jcjeant

Hi,

This french stance is very interesting to analyze word by word ....
I will only focus on 3 words "gauche" (left) .. 30 (30) .. and "secondes" (seconds)
So the stick was maintened on a full deflection to the left ... and yet .. the plane will not go in a left turn
Think ... maximum deflection to the left (for 30 seconds .. what a very long time) and the aircraft would not embark on a sharp left turn (or bank angle) ... very strange ..
What can prevent him from taking a left bank angle ?
Remember that it is to counter a bank angle to the right ..
So the force pushing the plane to a bank angle to the right should be very strong !!!
Wat can be this force ? .. what can produce this phenomena ?

HazelNuts39

A stall ...?

PA 18 151

Spot on.

The question jcjeant should really be asking is why was the PF trying to use aileron to counter wing drop.

RR_NDB

Machinbird

CONF iture There is also a technical reason for this: Technically speaking i´m not sure it would be possible to go deep by lack of data (recorded information). During the critical phase: The AS was lost (simply was not measured properly, despite the "redundancy"), and the mysterious WRG failure could also be impossible to be "understood" by lack of enough information to perform the analysis.

HazelNuts39

IIRC 'low speed stability' is the only anti-stall feature left in ALT(1). Is is a 'soft' resistance to stalling, overridable by the pilot. It is lost in ALT(2) in case of failure of 2 ADRs.

RR_NDB

mm43

It sounds very good for short duration AS issues at cruise.

Other (even less precise redundant method) AS information could "help".

henra

Can we agree that if the Pilots would have followed this advise it is very highly likely that we would know nothing about an air France Flight 447 today ?!

Re your first paragraph:
I still haven't seen any evidence/indication that 'HAL' has significantly contributed to the entry into the unfortunate attitude of the flight that was not commanded for by the Pilots.
Even the THS moving NU I would consider commanded to some extent as thirty seconds full NU Elevator indicates the desire for more NU and thereby for support by the trim. If you consider the response of the AC sufficient you won't try to pull max aft for 30s (that's an eternity, when was the last time you pulled just 5s or 10s full aft ???).
This to me indicates to some extent they wanted more NU.
And asked for that.
And got that.
Tragically.

Safety Concerns

I hate to disappoint you but the sales talk about a self diagnosing aircraft just doesn't hold up. Very basically now just for pilots:

you have a wire that goes from A to B. Along the way the wire connects to components within a system. Quite often when 2 components attached to the wire lose contact with each other a fault message will be sent.

the message will name the 2 components that are unable to communicate and quite often the message will add wrg: (wiring).

This is not a statement of a definate wiring fault just a statement that a very basic fault detection system can't work out who is at fault

RR_NDB

DozyWannabe "Redundancy is "powerful" when critical elements do not fail simultaneously. And UAS cases show clearly simultaneous "failing" (due product limitation) Simultaneous "failure" of critical elements should be reported immediately.

To the crew, and in Real Time.

Now, after clarifying to you i would like to repeat the Question (yet posted earlier): Why they put this redundancy? For what reason? What benefit?
When reviewing earlier posts i found this and i am now clarifying to you what i said.

RR_NDB

Safety ConcernsAnd in your opinion what this msg could indicate? The FDR is "not designed" for this type of data. So, BEA and Airbus SAS could never understand what caused it?

Safety Concerns

You can't determine what it means without having performed some form of diagnosis on the affected system.

For example an A320 apu starting circuit has a 400 ampere fuse. When this fuse blows the fault messages do not refer to it because the system is unable to determine where the loss of voltage occurred. The system often quotes the next relay in the chain as defective. So without the fuse its difficult to determine what was actually at fault.

These mesages are just a rough guide for maintenance as to where to look and also indicate that it isn't a good idea to use the mel based upon low intelligence fault messages

RR_NDB

Another "great feature" of "advanced planes". In some situations you may not be capable to even imagine what happened when analyzing a "mysterious" case.

I hope we will be able to understand the reasons for the "persistent" NU from PF.

A reason to "rethink" the FDR concept. Probably it currently lacks the capability to provide the required details perhaps not allowing definite conclusions.



PS

Probably Safety boards have:

1) All required resources to put responsibilities on the crew.
2) Not enough (resources) to verify why the crew made some actions.*

A perfect organizational "configuration" to conclude "crew error". And in this case, something that seems going to happen based in what BEA and Leaks yet emphasized.

(*) RHS not recorded may "help"?

Chris Scott

Quotes

jcjeant:
"Think ... maximum deflection to the left (for 30 seconds .. what a very long time) and the aircraft would not embark on a sharp left turn (or bank angle) ... very strange ..
What can prevent him from taking a left bank angle ?
Remember that it is to counter a bank angle to the right ..
So the force pushing the plane to a bank angle to the right should be very strong !!!
Wat can be this force ? .. what can produce this phenomena ?"
HN39:
"A stall ...?"
PA_18_151:
"The question jcjeant should really be asking is why was the PF trying to use aileron to counter wing drop."

Absolutely, but jcjeant's question is a perfectly understandable one from a non-pilot. For the benefit of the others, let me explain that, when we do our basic flying training on conventional light aircraft at a flying club or wherever, our ability to recover from the approach to the stall and the stall itself has to be demonstrated before we can be sent off solo. It is drummed into us that the use of aileron to keep the wings level approaching or in the stall is an absolute no-no. Roll must be countered with rudder.

Many light aircraft are fairly forgiving in that respect, perhaps because the wing/body angle washes out towards the tip-mounted ailerons. Others will react by rolling against the pilot input as the downgoing aileron stalls, usually leading to a spin. This is the classic trap from the early years of aviation.

In jet transports, when practising the approach to the stall, use of aileron (and associated roll-spoiler) is normally permitted. At the point in question, AF447 was already in a super-stall, although we have to assume that the PF had not diagnosed it. Quite what effect this was having on lateral stability is unclear, bearing in mind that the PF had been working hard on lateral control long before the aircraft stalled. Now, the whole wing was super-stalled.

I am reliably informed that, in addition to the fact that the aileron inputs might have been counterproductive (we don't really know), the associated roll-spoilers would have had little or no effect. Well aft on the wing chord, I guess they would have been in relatively stagnant air.

HazelNuts39

Chris Scott;

Just to add a few elements from the Upset Recovery Training Aid to your as usual excellent explanation:

The use of rudder to correct roll is not advised on large jet transports. The correct stall recovery procedure (not to be confused with reversing an approach to the stall) is:
1. Pitch down to reduce AoA in order to 'unstall' the airplane, then
2. Roll wings level using ailerons, and then
3. Add thrust to regain original airspeed and altitude.

Machinbird

Quote:
Originally Posted by BOAC

Henra
Henra,
If it had been Alt 1 law yes. Unfortunately this was in Alt 2 law. Nothing is leveling the wings in Alt 2 law but the pilot. There are indications that there was either a wing heavy condition or turbulence causing roll inputs to be required to avoid an extreme roll attitude.

From the BEA note

The aircraft was not yet stalled while the above was happening.
If we knew the reason for the nose up pitch inputs, we would be well on the way to understanding the accident. IMHO, inadvertent pitch input while controlling the roll axis seems to be the most likely cause.
Alt 2 is a funny law. You have to stay off the pitch axis, but fly the roll axis. That does not seem trivial to me.

wallybird7

JCLEANT

What can prevent him from taking a left bank angle ?
Remember that it is to counter a bank angle to the right ..
So the force pushing the plane to a bank angle to the right should be very strong !!!
Wat can be this force ? .. what can produce this phenomena ?

ans: Strong and violent updrafts in a thunderstorm!

OK465

In swept wing aircraft at high AOA, "aileron drag" or "adverse yaw" is due to increased induced drag (as well as parasite drag) from the down going aileron causing yaw opposite the roll input. Not because the aileron is stalled.

I flew the "poster child" aircraft which exhibited this nasty tendency, the F-100 Super Sabre. At high AOA it was a "rudder" aircraft for roll (if you felt lucky, you could "tweek" in some opposite aileron to assist the rudder).

The F-4 was also prone to this.

Conversely, when unloaded or pushing over establishing a lower AOA, or simply at a lower AOA, in any swept wing machine, the rudder is relatively ineffective for roll anyway, and ailerons must be used for roll control.

Rudders can roll a swept wing aircraft at high AOA, because of the large differential in lift being produced by the wing being yawed forward versus the wing yawing aft relative to the vertical axis which results in rolling. Straight wing aircraft just don't exhibit this to any great degree.

At 0 AOA in a swept winger, you get no roll from the rudder, only yaw. At negative AOA, you can get "proverse" yaw from the up going aileron.

Post stall, swept wing jet transport aircraft can also exhibit "adverse yaw" to varying degrees with aileron/spoiler inputs. Generally it is minor, but it is disconcerting to be commanding roll in a particular direction and not be getting what you think you should get.

*********************************************

In the A330, in ALT1 or Normal, nothing is leveling the wings but the pilot either. The wings are not automatically leveled when you release the SS input in either ALT1 or Normal laws. If the bank is more than 33 degrees in either law, ALT1 or Normal, true, the aircraft will return to 33 degrees of bank if you "let go da SS".

In ALT2, while hand flying pitch, the A330 is not the "squirrel" it is being portrayed as. Nor in roll.

henra

Agreed we do not know how violent the right bank tendency was.
Looking at the fact that even deep, deep in the stall, at AoA > 45° the roll attitude always appeared to be relatively straight and controllable (for that AoA). This does not point to a dramatical static imbalance. Nor does it point towards a tendancy of the aircraft to roll dramatically in general.
Therefore we will have to wait for more details to judge if aileron input was required to correct this or if it would have been rather a case for rudder trim or even waiting, riding out the turbulence.
We simply do not know the rate and duration of the beginning roll attitude.

On the other hand I agree that a cross control effect from correcting roll attitude is a prime suspect for triggering the initial Nose Up command.

(The subsequent NU's are different, I mentioned already my somewhat favourite theory of somatogravic illusion of falling when reducing the RoC from 7k fpm to 700 fpm close to the stall speed.)

What could have caused the roll excursion:
- Fuel imbalance:
Not very likely, would likely have caused severe Wing Drop after the stall, putting the Aircraft probably either Upside Down or even more likely in a spin as consequence of continuous Left aileron at high AoA. Moreover we have no indication of a pump failure or some such.
- Rudder Trim:
Cannot be completely ruled out. Although there is no mention of Pilots correctiong rudder trim. Would only make sense in combination with asymmetric thrust. Apart from this it is still unclear to me, what would cause this.
- Asymmetric Thrust:
AT should prevent that. Even after disconnect, behaviour should be rather symmetric. I would consider it not extremely likely.
- Aileron mistrim by AP to correct for wind shear, lateral air movement:
Could be a potential reason but it would be quite a concidence if the conditions would just change at the same time as the AP quits.
Ányway amount of roll should be small in this case (< 5° bank)
- Turbulence / Up-/Downdraft:
Looking at the general environment where AF447 was traveling and the relative unlikelyhood of the previously mentioned causes this is probably by far the most likely reason.

Did I forget a reasonable case?

If it was indeed the consequence of turbulence I would assume that even when not touching the SS for 20s, you would still find the aircraft in a reasonable attitude. Not perfectly S/L but probably also not in > 30° bank (Yes Alt2 would not prevent exceding this, I conclude this rather from the general behaviour of the AC in roll- might be wrong though).
With Autotrim trimming for 1g you would even not start entering a spiral dive by doing so. But we will have to wait the next BEA report for details which would confirm or contradict this assumption.

Mr Optimistic

As a non-pilot, my feeling is that this is more about the effective presentation of information to people under stress and psychology than about software or the absence of aoa readouts.

PJ2

RR_NDB;
You say that the industry should, "'rethink' the FDR concept [/I]".

On the contrary, the industry has come a very long way in flight data recording.

I'm genuinely interested in any push to improve flight data recording, (but with CNN around, not video recording) because the present legal standard in Europe, the United States, Canada and I believe Australia is the recording of 88 parameters for aircraft of the A330's certification date. For those certified prior to 1991 (October to be exact), only 17 or 18 parameters are legally required and if the aircraft is modified with a DFDAU [Digitial Flight Data Aquisition Unit], 22 parameters.

Any FDR & QAR data gathered above this regulatory minimum is voluntary. That should give some sense of the extent to which flight data is valued and gathered for investigation and accident prevention in FOQA Programs.

Nevertheless, many working in the industry have been battling this lack of data for decades, but with little success in improving the regulations even for the protection of data that is required to be recorded.

To some, data is indeed "inconvenient" because, in an accident the data can illuminate failures somewhere in the system. The will to deny unwelcome flight data is powerful among those who are commercially-motivated and such short-term thinking does drive the industry to a certain extent.

If the airline lobby (through IATA) is unified at all, it is not around sustaining such "mystery" but more ordinarily around the very high cost of such systems which will always include the costs of retrofitting aircraft to exceed the legal minimums. Where STCs [Supplemental Type Certificates...the legal authority that permits any aircraft to be modified), alone are unbelievably expensive, time consuming and resource-hungry, for smaller organizations.

That many current transport aircraft record over 1000 parameters is a testimony to the digitial revolution and the both the manufacturer's and the airline industry's established intention to monitor their aircraft, primarily aircraft systems but operationally as well.

FOQA Programs, now regulated with certain FAA protections for the collected data in the US but not in Europe, Canada or Australia, use QARs which can be programmed to record many thousands more parameters and at higher sample rates than available in FDRs and certainly exceeding the legal minimums.

The AAIB document, "CAP 731" is well worth examining, especially Appendix B which describes a bit of the complex recording process.

The BEA may have already used the QAR and other EEPROMs from the FMGECs, FCPCs, FCSCs mounted on the main equipment rack, "880VU" and since recovered. Hopefully they will have some comment on this recovery.

To return to your point, the fact that the #2 CAS is not a recorded parameter has nothing to do with a mysterious absence facilitating a 'mystery' and plausible deniability. Were that the case, many other parameters which hold equal importance would also be "missing". There are approximately 1300 parameters on AF 447's SSFDR. It is unfortunate that the #2 CAS parameter is missing, but it just isn't logical or even reasonable to believe that someone or some organization didn't want the parameter to be in the data.

One manual pitch-up theory says that the PF was pulling back in response to a high-speed indication on his PFD.

My own view is, the #2 CAS was almost certainly, roughly the same as the #1 and the ISIS-displayed speeds. Here is why I believe this:

Double or triple independent and concurrent system failures are extremely improbable. So first, let us set aside double or triple failures in other sensors and/or computers independent of the pitot failure and not merely a cockpit effect of upstream data or component failure causing an instant increase in CAS readings just on the #2 PFD.

By design the FMGEC > FCPCs cannot trigger the High Speed Protection Law or the Alpha Protection Law in Alternate Laws 1 or 2. And there is only one case where pilot ND response on the sidestick is inhibited and that is the High Speed Protection in Normal Law, and only until the speed is reduced below VMO +4kts, (IIRC). All other "protection" pitch-ups may be countered by the PF. So if the pitch-ups were executed by "rogue software/hardware", the side-stick should be effective in reducing the pitch and returning the aircraft to level flight.

This leaves us free to examine pitot failures in combination with the plausibility of an airspeed increase just on the #2 PFD.

There is no pitot failure mode which will cause an increase in airspeed in level flight.

A blocked pitot inlet and open drain hole produces a drop in airspeed and a blocked pitot inlet and drain hole requires an increase in altitude to produce an increase in indicated airspeed.

The increase in such a circumstance is mild - from experience (B767) it is not a huge increase - 20kts, perhaps. At FL350, the cruise speed was roughly 271kts (M0.80) and MMO (M0.86) is approximately 295kts. Nobody is going to pitch a transport aircraft up beyond 10deg to control that kind of increase in speed.

Regardless of amount of increase, logically the increase in observed speed must occur before the actual pitch-up because, with the pitot and drain hole both blocked, there is no increase in speed in level flight and the pitch-up by the PF would not have occurred in response to increasing speed. Therefore the PF was not responding to an increase in observed CAS on his PFD.

The argument is merely reasonable but of course is not conclusive.

So your, and everyone's question regarding "Why the pitch-up?" is, I think, the only important one at the moment.

I think the PF was executing the initial memory items of the UAS drill rather than stabilizing the aircraft (maintaining pitch-and-power settings) while calling for the QRH UAS checklist for the book settings but that's just a theory among theories and we'll soon know what actually happened and why.