macdee re: tight turn to starboard
 

The graphic you refer to in the BEA update of 27 May shows that the turns to starboard occurred after the plane had stalled and had already descended more than 3,000 feet, at which time "The airplane was subject to roll oscillations that sometimes reached 40 degrees." The turns also appear to have started shortly after the captain re-entered the cockpit, which may or may not be a coincidence.
Between disengagement of the autopilot (point 4 on the BEA graphic) and stalling (about 15 seconds after point 5), the plane shows no deviation in course.

cwatters writes
With an AoA of greater than 45 degrees, neither the wings nor the horizontal stabiliser will be acting as aerofoils, but simply as drag surfaces. Therefore it is unclear what the effect of any stick input would be.

It is certainly difficult to visualise how there would be any roll authority, and pitch authority could conceivably be the reverse of normal. Therefore theorising a fault with the stick seems a step too far.

@ mm43 (post #1567) : Thanks for pointing the FlightGlobal article. Wasn't aware of it, and based my hypothesis on the sole FCOM.
Thanks also to A33Zab :D

Thanks GraemeO, just wondered.

Can't be ruled out, just like SS armrest failure.
You need a proper adjusted and fixed base to perform coordinated SS inputs.

In such case PNF could take control but what if you are dealing with an emergency situation and fighting to get control while PNF is in the middle of QRH? Would one change roles?
All they had was only 210 seconds.

True.

But if one cannot be sure of the effects of a correctly operating stick in extreme AoA conditions how can one begin to conjecture on the effects of a faulty one?

Hi,

In the world of pilots of airliners ... is what the word panic is forbidden to use?
Do airline pilots are men with capabilities far superior to other men and therefore they would not be subject to the phenomenon of panic that often occurs in situations of great stress .. even among men the most seasoned
Panic is known to decrease the power of reason in a logical manner.
Note that the panic .. when it is externalized can be contagious (can influence people in contact with the person who panics)
Panic .. Perhaps an explanation as any other to explain the behavior of the PF and the sequence of events.
This can't be ruled out ... so far.

Not a bus driver.

There must be 500 pages on all the threads by now and I'm buggered if I'm going to read them all. Can someone bring me up to speed on this quickly with definitive answers, I'm trying to get my head round this:

The stall warning is inhibited below 60kts IAS, as per my understanding.

1. Does WoW factor in this decision?

2. Someone please assure me the logic is sane enough NOT to inhibit the stall warning at any IAS in the event of AIRSPEED UNRELIABLE?!

Thanks

Hi Ranger One,

Sorry - I can't assure you of that despite trawling through FCOM. It seems ground/flight logic has not been included with the Stall Warning - just the 60kt switch.

Contributions from A33Z - attaboy
 

I think all of us here appreciate the inputs from A33Z concerning more detail of the control laws, etc. So ATTABOY!

Some of us are still not clear what is taking place with the "trim", however. My understanding is that the 'bus attempts to achieve one gee ( roll and pitch corrected gear up, not sure when gear and high lift devices deployed). In other words, let go of the stick and the system tries to get to one gee. Pilots can't 'trim" for a gee, only command something other than one gee. So only reason I can see for the THS to move is to allow for dynamic pressure changes ( speed) that limit the elevator's effectiveness and/or allow it to operate from a "neutral" position versus an extreme angle. In other words, the THS doesn't move to achieve anything other than one gee with hands-off stick position. So I can see that as speed decreases or increases, the THS would have to change in order that the pilot gee command would be the same amount of stick displacement, regardless of speed.

Am I close?

Eee by gum Gums. That's so close - I can't spot the error.

gums

As in, trimming for stable climb rate? Some version of 'cruise' climb? As energy depletes, the THS increases Pitch Up to maintain climb? Something in this climb suggests a maneuver that the pilot felt was OK? I am lost how "trim" has any place in what one would think an emergency maneuver? The BEA say, "FROM 2:10:05, ".... They leave out a sequence of pilot displays. What information was the PF operating on to continue this climb draining energy merely to gain altitude? Was he trying to dump overspeed? Or was this THS movement unknown to him, he believing he had plenty of energy, and felt his AoA was reasonable? Had he thought he was merely initiating a "Little bit to the left" turn? Adding just a bit of NU that the THS decided to chase? When did he Know he was without A/THR, and auto trim wouldn't mess him up? It was eleven seconds after the a/p that both of them made note of the sitch.

Halp!

First of all, AIRSPEED UNRELIABLE may only be, indeed probably is, "crew assessed" not aircraft (if the aircraft assesses it it will "deal" with it).

Next, even if the crew do assess it, they do not, and cannot "tell" the aircraft.

So now we have a stall warning, driven by a airflow "vane". When would you like the stall warning to go off? Even at airspeeds below which the vane can work?? These vanes are substantial pieces of metal that can withstand the best part of 500K/1.0M. They are not little feathers that will read reliably at 20K IAS :ugh:

Your post expresses indignation at a warning that was not present, when maybe it should have been. The designers also have to consider not presenting warnings when they are not valid - indeed "Stall" is one of the highest level wanrings in the Airbus, and the consequence of an incorrectly presented one is potentially very hazardous.

The airframe / software designers have to work to some parameters. Knocking off AoA interpretation below 60K IAS seems valid enough to me - and also at that sort of airspeed you are not just "stalled" in the normal sense of the word - you are a falling leaf. I cannot see elevators having any effect at 50KIAS. The idea of the Stall Warning, and recovery actions, is to never get anywhere near that situation.

I am not relating the above to the accident as much as we know it, just the facts in the posts above.

NoD:On the other hand there is a huge difference between inhibiting initiating a stall warning at <60Kts and arbitrarily -stopping- an existing stall warning when the (sensed) speed goes from 60 to 59 Kts and then even worse
-resuming- the alarm just as corrective action is taken (nose down) and the sensed speed rises above 59Kts.

Do totally agree with you that consequences of screwing up the stall alarm can be very bad.

Hi NoD,

If I'm airborne - I'd like the warning to be available all the time please, especially with blocked up pitots showing Zero IAS.

not as hazardous as the warning being inhibited when stalled - apparently.

I agree. But if you walked onto the flight deck and found your colleagues were descending like AF 447 - do you think you might have had a better chance if the stall warning was still sounding?

The fully trimmed up elevator was doing a very good job at holding that trimmed attitude. Nose down input accelerated the aircraft - but the stall warning started again - which must have been very confusing. The real airflow air speed was much greater than 50 kts - it's the angle of attack at the entrance to the pitot which made them show so slow.

The BEA report places emphasis on the point that the unreliable airspeed lasted for no more than a minute. However it was in that minute of understandable confusion that the major upset occurred.

The big question is whether the flight crew where then already locked in a trap of no escape or did they miss a trick? As this is so far beyond the envelope of simulators, it is a question only an Airbus aerodynamicist could answer with any authority.

I cannot help feeling that the strong implication of inappropriate control input (in the later stages of the incident) from the PF may be misplaced if pitch and roll authority were all but gone and we know so little about the response he was getting to his inputs.

Don't Hang Up

Not only that (inappropriate inputs later on), who is to say that the pilot's initial (NoseUp, Left Roll) were anything but proper? The BEA has not expanded on the most important maneuver, the first one, flown after loss of AutoPilot.

He may have been under the mistaken impression that he could pull, roll, (for gums, bank and yank) with impunity, the a/c would PROT itself. Also, his climb (whether intended or no) may have been instigated whilst he was confident in AutoThrust. It was at least eleven seconds after A/P loss that the pilots acknowledged, "lost speeds, Alternate Law", (which Alternate Law? Did they know?).

A loss of autopilot does not automatically drop the a/c into ALTLAW. Loss of speeds does. A vanilla loss of A/P leaves the a/c in NORMAL LAW, full protections. Were the crew instantly aware of ALT LAW? NO, they acknowledged it eleven sedconds later. This is important.

One cannot merely assume that BEA released its sparse product in good faith. Given the Gallows attitude of the Press, was it not somewhat expected that "Blame" might be appropriated? Something their mission prohibits?.

Whether conscious or not, the BEA have some explaining to do, re: the entry into the Public record of such paltry data.

DHU

Strictly speaking, both are still airfoils, producing both lift and drag as long as there is relative wind, which there was. Control surface movement will still alter their lift and drag albeit with significantly less effectiveness approaching inadequate, particularly with all the nose-up stabilizer trim introduced. For example, recovery may require that the engines be brought to idle to minimize their contribution to the nose-up pitching moment, certainly not an intuitive response when descending out of the assigned cruising altitude. Lowering wing flaps (and landing gear) should provide a nose-down pitching moment. In any event, nose-down elevator shouldn't have been necessary for recovery. Absent a too rearward cg, T-tail, TOGA thrust from under-slung engines, and/or nose-up control input, a conventional airplane trimmed for level flight should be pitching nose down even at this angle of attack.

OK - even when the spec/manufacturer of the vane says the IAS is way below the validity of the vane's spec? I'd rather not...

Please go back to your basic training... if you are a pilot? The Stall Warning is a symptom (one of) the incipient stall - a "warning sign" as such. You react to it, but preferably prior to it, by flying the aircraft away from the stall. You, I hope, do not, when the warning goes away, say "oh that's all right then".

Cannot see where you get that from at all? We have no analysis from the BEA, and if the stall warning went off once, if only briefly, it brought the crew's attention to the situation and they should have been 100% sorting that out.

If we have go airline crews who recover from a stall purely in reaction to, and only while, the warning is going off, and consider recovery complete when it silences, we have a bigger problem that I thought :eek: NB this is not saying this what the AF crew did, just a reaction to your post.

No idea... but that is not relevant. We need to design as robust and reliable a system as experience shows we need - not redesign whole systems based on a release (unanalysed) of some data on one accident that will, inevitably, have a multitude of causal factors and recommendations.

But she was trimming for climb, automatically, so much so that the a/c stalled, having lost autothrust and assumedly, "Stall" (aPROT). So this is the deadly combination. A/P drops, a/c climbs, and Stalls. Thirty seconds over Oraro.

This time grab can be parsed at this point to weight a conclusion of wildly general and biased opinion. What happened after the Stall is extremely interesting, but as has been pointed out, minimally important. This a/c should not have Stalled, ever.

For instance, the initial maneuver was accompanied by two separate and brief Stall Warnings. What was the AoA the a/c had trimmed in at a/p drop? What was her Speed? Local airflow? Is ICE absolutely necessary, Willem Occam?

Nigel:

Well there was certainly an ADR DISAGREE... but I think you're right, thinking back to earlier discussion perhaps there isn't a specific ECAM for AIRSPEED UNRELIABLE - but the system knew there was a problem.

Yes absolutely, the last thing you want is nuisance stall warnings, I couldn't agree more - when you have *reliable* data. But please, if there's any serious dubiety about the data, I would doubt the wisdom of shutting off the stall warning based on IAS data that's known to be dubious...

Hi NoD,

If you had blocked pitots at FL 350 with IAS showing less than 60 kts - are you seriously telling me that you'd rather not have the stall warning system active?
Are you a pilot?

bearfoil:

I suppose it's a matter of opinion, but you do have a tendency to be a little inciteful...

"She was trimming for climb, automatically..." --- really? I can't find anything that states that is true.

Any THS trim effect (due to PF input) did not ramp in until nearly 1 minute (46 seconds) into the reported sequence of events. There is nothing that indicates anything other than demanded trim (PF nose-up inputs).

Care to explain it differently?

According to this article the french goverment has already assigned responsibility for the accident:

Rio-Paris : à qui la faute ? - L'EXPRESS

Another "leak" just in time for Paris Air Show (*)
 

Another (:suspect:) serious leak.

In this "war" the truth may be the victim...:sad:

(*) June 20-26, 2011 (12 days until the show)

Question: Who´s in control? :}

OK, I admit it - I'm confused as to the conditioning for the stall warning generation.

According to the information I have, the following is the conditioning for Stall Warning:

(my bold above)

The result of the warning is triggering of the cricket stall tone warning AND the sythetic voice "Stall" warning.

Can anyone expand on what happens in the case of unreliable airspeed with respect to the stall warning?

Garage Years

Yes. The point is that the pilot's first response to loss of auto pilot (not autoflight), may have been input while he's thinking the drop was into Normal Law, as would be the case if the auto pilot was lost due to its inability to trim within its own parameters. BEA is vague (eg: "From 2:10:05"), attributing the following displays and aspects without a specific time stamp. So if the pilot pulls the stick back, or, more importantly, it is deflected into a climbing Pitch inadvertently, then the continuous increase in Pitch, is inadvertent, and the increase in Pitch IS automatic, since the pilot is not expecting a climb, but the THS is trimming for it, and since the pressure is continuous, the Plane gains Pitch attitude inadvertently.

The Stick is an immediate unknown to the Pilot who starts to fly, but hasn't had a chance to assess the a/c orientation, and existing trimout. Rolling left may have been the only input he wished to make, but the third quadrant of the stick's deflection disc could have been entered by mistake.

This is extremely important if the AutoPilot let go due to its mechanical and aerodynamic limits being reached, and not merely due to AD rejects? We do not know that the AD losses were concurrent with a/p dropout, only that BEA reports the events w/o times, except for a starting point, eg: 2:10:05.

I am saying that if ICE was not the cause of discrepant reads, but perhaps attitude, aspect, or local flow were, it would explain what happened even more simply than what is generally accepted, absent the record.

I honestly don't mean to incite. I also don't have any conclusions to offer. What I think is important, and may well remain important even after the Final, is that certainty may be impossible, and mere Likelihood may have to suffice.

apologies for (perhaps) overly assertive style.

edit for GY you say: ":...Can anyone expand on what happens in the case of unreliable airspeed with respect to the stall warning..."

If I read Nigel correctly: "The Aircraft deals with it"

"...First of all, AIRSPEED UNRELIABLE may only be, indeed probably is, "crew assessed" not aircraft (if the aircraft assesses it it will "deal" with it)...."

FCOM :

http://i25.servimg.com/u/f25/11/75/17/84/af447_13.png

CONF iture Quote:
Originally Posted by GY
Can anyone expand on what happens in the case of unreliable airspeed with respect to the stall warning?

FCOM :

http://i25.servimg.com/u/f25/11/75/17/84/af447_13.png


IOW, with a Stall alert (2 of them) as the immediate environment after a/p drop, what was the cause of AoA being "At Stall" ALREADY? Prior to PF's altering the a/c's controls positions?

IF IN NORMAL LAW THE STALL WARNINGS MAY HAVE BEEN FALSE.

IF IN ALTLAW THEY WERE GENUINE.

THIS SUGGESTS THAT THE A/C WAS IN NORMAL LAW AFTER A/P DROP, AND HENCE A/P LOSS COULD NOT BE RELATED TO LOSS OF SPEEDS, BUT ONLY ITS OWN INABILITY TO TRIM THE A/C. OR, THE AUTOPILOT HAD TRIMMED THE A/C INTO A STALL.

A point I have been trying to make for nearly two years. perhaps badly?

If true this suggests that the A/C had trimmed almost completely into Upset, then Stall, then LOC. It also makes Pitot ICE a remote event.

Pas de quois, Thales

Yes, absolutely. But I think the question GY was asking, and I know the question I was asking, related to the interaction between IAS data and the stall warning being disabled below 60kts.

If you have unreliable data (sources disagree more than trivially) or NO data (all pitots stuffed) I would hope the system is smart enough NOT to inhibit the stall warning, irrespective of the (erroneous) IAS!

GY earlier posted that stall warning was triggered depending on, amongst other things, F/CTL law. Is it the case that stall warning is NOT inhibited below 60kts IF you're flying anything other than normal law?

R1:
This bring up another data point that would be of interest:

Over the full range of IAS and other factors taken into account to adjust raw AoA value what is the magnitude of the adjustment.

Is there the min raw AoA value that means you are stalled no matter what the speed and other factors?

Point I am thinking of is the "tweaks" are needed in close to normall flight conditions but are they relevant/needed in extreme attitude, does it matter if you are stalled by 15 or 25 degrees?

Also somewhat related to difference between "stall warning" (approaching stall) and "you are stalled".

What you really mean might be that IAS and warnings of any kind at low speed might be involved here. Wind Shear comes to mind.

I think the twenty-one second interval to include ten seconds prior to a/p drop and eleven seconds after ("lost speeds", "alternate Law") contain the meat of the matter. If she was actually indicating 60 knots (less), what's the point in the post Stall discussion? The focus for me is on two handovers. One, the a/p to the PF flying (disregard his seat), and the degradation of Normal to Alternate (any one of them).

a/p loss: after the loss of autopilot, there were two Stall warnings. They cannot have been genuine, for the a/c commenced a climb without adding thrust, two things that would have caused a Stall post warning, if genuine,
they were not.

Stall warnings are genuine in ALaw, and these two were false, thus the a/c was not in ALaw after a/p drop. Having preceded airdata loss, the a/p deselected itself, which means the a/c remained in NLaw, with all prots as advertised.

In the short interval between "I have the controls" and rejection of two pairs of airdata, (eleven seconds), the PF knew he was in NLAW, and handled the a/c "Exactly as he was supposed to". Whatever the cause of Pitot sensing failure, it likely was not ICE, and it certainly was not precedent to a/p drop, since loss of reliable speeds and switch to ALaw was diagnosed (reported) after loss of autopilot.

A33Zab, thanks for the BUSS reminder.
In 2007, it appears that the BUSS was not limited in its operation above FL250 as it is now the case. Would you know when that change took place ?


What are the vane's spec ?
Would you like to ref a manufacturer documentation … ?
Because to me, by merely direct observation of those vanes, they are pieces of metal which behave like feather …


I like the question, and it reminds me the axiom #1 : Nothing is simple in the Airbus World.
But to answer the question, I would say no : Stall warning has been inhibited below 60kt and airplane was most likely in Alternate or Abnormal Law.

Hi,

Indeed the Office of the Secretary of State for Transport Mariani know :eek:

Vive Airbus ..
Que justice soit faite

Google Vertaling

Even more pure horse manure from the "l'Express"

Note that the BEA report states the indicated airspeeds in the initial problem "fell to 60kt". It appears that the control system does this so that any protection that may be disabled if <60kt are not. Is it correct that this speed is so far below that required for flight this isn't mentioned in any manuals?

At least 2 out of 3 speeds recovered before the actual stall, and assuming they stayed in synchrony it is quite likely they reported a valid Indicated airspeed below 60kt at high angles of attack/ sideslip. Pitot tubes are designed to be work without significant error over at least +/- 20 degrees AoA, but they do roll off sharply above +/25 degree, and if the nonlinearity is based on the amount of air directly impinging cos (60 degrees)=0.5

Thanks to all for serious and honest thoughts and research on those sad events.
What is lacking in the last BEA report ? Almost everything. We need, at least, during the fall of the aircraft, the different control laws coming up then probably going away. We need to know which computers went off and which stayed with the crew. We need to know more about the aerodynamics of the plane (if, let us say, knowing what happened, therefore ready to act, unlike the crew, you rolled full down manual trim and reduced thrust to idle, would it work ?)
Even with the FCOM and some experience and technical knowledge, I am sure we cannot really know how for example the stall warning works. Only the engineers at Airbus or the Maintenance, with detailed wiring diagrams, can dig in this particular situation. Sometimes Speed, sometimes AOA, sometimes both, this alarm manages to work (Perpignan) with blocked AOA sensors…I flew on planes when we had a three months qualification course (yes, three months !) so then we could say we really knew the plane (which was far simpler).
But then, there were flight engineers, god bless them. On the A 320 I had a three days course on the systems, and no on line instructor ever had the courage to ask me a question on them…I probably would have done no better than the crew, but my first reaction, surprised by this unlikely event, would have been…to do nothing. The less you touch the stick, at high altitude, the better the plane flies. This rather stupid remark (forgive me) does not solve the questions, but might be of great help to youngsters on those planes. Sit on your hands, watch, wait…

AOA
 

R1:

That's not mentioned in the books !!!

Vane resolver sin and cos is input for ADR, ADR generates AOA from this input if CAS =>60 Kts if not AOA=0 and NCD. This is ADR output fed into the PRIMS/SECS and FWC.



http://i474.photobucket.com/albums/r...robes-a330.jpg

More on AOA:

Aviation Occurrence Investigation AO-2008-070
VH-QPA


Complete report including detailed description of polling logic.

Autotrim
 

Hi rudderrudderat:
In Post 1407 I wrote:
and I got beat up about it. I'm not wild about the idea of autotrim for the reasons that gums mentioned earlier and I still think that autotrim should drop out whenever the system becomes confused. :ugh: If it had, things might have been simpler for the crew of AF447.

There is a situation that occurs in jet aircraft that is extremely unlikely in your Cessna. Jets can stall both at low speed and at high speed. In a high speed stall at high altitude a lot of pilots pull the nose up instinctively in order to reduce airspeed. This increases the load and exacerbates the stall. The correct procedure is to unload the wing and reduce the thrust to slow the aircraft back into normal speed regimes...