FL380 with zero v/s.

[CONF iture]

That image looks a little... how should I say... "dramatic" for an accident report. It looks more like a sensationalistic press cutting to me - how are the BEA responsible for that (any more so than the investigators in other countries responsible for their press getting the wrong end of the stick)?
Problem, in their animation, BEA also 'forgot' to represent those 2 flights which adventured 'in the middle of the storm' … Are BEA and mass media so different …

Let me know what you think of the stuff I PM'ed you!
There is absolutely matter for constructive discussion on a dedicated thread to Habsheim.
You could be surprised I don’t actually disagree on everything you write …


NAV ADR DISAGREE
That message, time stamped 2:12, was received at 2:12:51. At that point in time the airplane was already in a full stall, with all airspeeds probably misreading due to pressure errors at the pitots and at the static sources due to the high AoA. I wonder if that could tell us anything about the speed displayed on PFD2 prior to stall, but can't figure it out. Anyone?

If I get you right, you would see the NAV ADR DISAGREE as a consequence of the stall rather than due to the pitot probes icing.
PFD2 and ISIS speeds may have agree for a while …
BEA : How long is it "then a few moments later" ?
275kt at FL350 is not slow either – How far is it from Vmax ?
What was the reason for AP disconnect then ?
ALT LAW maybe came 'then a few moments later' … until then protections were still avail.
Also, if 2 speeds sharp fall to 60kt one after the other … don’t they agree again ?

Svarin, what’s your take on this ?

Full FDR data … anyone ?


The recorded parameters show a sharp fall from about 275 kt to 60 kt in the speed displayed on the left primary flight display (PFD), then a few moments later in the speed displayed on the integrated standby instrument system (ISIS).

[Machinbird]

Rudderrat & Retired F-4
From the BEA Note:

At 2 h 12 min 02, the PF said "I don’t have any more indications", and the PNF said "we have

no valid indications".
At that moment, the thrust levers were in the IDLE detent and the
engines’ N1’s were at 55%. Around fifteen seconds later, the PF made pitch-down inputs. In
the following moments, the angle of attack decreased, the speeds became valid again and the
stall warning sounded again.

This tells me that the crew performed an "experiment" that indicates they could have reduced angle of attack using the elevators.

Even if there was a notch in the wing pitching moment, it is likely that this aircraft could have developed enough nose down momentum to proceed past the notch. Particularly if, as the aircraft began to pitch down, they then began to roll nose down trim in using the manual trim. In effect, as the aircraft began to pitch nose down, they began to match (approximately) the rotation rate of the aircraft nose down with the THS rotation. Success would have been signaled by getting light in their seats and return of rapidly accelerating airspeed at approximately 45 degrees nose down attitude.

If you will think back to Gum's Viper deep stall recovery, the recovery technique was to pitch the aircraft up and then develop sufficient pitch momentum nose down to pass through the sticking spot. In effect (with all that NU THS), AF447 was starting from the stick back position already and just needed to start the ball rolling forward.

I am wondering whether as the aircraft began to rotate nose down, the crew began to get light in the seat as the aircraft began to fall faster. That also could have been a discouraging factor to continued nose down input.
Of course in my F-4, best acceleration occurred at zero g, but the AF447 crew was undoubtedly not accustomed to the concept.

[cwatters]

I am wondering whether as the aircraft began to rotate nose down, the crew began to get light in the seat as the aircraft began to fall faster. That also could have been a discouraging factor to continued nose down input.

and not everyone was in a seat or at least not strapped into one.

[DozyWannabe]

Problem, in their animation, BEA also 'forgot' to represent those 2 flights which adventured 'in the middle of the storm' … Are BEA and mass media so different …

The BEA aren't reporting on or scrutinising those flights, they're only working on AF447. I'm sure that the presence of those flights will be mentioned in the report if it turns out to be relevant. It's worth remembering that Delta 191 proved that it only takes a couple of minutes for weather conditions to change drastically - what might have been mild for those other flights may not have been for AF447.


There is absolutely matter for constructive discussion on a dedicated thread to Habsheim.
You could be surprised I don’t actually disagree on everything you write …

I don't think another public thread would be worth it - not only because there have been so many and the issue done to death, but because it almost always ends up in a bunfight. I'm happy to participate in PM discussions though.

[Svarin]

CONF iture asked :

Svarin, what’s your take on this ? I will summarize very quickly here :

At the start of accident sequence, two apparently unrelated faults :
PROBE PITOT 1X2/2X3/1X3
WRG:ADIRU1 BUS ADR1-2 TO FCPC2

PROBE FLR is very likely due to icing of probes by high altitude ice crystals.
It triggers a specific 10 seconds ADR checking process by all three PRIMs.
Starts 02:10:05.

All three PRIMs work in Alternate 2 for 10 seconds.

WRG FLR affects PRIM2 by cutting it from ADR1 data. My take is this happens right at the start of above mentioned 10 seconds ADR checking process.

At the end of these 10 seconds, each PRIM will either confirm and latch Alternate 2 or revert to Normal law, based on differences between ADR values. This is the purpose of this 10 seconds checking. This happens at 02:10:15.

My take is PRIM2 reverted alone to Normal, while PRIM1 & PRIM3 latched Alternate 2. This is because PRIM2's set of ADR data, being different (only ADR2 & ADR3), would have allowed differences into the 50 kts range that triggers reversion to Normal law.

The sentence "the aircraft was in Alternate 2" makes no sense. The PRIMs compute flight laws, not "the aircraft". All three PRIMs are 99.9999% of the time in full agreement. Not this time.

NAV ADR DISAGREE is CMC stamped 02:12 and ACARS ground-received at 02:12:51, which means it happened on board between 02:11:58 and 02:12:43. Aircraft was downwards fast at that time. Therefore, the ADR DISAGREE condition was not in effect to latch Alternate 2 for all three PRIMs at the start of pitch-up sequence at 02:10:16.

At 02:10:16, the pitch-up sequence and "zoom-climb" happened with (according to my research and "flimsy theory") PRIM1 & PRIM3 in Alternate 2 and PRIM2 in Normal.

[BOAC]

Svarin - that is interesting. Excuse the non-AB brain here, but what are you are saying in terms of the Air Data presented to the crew at the outset? I'm afraid 'Prim 1 and 3 in Alt2 and 1 in normal' needs some translation for me. With 1 in 'normal' and the other two in 'Alt2' what would be the governing control law? Would the ASI's L and R now show a disagreement? Would they be 'invalidated'. Would the standby ASI be affected? What other baro/ADC sourced info is useable? IE Would the altimeters/VSI show the extreme climb?

Lastly, can you comment on what effect (if any) a low airspeed in the system might have on autotrim?

Apologies for all the questions, but your post is vastly more important than tailplane camber!

[Lonewolf_50]

Where I'm not so sure if I understand it correctly is the thing with the Sine.
I will give it a try based on how I understood your approach:

Henra, thank you for post 125, I crossed up where to use the axis of aircraft (external coord system) and the wing chord. It appears that I became a flesh version of one of RetiredF4's confusers. :)

I had experienced glorious moments . . and I now realize that it was all very beautiful. But I shall never again fly through a thunderstorm . ... unless the Fatherland should demand it. "
Source : Manfred von Richthofen's letters to his family
Hmm, an early definition of "Operational Necessity," and a consideration not consistent with operating commercial transport aircraft. ;) OKC< thanks for the bon mot, The Red Baron has provided us with many such. :ok:

[Svarin]

BOAC asked :
Air Data presented to the crew at the outset? [...]
Would the ASI's L and R now show a disagreement? [...]
Would the standby ASI be affected?BOAC, your questions are all related to what the crew is being provided as indications on their instrument panel. What I am dealing with is computerized flight controls.

ASIs were affected. But there is no way I could discern exactly how and when. Inside my own subset of research in this, however, ADR data (provided to Flight Controls Computers -> the PRIMs) was affected. In Normal law, PRIMs could effect ultimate authority on flight control surfaces. There are 3 of them. What happens when PRIMs are in disagreement ? Now that is a question...

FDR might reveal some of these parameters. It is however unfortunate that ADR2 data and associated RHS airspeed was not recorded on FDR...

[Svarin]

BOAC asked :

With 1 in 'normal' and the other two in 'Alt2' what would be the governing control law?A simplified approach to this problem is to examine which computer governs which flight control surface, and examine carefully how the different flight laws work, and how they interfere with pilot actions. Reality would tend to be even more complex.

Occam's try on these hypotheses is the triggering of an undesired normal law (full authority) overspeed protection by PRIM2 only which uses erroneous airspeed data after having unexpectedly returned to Normal law. This protection reacts to speed only, and simply adds a nose-up order, with authority superior to that of pilot sidestick. Alarms and warning triggers to the crew are a big unknown in this situation...

This situation I described is absolutely not what should happen on a FBW aircraft. Flight Controls Computers are not supposed to disagree at every turn. They are almost always in full agreement to the 0.0000001% precision.

I have determined, however, that in this particular instance, they were likely not in agreement. After such an unprecedented problem, all bets are off.

This is not fundamentally different from the unexpected rudder hardover in a 737, or the unexpected loss of all hydraulics on a DC-10 or 747, or whatever... These happen extemely rarely and are always specific to a certain design. No design is perfect, even this one.

Designs always contain certain trade-offs, carry certain basic assumptions, and try to cater for the unexpected in the best way possible. The more complex they are, however, the more prone to human-like error they become.

James Reason is always quoted for his Swiss-cheese model, but he could also be quoted for this : "intelligence and error are both sides of the same coin".

[SaturnV]

Dozy, the BEA seems to have had extensive conversations with the crews of the other three flights (Lufthansa, Iberia, and AF459).

The first interim report had this to say about AF459's use of radar. (AF459 was 35 minutes behind at 350)
After flying through a turbulent zone in the head of a cumulus congestus formation at the level of NATAL, without having detected this zone on the radar, he selected gain in MAX mode. At about 2 h 00, he observed a first echo that differed significantly depending on whether the radar’s gain was in CAL or MAX mode. The TILT was set between -1° and 1.5° ....the screen, which was set to a scale of 160 NM. The echoes were yellow and red when the radar was set with gain on the MAX position and green and yellow when the gain was on the CAL position

Lufthansa (20 minutes ahead at 325, (FL based on AMDAR trace))
The crew reported that it flew at the upper limit of the cloud layer and then in the clouds in the region of ORARO. In this zone they saw green echoes on the radar on their path, which they avoided by changing their route

Iberia (12 minutes behind at 370), the BEA report does not mention their use of radar.
These conditions were particularly severe 70 NM to 30 NM before the TASIL waypoint. They moved away from the route by about 30 NM to the east to avoid cumulonimbus formations with a significant vertical development, and then returned to the airway in clear skies close to the
TASIL waypoint.

As for AF459, same model airplane, presumably same radar system, similar flight crew training, and also had presumably received a similar message from dispatch re: satellite observation of Cbs vicinity ORARO-TASIL. (AF459 departed approximately same time as AF447, but from further south.) Did the captain of AF459 take a rest at the typical time? Did AF459 change comm frequency away from ATLANTICO immediately after a SELCAL check? Comparisons of behavior between the two crews could be quite enlightening.

Even in the absence of a full CVR, inferences can start to be drawn regarding the situational awareness of the AF447 flight crew, before and during.

In its brief note, the BEA did not mention any crew conversation about the radar setting. If they had changed the mode and discussed it, would not the BEA have mentioned such? The BEA does note that the flight crew notified the cabin crew that they would be going through an area of turbulence, but no indication of any specific instructions to the cabin crew to be seated, stop service, etc. The BEA does not note any attempt by the crew to communicate with ATLANTICO or DAKAR about its circumstances, even though it had looped in its heading. (If IB6024 had not deviated, it would be only eight minutes behind at the climactic point.)

My own feeling is that if there was exculpatory conversation on the CVR, the BEA would have referenced such in its brief note.

[HazelNuts39]

ATSB's 1st Interim Report on the QF32 pitch-down accident describes the role division between the PRIMs:
Review of PRIM monitoring functions
The aircraft’s flight control system included three flight control primary computers (FCPCs, commonly known as PRIMs) and two flight control secondary computers (FCSCs, commonly known as SECs). One PRIM functioned as the master while the other two PRIMs could take over as master if a fault in the current master was detected. The master PRIM processed and sent control surface deflection orders to other computers, which executed them using servo-controls. The two other PRIMs continuously computed control orders and monitored control surface deflections but those orders were not actioned.
Each PRIM consisted of two independent parts, a Command (COM) part and a Monitor (MON) part. The MON part monitored the performance of the COM part and the position of the control surfaces. If there was a discrepancy between COM and MON, then the PRIM would ‘fault’ itself. The fault could be for only a part of the PRIM (for example, pitch channel) or for the whole PRIM. A PRIM could not generate a fault for the whole PRIM unless it was the master. The PRIM Fault parameter recorded by the FDR was active only for a fault of the whole PRIM and not for a partial fault (for example, a pitch channel fault). However, partial faults were recorded by the PFR.
For elevator control, the active servo-controller in normal operation was PRIM 1. The servo-controller priority order was PRIM 1, PRIM 2, SEC 1 and SEC2. If PRIM 1 could not perform this function, then the servo-control function reverted to PRIM 2 and so on.

Table 4 provides a sequence of events for the PRIMs and is based on a review of the FDR and PFR data by the aircraft manufacturer and investigation team.

Table 4: PRIM sequence of events
(table deleted, does not reproduce in PPRuNe - HN39)

In summary, the PRIM PITCH FAULTs and PRIM 3 FAULTs that occurred during the flight were consistent with the system design. They were consequences of the pitch-down events and not the initiators of those events.

[A33Zab]

CONF iture asked :Svarin, what’s your take on this ?

I will summarize very quickly here :

At the start of accident sequence, two apparently unrelated faults
PROBE PITOT 1X2/2X3/1X3
WRG:ADIRU1 BUS ADR1-2 TO FCPC2

PROBE FLR is very likely due to icing of probes by high altitude ice crystals.
It triggers a specific 10 seconds ADR checking process by all three PRIMs.Starts 02:10:05.

All three PRIMs work in Alternate 2 for 10 seconds.

WRG FLR affects PRIM2 by cutting it from ADR1 data. My take is this happens right at the start of above mentioned 10 seconds ADR checking process.

At the end of these 10 seconds, each PRIM will either confirm and latch Alternate 2 [B]or revert to Normal law, based on differences between ADR values. This is the purpose of this 10 seconds checking. This happens at 02:10:15.

My take is PRIM2 reverted alone to Normal, while PRIM1 & PRIM3 latched Alternate 2. This is because PRIM2's set of ADR data, being different (only ADR2 & ADR3), would have allowed differences into the 50 kts range that triggers reversion to Normal law.

The sentence "the aircraft was in Alternate 2" makes no sense. The PRIMs compute flight laws, not "the aircraft". All three PRIMs are 99.9999% of the time in full agreement. Not this time.

NAV ADR DISAGREE is CMC stamped 02:12 and ACARS ground-received at 02:12:51, which means it happened on board between 02:11:58 and 02:12:43. Aircraft was downwards fast at that time. Therefore, the ADR DISAGREE condition was not in effect to latch Alternate 2 for all three PRIMs at the start of pitch-up sequence at 02:10:16.

At 02:10:16, the pitch-up sequence and "zoom-climb" happened with (according to my research and "flimsy theory") PRIM1 & PRIM3 in Alternate 2 and PRIM2 in Normal.

If so, this would be a serious system flaw and therefore I disagree.

There’s only 1 PRIM in control and that will be the one which can deliver the highest possible law and in the sequence PRIM 1, 2 & 3

If PRIM 2, in your view, is in control and able to deliver NORMAL LAW this means that PRIM 1 was not able to compute NORMAL LAW protections and doesn’t say anything about the ability of PRIM 3. If PRIM 2 is in control in ALTERNATE LAW this means PRIM 1 and PRIM 3 could not deliver NORMAL LAW and PRIM 1 was also unable to deliver ALTERNATE LAW.

I’m sure if PRIM 2 was in (NORMAL or ALTERNATE) control this had already been mentioned in the BEA update.

ALTERNATE 2 mode is latched, so once it is declared (at the time message F/CTL ALTERNATE LAW appeared on ECAM and PF called it out) there can only be a reset on ground with all hydraulic systems depressurized.

In my view, at that time PRIM 1 was in control, PRIM 2 and 3 couldn’t compute NORMAL LAW either.
Due to the PITOT problem they had all the same ADR information.

PRIM 1 declared ALTERNATE 2 LAW and no PRIM can revert to a higher law anymore even if ADR becomes normal thereafter.

[GarageYears]

savrin:

How do you fit your analysis with this from the BEA note?

At 2 h 10 min 16, the PNF said "so, we’ve lost the speeds" then "alternate law […]".

You state:

My take is PRIM2 reverted alone to Normal, while PRIM1 & PRIM3 latched Alternate 2. This is because PRIM2's set of ADR data, being different (only ADR2 & ADR3), would have allowed differences into the 50 kts range that triggers reversion to Normal law.

<snip>

At 02:10:16, the pitch-up sequence and "zoom-climb" happened with (according to my research and "flimsy theory") PRIM1 & PRIM3 in Alternate 2 and PRIM2 in Normal. 20th Jun 2011 05:55

Unfortunately, based on the voice comment above, I seriously doubt the control law in effect was Normal during the "zoom-climb" event, unless you are saying that the PF controls were in Normal, while PNF was seeing Alternate (but my understanding is that this is not possible).

[Lonewolf_50]

GY

As the system is supposed to work (based on what I understand of it) that seems to be the case, but what has Svarin concerned seems to be the system not working as advertised ... which would lead the cockpit crew into some confusion regarding "what instruments are right?" and "what is it doing now?" In that state of confusion, attention to flying can easily be diverted, and diverted attention is one of the few hypotneses available to explain an inadvertent 3000 foot climb. <-------- That makes little sense.

(Pitch and power chorus in three, two, one ... ) ;)

If there's a bug, how to you find it? How much time and effort, and what initial conditions are necessary, to replicate it?

And, cui bono? (Small bit of tinfoil on me head for the following)

What incentive does AB have to pursue a very low probability bug with only circumstantial, and possibly second or third order, evidence? From a practical point of view, how do you determing the initial conditions (internal to the system) to try and trigger it? With a certain amount of information missing ... it may not be doable.

[ChristiaanJ]

With a certain amount of information missing ... it may not be doable.Please....
Don't confuse what we are trying to deduce here, from just an initial 'text' description of the 'events' from the BEA note, with the work being done by the BEA, with (one expects) AB support.
THEY have the traces, the full FDR and CVR records, and the full documentation of the FCS.
I'm sure they're going through some of the same discussions, but with a far greater amount of information at hand.

I've seen no mention of the QAR having been recovered. It might have provided even more info.

[Lonewolf_50]

Understood, ChristiaanJ, and I'm not trying to get too far ahead of the problem.

The problem with electrical ghost faults is the damnable time one has in replicating them. I've had gripes and write-ups that repeatedly got signed off as "could not replicate on deck" which sometimes, because they could not be replicated, were very difficult to trouble shoot.

That is also true for spurious signals and strange behavior in computer driven equipment in aircraft. Sometimes, the only maintenance action was to reboot the damned thing.

If the QAR is available, that raises the odds for successful analysis, but it is no guarantee.

[Svarin]

Lonewolf_50 writes :

the system not working as advertisedYes, my set of hypotheses sees it having worked as programmed, not as expected, even by its designers, making the expression by design completely meaningless.

Lonewolf_50 again :

electrical ghost faultsGoing further into detail, I am of a mind that the WRG message is not a physical wiring fault, but a communication problem, software-related, between PRIM2 and ADR1. This would therefore not be a ghost fault. It would be reproductible.

One would need to mount the whole identical system (all computers, soft versions, P/N, etc...) on a simulator and introduce the specific 10 seconds monitoring process, and see what happens. This 10 seconds process being likely the only case for which communication breakdown would occur between these two computers.

GarageYears remarked :

I seriously doubt the control law in effect was Normal during the "zoom-climb" eventYou are certainly welcome to show this healthy skepticism regarding the whole PRIMs sour business I have posited.

May I summon A33Zab great technical information for help ?

There’s only 1 PRIM in control and that will be the one which can deliver the highest possible law and in the sequence PRIM 1, 2 & 3 (emphasis mine)

This means Normal law is preferred over any inferior laws. This is the very core of this flight controls design.

If PRIM 2, in your view, is in control and able to deliver NORMAL LAW this means that PRIM 1 was not able to compute NORMAL LAW protectionsQuite right, except PRIM1 is not faulted before 02:13. Which PRIM becomes master if PRIM1 is ALT2 and PRIM2 is Nz ?

In my view, at that time PRIM 1 was in control, PRIM 2 and 3 couldn’t compute NORMAL LAW either.
Due to the PITOT problem they had all the same ADR information.With all due respect to your extreme technical knowledge, may I introduce this disagreeing parameter ?
WRG:ADIRU1 BUS ADR1-2 TO FCPC2

This means PRIM2 does not have the same ADR set available to it than the other two PRIMs have.

[henra]

At the end of these 10 seconds, each PRIM will either confirm and latch Alternate 2 or revert to Normal law, based on differences between ADR values. This is the purpose of this 10 seconds checking. This happens at 02:10:15.



Hmm, but if Normal law was available again, wouldn't there also have been the stall-protection active again ???
From what is happening in the next two minutes I do not see much what resembles the expected reaction of Normal Law, given the AoA and speed.
Could you point me to any single behaviour that you would logically expect in Normal Law ?

Somehow, this theory strikes me as a bit strange.

[BOAC]

We have a whole bunch of disagreement on this from those who claim to know the system, and I wonder what chance two lads had in the middle of the night in the ITCZ?

Cold anyone venture a definitive answer to what they had left in baro terms with those ACARS warnings?

[MurphyWasRight]

the probable cause of this accident was preoccupation with an inflight malfunction and failure to monitor properly the airplane's flight instruments which resulted in losing control of the airplane.



The flightcrew's statements about the ADIs failing were not substantiated by the facts. It is most likely that the flightcrew became spatially disoriented during the upset.

They were unable to believe the information displayed on the ADIs, did not recognize the unusual attitude of the airplane, and were unable to take the correct action to recover the airplane until it began to emerge from the clouds.


The increased automation has not necessarily reduced pilot workload, however, but has shifted it to monitoring tasks which the pilot formerly had to perform, and there is evidence, from both research and accident statistics, that people make poor monitors.


Research also indicates that the excursion from a stabilized condition might be exaggerated even after a system anomaly is detected, because of the period required for a pilot to transition from system monitor mode to system controller. Time is needed to "ascertain the current status of the airplane and assess the situation," 13/ before the pilot can reenter the control loop and take corrective action.


The above is not a new AF447 leak or even Airbus related. It is from the 1985 NTSB report on the China Ailines 747 loss of control initiated by a single engine failure at altitude. (my bold)
China Airlines B747SP Loss of Power and Inflight Upset (http://www.rvs.uni-bielefeld.de/publications/Incidents/DOCS/ComAndRep/ChinaAir/AAR8603.html)

Fortunatly they did break out of clouds in time to recover, although with some injuries and extensive damage.

[wozzo]

My take is PRIM2 reverted alone to Normal, while PRIM1 & PRIM3 latched Alternate 2. This is because PRIM2's set of ADR data, being different (only ADR2 & ADR3), would have allowed differences into the 50 kts range that triggers reversion to Normal law.

So PRIM2, which is missing on set of ADR data completely (according to this interpretation of ACARS messages), doesn't care about the missing data and computes with only 2 sets as if there had been always only 2 available and never 3? I'd rather expect that, given the importance of the these data, PRIM2 would be less likely to return to NORMAL.

[mm43]

MurphyWasRight;

Well said.:ok:

It is time to stop trying to apportion blame to aircraft systems, and time to take a long hard look at the automation ~ human interface. It has often been said in these threads that the automation is 99.9^n% right, and now is the time to look at the human interface and the the well known fact that the same degree of reliability is and can not be expected of mere mortals.

We have spent endless pages arguing over which pilot was seated where, and which of the two F/O's was PF. There is no need to go there in the meantime, but the outcome may help in determining who was where.

I know this Aviation Weekly & Space Technology (http://www.aviationweek.com/aw/generic/story.jsp?channel=comm&id=news/awst/2011/06/06/AW_06_06_2011_p36-330706.xml&headline=null&next=10) link has been posted by PJ2 previously, but it drew no response. I suggest that it is read and digested, compared with the China Airlines incident that MurphyWasRight has mentioned, and let the discussion start on training and how to keep the pilot in the automation loop.

I suspect that the crux of this accident is not hardware or software, but a human condition, i.e. industry, management, training and CRM, that needs to be addressed.

[HazelNuts39]

At the end of these 10 seconds, each PRIM will either confirm and latch Alternate 2 or revert to Normal law, based on differences between ADR values. This is the purpose of this 10 seconds checking. This happens at 02:10:15.

My take is PRIM2 reverted alone to Normal, while PRIM1 & PRIM3 latched Alternate 2. This is because PRIM2's set of ADR data, being different (only ADR2 & ADR3), would have allowed differences into the 50 kts range that triggers reversion to Normal law.

From BEA's explanation of PROBE PITOT fault message (Interim #1 p.51): This message, transmitted by the FCDC2 (EFCS2), means that the FCPCs (or PRIMs) triggered one of the speed monitoring processes: they have detected a decrease of more than 30 kt in one second of the “polled” speed value. The three ADRs were considered valid by the EFCS2 at the time the monitoring was triggered, because the prior rejection of an ADR would have generated a class 2 fault message and there would therefore have been an asterisk in front of the source. In this case, the “polled” value is the median value.

At the time this monitoring is triggered, the FCPCs open a window during which they operate with alternate 2 law (see following graphic). The rudder deflection limitation function is also frozen, but the associated alarm is inhibited. At the end of the window, if the difference between the values polled at each end of that window is less than 50 kt, the FCPCs return to normal law. Otherwise, they continue in alternate 2 law, the rudder deflection limitation function remains unavailable and the corresponding alarm is generated.
Note: the alternate 2 control law is a load factor law for pitch and a direct law for roll. Only the load factor protection remains available. In certain cases, the high and low speed stabilities may also be lost.

[Teddy Robinson]

BEA put out their initial findings to quieten the press clamour, as predicted in thread 3, all is quiet as they complete their investigation, at which point speculation can stop, and discussion begin .. I say that with no disrespect to some very informed input from some people here .. others less so.

If we, as aviators, look back to brass tacks, not one of us would want to be in the situation whereby we are trying to figure out how an aircraft is trying to stop us from having an accident, whether that is through protections, stick shakers, buffet .. whatever, brass tacks is about making the decisions to avoid being in a position where those protections kick in, here I err towards the rational of those pointing out the deviations of previous aircraft, and the personally observed fallability of AWW. From first reading, there are issues that will become clear when the final report is issued, they will probably focus on training and awareness rather than discuss big red buttons to give the airplane back to the pilots.

The wider point is surely, that loss of situational awareness took THIS aircraft into a regime that it could not make sense of, and took the crew and passengers with it, where as other aircraft did not encounter the same conditions, primarily by not being there in the first place.

Whilst all the technical discussions as to what the aircraft, crew, and training culture subsequently did to neautralise a dynamic situation are appropriate to patching the responses in future, (even if the effect turns out to be that under stress all those protections cancelled one another out) it is surely the avoidance of the situation in the first place which closes the first hole in the Swiss cheese model.

Lets wait for the report.

[lomapaseo]

Whilst all the technical discussions as to what the aircraft, crew, and training culture subsequently did to neautralise a dynamic situation are appropriate to patching the responses in future, (even if the effect turns out to be that under stress all those protections cancelled one another out) it is surely the avoidance of the situation in the first place which closes the first hole in the Swiss cheese model.



Of course well spoken (all the stuff I left out of the quote), but it is well to remember that it is harder to prevent the initiating causes then it is to provide mitigation or accomodation to these surprises.

I think about all the stuff that can go wrong every time I fly, but in the end I convince myself that even average pilots can save the day.

[john_tullamarine]

but in the end I convince myself that even average pilots can save the day.

Indeed, most of us would be far more concerned about getting killed driving to and from the airport - although I sometimes shudder at some of the third world operators into whose tender care I have entrusted my life over the years.

Risk principally is tied up with probability/seriousness of outcome/exposure period/frequency. It is human nature to accept mitigated risk. Caveat - young men, being programmed genetically to be less worried about risk than the rest of us, might tend to be less concerned about the mitigation bits and more about the adrenaline buzz.

We just have to accept that, every now and again, the ducks all line up nicely to give Lady Luck a clean shot at the heart. It is this phenomenon which drives much in the way of regulated technological development.

There are no guarantees, only probabilities.

[PA 18 151]

There are no guarantees, only probabilities.Absolutely 100% spot on.

If when something unusual happens (figures plucked out of the air for effect)

the plane is 99.99999% reliable
SOPS are 99.9% reliable
A captain 98 % reliable
A first officer 90% reliable

You are going to get occasions when you lose the aircraft. Fortunately on the rare occasions that the plane gets it wrong one of the pilots has a clue, it should come as no surprise that sometimes both of them don't.

[PJ2]

john;

There are no guarantees, only probabilities.

Yes, and possibilities. The notions can be complementary or independently considered. It may be possible but not probable, it may be possible and probable and the other way around. The thinking opens previously blind paths to understanding an event. The discussion surrounding the Fukushima accident is one such event where the probability was considered "rare" that the generators would be flooded by a tsunami, but, despite a number of engineers calling for consideration of the "possibility" of flooding (and following through with an examination of the outcomes of that possibility), it was never done and within an hour or so, the pathway to the worst nuclear disaster in history was set at Fukushima.

Lee Clark, Possibilistic Thinking (http://leeclarke.com/docs/clarke%20thinking%20possibilistically%20-%20Significance.pdf)

[DozyWannabe]

We have a whole bunch of disagreement on this from those who claim to know the system, and I wonder what chance two lads had in the middle of the night in the ITCZ?

It's not disagreement over the systems behaviour, it's a group of people who know the systems and are willing to stick to the facts as presented versus a group of people who may know the systems, but are bending every bit of minutae of that knowledge to advance their agenda that the aircraft (and consequently the manufacturer) must somehow be completely to blame.

Quite how this seemingly endless reservoir of venom towards Airbus and the BEA came about I'm not entirely sure...

[BOAC]

DW - that's accepted, but you need to open your horizons and consider the possible effect the whole 'automation' mentality is having on the industry - read mm43's post above and my somewhat extinct thread on 'Safety and CRM and Q&A'.

In addition, perhaps you could give us a definitive analysis of exactly who is providing the accurate information here and who is 'bending every bit of minutae of that knowledge to advance their agenda' and we can then disregard these other posts, since for those of us who are not 'experts' it is exceedingly difficult to know who is spouting rubbish.when statement after statement is challenged by someone else 'claiming' to be expert.

With your expertise, could you also tell us (as I have asked) exactly what baro information would be available to which pilot with the given failures? I also have not had an answer (to Svarin) in post 207 "Lastly, can you comment on what effect (if any) a low (reported) airspeed in the system might have on autotrim?" - perhaps you could oblige? Is the autotrim function SOLELY based on elevator demand or is there a speed input?

[lomapaseo]

In addition, perhaps you could give us a definitive analysis of exactly who is providing the accurate information here and who is 'bending every bit of minutae

I hope that no names are given as it defeats an open forum discussion.

Point and counterpoint will suffice for most of us.

Often I have no expertise in the discussion but do admire a posters ability to rationalize arguments with facts rather than just analysis of minutae to furthur a suspicion.

[Lonewolf_50]

I know this Aviation Weekly & Space Technology (http://www.aviationweek.com/aw/generic/story.jsp?channel=comm&id=news/awst/2011/06/06/AW_06_06_2011_p36-330706.xml&headline=null&next=10) link has been posted by PJ2 previously, but it drew no response. I suggest that it is read and digested, compared with the China Airlines incident that MurphyWasRight has mentioned, and let the discussion start on training and how to keep the pilot in the automation loop.
With respect, I disagree with that last, and suggest to you that the point is to keep the pilot in the control loop. Automation is already there, and either substitutes for, or aids, the pilot in a variety of cockpit task areas.

The pilot is required to be in command. What may need some tweaking is making a better fit for him in the control loop to better enable his exercise of command. It come's with the pilot's badge.
I suspect that the crux of this accident is not hardware or software, but a human condition, i.e. industry, management, training and CRM, that needs to be addressed.
Which ones will be addressed, and how thoroughly? We shall see.

[BOAC]

loma - don't forget that 'names' here are not 'names' and an accusation that posters are 'misleading' folk is potentially equally wrong, may yet turn out to be part of a different 'agenda' and needs to be clarified.

"Point and counterpoint will suffice for most of us" - therein lies the rub - whose 'counterpoint' do we listen to? So far, AB software 'expert' opinion has been challenged repeatedly so that none of us know who is right, and as I said earlier, what hope for a crew in an upset?
.
Now then - anyone willing to answer my queries - definitively, correctly and without fear of contradiction?

Lonewolf - spot on.

[Tailspin Turtle]

In addition, perhaps you could give us a definitive analysis of exactly who is providing the accurate information here and who is 'bending every bit of minutae of that knowledge to advance their agenda' and we can then disregard these other posts, since for those of us who are not 'experts' it is exceedingly difficult to know who is spouting rubbish.when statement after statement is challenged by someone else 'claiming' to be expert.

It ain't that easy to tell. I'm by no means an armchair aviator: aeronautical engineer, F-4 and helicopter flight test engineer, pilot (ATP but not used as such), 3,000 total hours, part-time flight/instrument/aerobatic instructor, rated in airplanes/gliders/helicopters, flown in national glider contests, etc. I was absolutely certain that they had lost control into a spiral dive, figured it out, and were just too late on the recovery and pullout. Posted that three times, different ways. No way they were in a flat spin because they almost certainly could not have recovered and therefore would not have crashed into the water with no significant yaw rate ("en ligne de vol"). Didn't even consider that they could not only have stalled it but managed to keep it stalled for 35,000 feet without falling off on a wing into a spiral dive or a spin. Impossible. Gob smacked I was.

I once convened my production test pilots, all very experienced, to discuss a simple power margin procedure that customer pilots were using and getting different results from ours. They all wondered why I asked since it was so straightforward and well known. Then each of the six told me how they did it, none exactly the same, maybe eight ways in total since a couple of the guys had alternatives. And the dispersion of the results when I had them fly the same aircraft was unacceptably large. It took a while but we finally agreed on and documented a specific procedure that gave reasonably consistent results. So it's no surprise to me that pilots disagree amongst themselves. Also see the debate about what controls airspeed and flight path (n.b. I use stick and throttle together or one or the other as required.)

In that regard, one of the things that is apparent from the discussion is that the Airbus protections and fall backs are very, very complicated. Even the knowledgeable and/or formally trained appear to disagree (or at the least quibble) about how the software and computers work. That doesn't seem like a good thing...

[Lonewolf_50]

Tailspin Turtle:
I use stick and throttle together or one or the other as required.

Likewise.
next bit Deleted as irrelevant to A330.
Part of the idea behind teaching that way was to keep the student from trying to fly by just moving the stick. (I wonder if autothrottle lends itself to that functional habit? )
Next bit deleted as irrelevant to A330.

EDITED per BOAC's valid complaint.

[BOAC]

This is so irrelevant as to be unbelievable! Please continue on the Flying Instructors' thread? What ON EARTH has this to do with AF447?:mad:

[A33Zab]

SVARIN:

This means Normal law is preferred over any inferior laws. This is the very core of this flight controls design.
Quote:
If PRIM 2, in your view, is in control and able to deliver NORMAL LAW this means that PRIM 1 was not able to compute NORMAL LAW protections

Quite right, except PRIM1 is not faulted before 02:13. Which PRIM becomes master if PRIM1 is ALT2 and PRIM2 is Nz ?
Quote:


Both NORMAL LAW as ALTERNATE LAW are Nz laws, to keep it ‘simple’ you mean PRIM 1 is NOT able to compute NORMAL while PRIM 2 is capable.
Then I agree with you PRIM 2 will be in control.

And actually this happens once in a while in normal flight, our crew reports in such a case: 'Alternate Law message was transient'.
F/CTL ALTERNATE LAW did appear on ECAM and after a while (when system reverted to NORMAL again) the message disappeared from ECAM.

Technically speaking ‘ALTERNATE LAW was triggered at the start of the monitoring window and at the end the monitoring window the system reverted to NORMAL again’
A correlated failure, which triggered the ALTERNATE LAW may be present in CMC, maintenance action however is not required when crew reported ‘transient’.

When the system later on triggers an ALTERNATE LAW again the F/CTL ALTERNATE LAW warning will be set again with the correlated fault message at that time.

Back to AF447,
If PRIM 2 reverted to NORMAL the NAV ADR DISAGREE would set ALTERNATE LAW a 2nd time (at the start of the monitoring window) and that did never occur because system was already in ALT 2 and PRIM 2 NOT in NORMAL!

What we know is that F/CTL ALTERNATE LAW was triggered after (PITOT ICING) and latched to ALT 2 (due missing speed information) at the time of F/CTL RUD TRV LIM FAULT and confirmed by frozen position of Rudder Travel Limiter.
This mode is latched – for the remainder of the flight - to prevent cycling between the modes.

From the BEA report:


Like the FMGECs, the PRIMs consolidate the parameters that they use by
means of monitoring mechanisms.
Concerning the airspeed, it is the voted value that is used.
In normal operation, this is the median value.
When one of the three speeds deviates too much from the other two, it is automatically rejected by the PRIMs and the polled value then becomes the average of the two remaining values.
But if the difference between these two remaining values becomes too great the PRIMs reject them and the control law switches to alternate 2.
Furthermore, another monitoring procedure is applied to the value of the voted
airspeed and triggers switching to alternate 2 law when it falls by more than 30 kt in one second.






In my view, at that time PRIM 1 was in control, PRIM 2 and 3 couldn’t compute NORMAL LAW either.
Due to the PITOT problem they had all the same ADR information.

With all due respect to your extreme technical knowledge, may I introduce this disagreeing parameter ?
WRG:ADIRU1 BUS ADR1-2 TO FCPC2

This means PRIM2 does not have the same ADR set available to it than the other two PRIMs have.


Agree, but doesn’t necessarily mean this set gives a more reliable outcome to the median speed calculation. On the contrary 3 ADRs is more reliable, as long as you can speak of reliable in this case, than 2 sets of signals.

I can’t find it in the documents but it could be that the logics take the amount of available ADR signals in account.

Then for this Maintenance message:

The problem area:

http://i474.photobucket.com/albums/rr101/Zab999/ADIRU1WRG.jpg


As preface on my argumentation keep in mind, EFCS is the complete Electronic Flight Control System and consist of FCPC1, 2 & 3, FCSC1 & 2 and FCDC 1 & 2.
When a failure is detected, PCPCs and FCSCs take operational reconfigurations and sends failure information to FCDCs, FCDCs analyzes the failure and after confirmation generates a maintenance message. The message source will be EFCSx(FCDCx) as will be the identifier EFCSy(FCDCy).
If both FCDCs are available, FCDC1 sends external and internal (ATA 27 ex S/F) failures while FCDC2 sends all failures as external to avoid double storage within CMC.
ADIRU failures are considered as external failures, however because only FCPC2 reported this failure it is rated (ATA27) internal.
Furthermore FCPC 1, 2 & 3 are interchangeable units (as long as they have the same P/N and OBRM*) only their position in the racks determines if they act like FCPC 1,2 or 3.
*OBRM = On Board Replaceable Module

There are extensive BITE* tests available in the system, the logic however is not detailed.
(Wrap Around, SSM, parity and refresh rate tests may be expected within BITE)
*BITE = Build In Test Equipment

A is not B,
In A. a wiring failure is suspected only, if BITE analysis has revealed its fault. AMM states: “As soon as there is more than one component (more than one part) in the message, A/C wiring can be suspected but it is not indicated in the messages except if it is clearly identified as being the origin of the fault.”

The failure messages in EFCS are always declared as HARD, even when it was INTERMITTENT (t<=2.5 s) or HARD due to several INTERMITTENTS during flight.

Based on this it should be considered as wiring fault i.s.o software flaw(OBRM) but can't conclude if it was of any influence (HARD) or not (INTERMITTENT) in law reconfiguration.

[A33Zab]

SVARIN:


Lonewolf_50 again :

electrical ghost faults

Going further into detail, I am of a mind that the WRG message is not a physical wiring fault, but a communication problem, software-related, between PRIM2 and ADR1. This would therefore not be a ghost fault. It would be reproductible.

One would need to mount the whole identical system (all computers, soft versions, P/N, etc...) on a simulator and introduce the specific 10 seconds monitoring process, and see what happens. This 10 seconds process being likely the only case for which communication breakdown would occur between these two computers.



If not already done in the last 2 years,
'IRON BIRD' facility, keep left on leaving the Toulouse training centre.

A33Zab.

[wmelvin]

From Wikipedia I find: "On Airbus planes pushing throttles to TOGA detent does all regarding flight path and speed", and in Thread 3, post # 16, I find from jcjeant the English translation from French: "At 2H10.51 this is again a stall alarm.the power levers are put on the position TO/GA and the PF keep the command to climb. Incidence angle of 6° when sounded the stall alarm continue to rise.
The adjustable horizontal stabilizer go from 3° to 13° (climb position) in about 1 minute.
He will stay in this position to the end of the flight."

I assume the reference to "incidence angle" is what Americans call "Flight Path Angle"; so my question is: "did the TO/GA selection command a FPA of 6 degrees and drive the elevator trim to full nose up to attempt to accomplish that?

[ChristiaanJ]

I assume the reference to "incidence angle" is what Americans call "Flight Path Angle"...I think you're a victim of Google translatese.... it almost certainly refers to AoA, and not FPA.
My apologies, but I really can no longer be bothered to spend twenty minutes to go through the French original to find the text in question AND the misleading translation, every time one of these issues crops up, unless somebody pays me.
Most professional translators will give you the same answer.....


No insult to Google translation meant... they have become a lot better the last couple of years!

[rgbrock1]

ChristiaanJ:

As a fluent speaker of German I can state that Google translations, although somewhat better in recent years as you wrote, still leaves an awful lot to be desired. There are just certain nuances and ideas in other languages that suffer in translation to English.
I'm sure you know that as well.

[AlphaZuluRomeo]

I've seen no mention of the QAR having been recovered. It might have provided even more info.
I think I remember an avionic bay was recovered (coded name, on the wreckage map), and the QAR was (hoped to be?) part of it (or did I assume that? Anyone who know better, please correct me if I'm wrong with this assumption).
Didn't hear about the QAR since, though...

[Lonewolf_50]

@BOAC
This is so irrelevant as to be unbelievable! Please continue on the Flying Instructors' thread? What ON EARTH has this to do with AF447? http://images.ibsrv.net/ibsrv/res/src:www.pprune.org/get/images/smilies/censored.gif Based on two years of reading PPRuNe threads on the mishap ... pitch and power. But I get your point on maintaining focus. :O So I deleted it.

A33Zab
Lonewolf_50 again :
If not already done in the last 2 years,
'IRON BIRD' facility, keep left on leaving the Toulouse training centre.
The post you qouted for that reply was not written by me. Perhaps Svarin?

[wmelvin]

Although the Wright brothers used angle of incidence for what we now call AOA, my question still stands; Does the GA mode also command a flight path? If so, can it use elevator trim to accomplish this? The original French is in Thread 3, post 16.

[mm43]

wmelvin;
Does the GA mode also command a flight path? If so, can it use elevator trim to accomplish this?Be careful when reading material which is not in context. The control law being used by AF447 at the time you refer to was not Normal Law. It was Alternate Law No.2 and pitch protections available in Normal Law had been lost.

The elevator NU command was initiated and maintained by the PF and the Trimmable Horizontal Stabilizer moved accordingly to preserve the elevators commanded effectiveness. Thrust was also PF controlled.

Check this summary of A330/A340 Flight control laws (http://countjustonce.com/a330/a330-flight-laws.html) for further info.

[HazelNuts39]

Although the Wright brothers used angle of incidence for what we now call AOA, my question still stands; Dassaults Lexique bilingue of aeronautical terms gives these translations:
angle d'incidence = angle of attack/incidence
angle d'incidence critique = stall angle
angle de pente = flight path angle

At 2:10:51 AoA was increasing, FPA was passing through a minimum of about 1.5 degrees up.

For what it's worth, wmelvin's question reminds me that at that point the stall warning computer (the computer exercising that function) must consider its polled airspeed to be valid, to calculate a stall warning threshold of 6 degrees. Iow the ADR DISAGREE condition occurred later.