Elevators alone
 

From JD-EE

Um no - you miss my point :). Even if he had bent the side stick backward, and held it there, on many aircraft that would have only resulted in nose drop into a stall - then after an acceleration downward nose back up then drop into a stall and so on. He was a glider pilot - that's what they do - its what most aircraft do. However, this aircraft did not drop its nose it sat back on its tail and dropped in a nice smooth (call it something no one will complain about) balanced (?) stall. As I have said - can that be done in other aircraft?
If the trim wheel had been wound forward to a cruise position - would the result have been a nose drop regardless of elevator position?

Should stall recovery procedure include manually trimming nose down?

NO!!

Only some "genious" f..... engineers are able to produce such a mess.

Ian. Not even. The Pusher would not allow the break, NO STALL, that is the design consideration for the Pusher. Now to find out the 330 can MUSH, and sidestep the STALL completely is infuriating. This is not a magic. This is STOOGES Three stumbling , and cutting class in aero UA.

Hi,

I presume that requirement has been waived in consideration of the high-angle-of-attack protection in normal law.
Ok .. so nothing new added .. question not answered.
Be sure (for those who want an answer .. of course) that this question will be on the center of discussions and answered .. but in another room than this one
Personally I have a doubt about the implication of the different flight laws of A330 for be an (good argument) reason for a derogation of the stall alarm certification.
Certification "a la carte" ?
Seem's .. the "menu du jour" is better

What type of weather radar did they have, with or without multi scan?

A33Zab,

I've created this graph that could be used to illustrate your references in a perhaps easier fashion.

The graph is indicating the separation between Normal (Phase 1 in BEA Report) and Alternate Law (Phase 2 and 3 in BEA Report). It further separate the Alternate Law section in two, first corresponding to the Climb (Phase 2 in BEA Report) and second corresponding to the Stall and Stall Recovery (Phase 3 in the BEA Report)

Also the graphs illustrate well a number of observations, of which at least one were a surprise (to me):

1. During a large part of the climb - marked area 1 (blue rectangular area)
there is a continuation of a strong Normal Accelaration variation - perhaps the strongest in the graph, indicating the strong (est) Turbulence - that started during Normal Law.

2. The Elevators (resultant) move during this area (1) seems (if any) to be ND (the graph is slightly above -3 degree line). This seems to be in spite of the Pitch Commands mostly NU. - surprise Elevators don't move NU with Pitch Command NU Is this a consequence of a strong Updraft?
It coincides with Normal Acceleration above 1g - confirms A33Zab observation.

A more accurate, or higher granularity (DPI) graph may show this a lot better.

3. The Elevators started to move NU more significantly only in the last 2/3rd of the Climb period - marked area (2) on the graph. This comes later than the Pitch NU commands, and coincides at the beginning with the slowing of the Vertical Speed, all the way, until the Vertical Speed becomes Zero, soon after which the Stall state is entered - same surprise - why the delay in significant Elevator move?

4. The most significant Elevators NU move occurs during the beginning of the Stall, particularly during the Section Marked 4. The move It is followed by THS NU from -3 to -13 degrees. This coincides with a strong Pitch NU command, while the Vertical Speed (falling) goes from 5000 to 10000ft/min. - Normal acceleration is at approx constant level, of approx 1g.

Indeed.

The graph does not have the granularity to show this clearly.

Limiting the THS NU when Stall Started, would have helped with the THS never getting to max NU. as it has been discussed recently a couple of pages back.

http://farm7.static.flickr.com/6199/...ab604868_b.jpg

I'm not a pilot, but I have a hard time understanding why the stall warning did not operate below 60 Kts, unless the engineers just didn't see the forest for the trees. The logic seems to be that since the AOA values aren't valid below 60 Kts, there's no way to know whether the airplane's stalled.

But if the airspeed is below 60 kts, how could the airplane not be stalled? I guess you don't want a stall warning sitting at the gate, though surely there are other ways to prevent that.

First question, basic, and re: DIRECT, for now, so, ROLL.

The a/c shows ROLL(ED) left 10 degrees as well as ROLL input 10 degrees. Then ROLL(ED) RIGHT 10 degrees, with Roll input 10 degrees.

What am I missing, I know the a/c is responsive in ROLL axis, but congruent position with input? Should we not see some displacement in response?

Next. The accel and vert speed are as expected, but the a/c does not gain altitude until the Vertical speed and accels go negative. She got an initial serious BUMP in accel, increase in VS, but did not climb until after they subsided. Que? There is a lengthy <1 accel, so the "feel" would be as though "falling" almost all the way UP. This explains (perhaps) why, if PF felt the "drop" (though climbing) he maintained (increased Back stick? It also may explain why the Elevators (and THS) rolled in via FCS to recapture 1g?

During this exact time, the PITCH ATT. is decreasing, along with the accels and the VS. Here is where he may have started sensing "Overspeed"? Nose dropping, (negative) accel, and rapid decrease in VS. ("I am descending")

All the while climbing. Wish I knew how to index the CVR with this litany of confusing prompts. ;)

HN39, Thanks for that explanation. Sorry to put you through that.

Just that Airbus appears to be using C* control logic, not a pure Nz Law.
As a C* aircraft slows, it switches from Nz to pitch rate over a small range of airspeed assuming I understand C* correctly.

There are questions on my end how Airbus handles invalid ADR inputs, and I'm guessing in that case it goes with pure Nz. But later as airspeed begins to return before the stall, it should transition the aircraft to pitch rate. The A320 transition speed is supposed to be around 210 knots. Probably not much different in the A330.
Zero pitch input on the stick below the transition speed should then result in a zero pitch rate (i.e. an aircraft trying to hold an attitude). Take a look at C* in this article:
Fly-By-Wire A Primer for Aviation Accident Investigators

Later in the deep stall with airspeed invalid, it should again be a Nz aircraft if my guess is right, except a substantial portion of the "lift" is no longer generated by the wing but instead by the fuselage drag.

Hi Ian_W,

It will do very shortly once this procedure is adopted. (Produced in May 2011)
The crew must be able to identify the symptoms of a stall, else it would be of no use.
http://fucampagne2008.unblog.fr/file...lprocedure.pdf

Generic Stall Recovery Procedure.

"2.
a) Nose down pitch control… Apply until out of stall (no longer have stall indications)
b) Nose down pitch trim…….………………………………………..As needed

Rationale:
a) The priority is reducing the angle of attack.
There have been numerous situations where flight crews did not prioritize this and instead prioritized power and maintaining altitude. This will also address autopilot induced full back trim.
b) If the control column does not provide the needed response, stabilizer trim may be necessary. However, excessive use of trim can aggravate the condition, or may result in loss of control or in high structural loads."

Lone, "Any pilot would do that, if he or she recognized being in a stall."

Um, on the AF447 flight was there even one pilot if we take that as a definition?


I harp what feels like too much about the "training? TRAINING? I don't need no (censored :uhoh::\ censored) training!" culture that seemed to have evolved around the AirBus FBW systems, at least at Air France. (OK OK, I exaggerated for effect but....)

(By the way - would folks commenting to lyman please note this at the start of their message so I can safely ignore it. It'd speed up reading.)

Updrafts and controls
 

The increase of normal acceleration between 02:10:07 and 02:10:15 is strongly correlated with the increase in pitch attitude in the same period. The increase in pitch was in response to elevator movement (quite sensitive at M=0.8) which in turn corresponded to sidestick command - see BEA#3 top of page 41 (english version). [Sentence DELETED. Reason: The graph on page 42 does show the elevator angle]

In an encounter with an updraft the airplane would also experience an increase of normal acceleration, but it would not pitch up in this manner.

One other aspect of that concept of weather concern to figure in is his glider training. Unless you're landing altitude is a good thing until you run out of air, O2, or warmth.

Presuming we know all the speech type noises in the cockpit, you're dead right there's only the thinnest of suppositions there. (And I don't know why BEA might skip anything.)

Anecdote warning... I had a good physics instructor in college. By the fourth semester of honors science we'd winnowed the class down to maybe 30 people in a massive large lecture hall. He'd spend an hour on course material. Then he invited us to stay afterwards for his ramblings, never to appear in any tests. We all stayed. He was feeding a savage hunger to learn, at least in my case. I learned as much that way as I learned of the physics being taught.

Anyway - my thought is that there should be some periods after flight school for passing along hanger flying stories in ways they'd stick in pilots heads. This might make a good, "Nobody knows if this really happened or not but..." sort of story. It'd be an anecdote about losing focus on the real problem, avoiding stall, rather than the wrong problem. And the PF really did lose it big time with his remarks about the extreme wind noise possibly being over speed.

He reacted before he thought.

Stall Recovery Procedure
 

The following Airbus Safety Magazine - Safety First - January 2011 link has a fully illustrated description of the new common Stall Recovery Procedure.

Lone & HN, "If you feel that the language used in JE-EE's musing, or speculation, is over the top I understand."

It was rather intentionally over the top to beat its way into discussion. The resources at hand appeared to have been ignored and the PF and PNF seemed to be in different worlds. That's why I grinned like a manic monkey and dreamed up the pizza metaphor. (Um, won't cause injury except to ego and cleaning bills, will get the attention, and won't damage the plane. So that leaves out things like nerf ball guns, paper planes, and clipboard edges.)

I've developed a bad habit of dealing with one chronic problem I have using graveyard humor. If I can laugh I feel better - and if I feel better I think better. I was hoping the image was ridiculous enough to telegraph the over the top intent. Get them communicating. Get them paying attention to data they had in the cockpit. Question why some data was not present. And perhaps the most serious issue that is very human, focusing too tightly - forest and trees metaphors are traditional here.

(And, of course, way too little hands on crisis management training.)

HN - I'm just groping for an alternate explanation for his pulling on the stick and climbing the way he did. Maybe he was trying the low level stall warning recovery. But with the storm vividly on their minds enough to mention it in that taciturn cockpit setting, suggests he wanted to be up higher if he could get there. Would that have filtered into his arm's motion when he took control and pulled 8 degrees on the stick? "Maybe" His response was so quick it was almost an instinctive reaction.

Chu Chu, it appears the plane's speed stayed above 100 kts for its vertical component. The horizontal component was maybe half that.. The probe's angle was so high it was reading artificially low compared to the 120-130 kts the plane was actually traveling. The AoA probes had more than enough wind speed over them to work properly.

Hi
Spot on. ALTER is the union, in fact. BSPN is its publication. The rest is perfectly analysed.

Mhmmm, OK. Perhaps, indeed ALTER implyied that. I don't think it's true either, AFAIK the crew did not recognize the stall (so no reason to try to apply a related procedure).
Will have to try to correlate what's in this old procedure and the FDR traces, but I'm pretty sure I'll find far more differences to agree with ALTER's views.

I'm very interested to know if you can find something : please let us know :)
Will do the same, of course :)

Uh? Can't remember having tried that much insane things... :E

Simple : Speed indication failure.
Let's imagine... Ash cloud or severe pitot icing. Speed drop to < 60. Is your aircraft stalled? No, only wrong speed indication. Does the stall warning sounds? Yes if we follow your idea. Then crew applies stall recovery procedure => nose low => speed increase => overspeed. Miiiiiip bad idea.
Never the less, I do agree with you that cutting off the stall warning under (sensed/indicated) 60kt in AF447 was not good. But I didn't find how to do better, for now.
As JD-EE noted, the real airspeed was more in the 100kt range. Pivotable pitots (to let them follow the airflow) would have helped, there. Those exist on the Rafale fighter (and perhaps others), as shown on this pic.
By the way, does anybody know why such combined probes (pitot+AoA) seem to be rare? Not so easy to make/maintain? Too expensive? Never tought about? Not deemed useful on a liner, which should not reach such exotic AoAs?

That's more simple: check aircraft on ground (sensors on the undercarriage, already there for other purposes, like auto-brake, spoiler deployment...) = no stall warning.

New stall recovery procedure
 

Thank s very much mm43.

Now that is really new?
That´s what worked in the fiftieth already, how come they have to invent it again 60 years later?

At least they didnt need to look it up in fiftieth documentation, to look in the 2004 reference stall ecovery zip was probably enough.