infrequentflyer789

Thanks for your response. Help me to understand how "relieving the elevators" does not acquire a massive amount of PITCH Authority in one or the other direction? The HS is acting as a lever to "neutralize" the effort required to deflect the elevators, ever producing more baseline AOI. Angle of Incidence in this sense is an artificial "neutrality". The FCOM alerts the operator to the THS preventing some PITCH authority above 180 knots? How is this alert benign to the efforts of PF447 recovering the a/c?

Yes, I think elevators alone would have been preferred. Once above a certain angle, and at >180 knots, with All Trim Fuel aft, recovering seems a tall challenge? In this sense, the elevators are "Trim Tabs", ineffectual at "boosting" the control surface (THS), to allow maneuvering. Instead, they have a range of effect that traps the a/c in fluctuating PITCH excursions. Neither direction is satisfying, to the extent that AoA is "never below" 30 degrees (BEA).

Not even Flare should top 17 degrees, yet here the a/c is at "cruise" and 16 degrees! Do we set out to castigate the pilot ("Je ne comprends rien").
Do we attack the machine? Why do either? The a/c in some way got away from her pilot. Some random misfortune is trying to tell us something. To the extent we make up our minds, we lose the opportunity to learn.

As an aside, you say "Elevators alone were sufficient to get her to 16 degrees and fully STALLED." In ALTERNATE LAW, isn't the airframe protected from such a thing?

Static Port Placement
 

grity: Interesting graph, Grity; thanks.

I should have explained that my surprise was not at the fore-aft placement of the static ports on the A330, but their placement so low on the fuselage, maybe only 30-40 degrees off vertical centerline.

By comparison, the DC-10 static ports are at least 60 degrees off the vertical. FWIW, they are also about 47 feet aft of the nose, beneath the mid-cabin doors. The low speed static correction is less than the 300 feet of the A330; more like 240 feet, IIRC.

Also noteworthy is the lack of an airspeed input to the Stall Warning on the DC-10. It is based on AOA sensor, modified by flap discretes.

CJ;
FWIW, I agree CJ - it is a mistake to read too much into each word in an attempt to squeeze this or that interpretation, (ours!), from it. Finding agreement when far more is known is going to be equally difficult as well.


Interesting post and link Machinbird, thank you. I don't disagree that a form of cognitive overload may have occurred. In keeping with this discussion, I would like to broach a human factors aspect about which thus far nothing has been noted or asked. The observation concerns SOPs, CRM and cockpit discipline in the handling of this event.

The creation, implementation, training and checking of SOPs, CRM processes and their outcomes, cockpit discipline, are intended to avoid or delay cognitive overload, or cognitive dissonance.

While cockpit design, cockpit displays, ergonomics, and aircraft/system drill and checklist design will influence outcomes, a "Standard" way of accomplishing complex tasks under both time and operational pressures reduces the potential for error. I know very well that it doesn't always work that way despite best efforts.

I would like to preface this observation by stating that this is not just about "the crew", but it is a question which will arise in every airline pilot's mind because SOPs, CRM and cockpit discipline are heavily emphasized in recurrent training and line checking.

This is not about finding blame because that is not what the investigative process is about. The investigative process is about "finding things out" and must be free to ask all questions regardless of how uncomfortable they may be. In other words, the investigative process is unlike ordinary discussion in that when a question is asked, it does not imply anything other than a need to know.

So asking a question about training does not imply that "training is under suspicion" nor does asking questions about the crew's actions imply that such are under "special scrutiny".

I would like that distinction to be as clear as possible. This has nothing to do with complementing or criticizing the crew. The notions of "defending the crew" or "criticizing the crew" don't apply in the investigative discourse. Criticizing or defending the aircraft may have a place in other venues and so may such an examination of the crew have a place, but that would be a legal, political, economic discourse, not the investigative discourse.

I make this distinction partly for the reason CJ posted his comments on parsing language - the same words will mean different things, depending upon the discourse. "Investigate" means different things in politics and law than it does in safety work.

That said...

The question of SOPs and standard responses to abnormalities arises out of the almost-immediate response just after 02:10:05 by the PF to the autopilot and autothrust disconnect and the ECAM messages. We know that the BEA Update states that almost immediately there was a left and nose-up SS input and pitch increased beyond 10 degrees.

Abnormal and Emergency SOPs are created and trained to:

a) ensure control of the aircraft is firmly in hand and that stable flight was being maintained and to ensure the flight path (track and altitude) was being maintained as needed;

b) ensure that all crew members on deck were alerted to the problem, and to who was in control of the aircraft and who was to execute drills and/or checklists - ATC would be notified when there was time;

c) ensure that all crew actions are predictable and understood by all cockpit crew members and to avoid "individual variations" leading to unexpected responses and confusion.

This is the standard Aviate, Navigate, Communicate rule that we have seen mentioned.

In terms of CRM and responding to abnormals and emergencies, no action, no drills, no checklists are started until the aircraft is under control and stable.

Most of the time that is a quick assessment and emergency drills are done without delay. For an abnormal such as this one, there is no hurry. The airplane was stable before the event and other than minor variations perhaps due to turbulence, would continue that way in the absence of input.

The pilot flying always initiates a drill or checklist by announcing the drill being executed or the requesting the appropriate checklist. This communication alerts everyone in the cockpit (usually just the two), to what is coming next and to prepare for response, as trained.

If a memorized drill (such as the Rapid Depressurization, Emergency Descent), the PNF monitors the drill for accuracy and correctness while the PF flies the manoeuver and where necessary calls out any deviations. If required (in a EGPWS manoeuvre for example or Stall Recovery), the PNF may call out primary parameters such as pitch, altitude, speed, rate of climb and so on. Here, such feedback would be of critical importance to help avoid "tunnel vision" and increase situational awareness for both crew members, and to help the PF who would be concentrating on flying the aircraft with a possibly-reduced scan.

As I have said, this wasn't an emergency such as a depressurization, an engine failure or fire or a GPWS event which requires timely, immediate action but it did have two memory components to the drill, (as shown in the BEA Interim Report #1), before the UAS QRH Checklist was to be called for. I have discussed this memorized drill and checklist a number of times on previous threads.

The next step in the SOPs for handling an abnormal is to execute and clear the ECAM messages. There are SOPs for this process. Once clearing all the ECAM messages is accomplished, the STATUS page is checked for equipment and capability lost. The ECAM is then cleared. Then, if still necessary, the "paper" checklist, (QRH) is called for by the PF and brought out by the PNF and referenced. Usually the QRH checklist most commonly used are the Landing Config/Approach Spd/Landing Distance Following Failures and Landing Distance Without Autobrake CONF 3 or CONFIG FULL checklists. These kinds of items might be delayed until closer to destination - I'm just using this as one example.

When all ECAM actions and Status reviews are complete and when the required QRH checklists are completed and when the aircraft is considered secured the process of communicating to ATC any problems with capability in terms of maintaining flight and navigation, then communicating with the Flight Attendants, then, where necessary, the company. Any diversions, changes in altitudes are considered at this time.

This is one example of how it's done and there will be variatons, slightly different priorities and more or less memorization of drills depending upon the airline.

The key point here is, this process must be thoroughly trained and checked during recurrent simulator and line checks. SOPs, CRM and cockpit discipline are for those times when things start come unglued in a hurry. The training focuses one and gives a structure to the inevitable initial chaos which can unfold rapidly as we have seen here; - at 02:08 they were talking about deviating to the left of course, and just over six minutes later they were gone.

The question of SOPs goes beyond the crew and must be asked "upstream" of just this crew. What support for these areas was there and how robustly was it carried out?

A more complete update from the BEA will hopefully shed some light on this aspect of this accident because it may be related to the reason for the pitch-up and stall.

The "Airbus Golden Rules" came out a very long time ago - around 1998 IIRC. The later 2004 document discusses the above process in greater detail and is well worth examining.

Hi Bear,

See you are back and posting once again. There was a very interesting post awhile back on the other AF447 thread in R&N. Take a look at it, would like your thoughts. You can find it at: Post #1222, Pg.62 by The Shadow.

I would sure like to know what was happening in the two minutes or so before AP/AT disconnect. If the pitots were beginning to clog, what was the autos doing to adjust for sensing slower speed that really wasn't slower.

Hi TD!

I've read the shadow, and more than once. As to a gradual packing up of the Pitots (and three apace), the a/c would trim ND/NU with ThrottleUp and/or THS/N/U/D. This would satisfy the computer, and completely bollox up the actuals. At some stage, and not necessarily due ICE, the a/p would quit (cannot hold 'Altitude'), or for any of a number of easily understood reasons. If so, and prior to "Disagree", the a/c may simply handover the goods and continue in NORMAL LAW. If agreement then degraded to the extent of AD needing "Timeouts", there is a separate kettle of fish, eh? That is UAS, certainly, but as a follow on, not as a priori. I am not sure if this mode is represented in the prior incidents of UAS, but I also do not know if our crew had experienced an actual UAS of the garden variety either.

I am amazed that so distinguished a group as this one hasn't thought the PF was indeed acting on bad data (not his "fault"), and was doomed to a guessing game either way, No?

nice to meet up again, my friend.


Can we call this mode: "Triply redundant bogus airspeed, but who knew?"

TRBA/bwk........

bearfoil;
Where is the guessing game here? It was an airspeed indication failure. Engine thrust including all engine indications, the altitude indication, the IVSI, the aircraft attitude indicator and the ship's GPS (for groundspeed and actual altitude) were all functioning and available, and the use of these indications is guided by the UAS memorized drill and checklist. There was no need to do anything but wait.

There is no guessing game which is probably the reason that it hasn't been brought up for serious discussion. It can be brought up I suppose but where is the evidence and what is the defence for the notion that this was a guessing game?

"Je ne comprends riens." From a pilot flying a wide body with 280 souls on Board. "No valid indications". "We have no indications".

Absent anything of value to rely on, my 'guess' is that "guessing game" is a kind sort of way to describe epic fail.

"There was no need to do anything but wait". The Pilot flying thought differently, and acted accordingly. You have evidence to prove he was mistaken?

If The Shadow is correct, and he makes a compelling case, the a/c was guessing too. Acting on bogus data might be a fair way to frame it. It passed along its blunder to the flightcrew.

Hi Bearfoil,
In fact, the FCOM is underlining that the aircraft "PITCH authorithy" is better than in "conventional aircraft". What is mentioned is about a very remote case: a fully "jammed" THS - without any manual or electrical trim available. Above 8 deg NU, elevators will still be effective below 180 knots: this is the limit speed mentioned for safe flying in this case.
But now, do you think that AF447 had a jammed THS? Should I remember you that autotrim and manual trim were both very likely available. In fact, it seems that the main flight-control issue about AF447 "recovery" attempt, so far, was rather the complete lack of PF sustained nose down imputs...

By itself, "All Trim Fuel aft" means nothing... there is a variation from 0 to +0.5% aft of the target CG due to central tank fuel use which is completed by transfer from the THS to the central tank during cruise; add also a 2% safe margin below certification aft limits (if target is at max) and a push-button for forward fuel transfer, just in case.
Recovering from a stall without sustained nose down imputs seems rather quite more challenging than any "all trim fuel aft".

bon soir, takata.

"Better" Pitch authority than conventional a/c? Describe Better? More powerful (Stubborn)? More "Travel" (~14 degrees)? Impossible to recover at certain speed and exaggerated deflection? Jammed or not, The THS is a formidable force. Pitch Authority is more complicated than mere power, though admittedly Airbus has that ground covered. How quickly is it available? If it transits to its NU limit, are elevators enough to recover from some unusual attitudes? I think the answer is patently NO.

Without assigning a responsible, the THS can prevent a predictable planform. yes?

I think there is a working theory that has fossilised into a narrow ledge for some.

I doubt seriously the pilots will be found to have handled this a/c correctly at every turn. Likewise, the a/c itself became a riddle....... "Je ne comprends riens"............

Airbus once again must confront: What is it doing? It seems the THS happily wormed its way to near Full NU, and remained there. Likewise, the a/c climbed careeringly to STALL.

Regardless the upshot of the piloting, A330 has a legacy of explanation of its own, imo.

Absent anything of value to say/add in your many posts, you'll better refrain from bolding quotes out of your FANTASY LAW universe.
"Je ne comprends rien" is not part of the tanscript, neither in French or English. (Beside, you dramatically added 52 "souls" on board).
Here is the good quote from the note in English.
And some part of your answer (about what indication was lacking) is certainly to be find in the following sentences:
So, fifteen seconds before this point, the speeds were invalid (AoA was above 40° and the PF was applying full back stick for 30 seconds)... then, maybe no speed at all were displayed on either PFD, nor Stand-by gauge...

I took a break. Several things are obvious upon reading to catch up. So I've caught up and herewith deliver this word bomb.

Nothing's changed, at least through the morning of the 6th.

(wordy annecdote)
There seems to be a lot of oversimplification in fields I work or have worked. For example, in some systems you cannot use "voting" type redundancy. The GPS satellite FSDU, Frequency Synthesizer and Distribution Unit, that takes the Rb and Cs clock 10.23MHz signals and applies carefully modifications before distributing the system clock to the two independant clock busses on the bird. There was a passive frequency distribution network from the clocks to the frequency synthesizer inputs. It combined all clock outputs into the three synthesizers' inputs. Since a 10.23 MHz sinewave is not something you can vote about only one clock and one synthesizer is powered at a time. (Otherwise you would get unacceptable spurious signals even with very high quality electronic relays to select clock signals.) I had to design for about 120dB isolation hoping to get at least 100dB. That's a 1 trillion to one isolation ratio as the target and 10 billion to one as a target. Trust me, clocks and synthesizers not needed were turned off. The outputs were all in parallel with a circuit designed to not short out the output bus if a transistor failed short circuit or open circuit. So that was, in essence, a passive voting system. The powered synthesizer drove the bus. With two busses a failure on one bus let the other survive and take over the full job. (/wordy)

PJ2 likes the first step of the UAS scenario concept. It fits very well when you consider it's claiming coincidence for all three pitots to fail the same way at the same time. If the PF side, which we do not see, showed even a brief overspeed just as the aircraft disconnected that would be the last thing the PF saw and he might have over trusted it. The off-again-on-again-finnigan stall alerts certainly could foster the view that they were spurious when overspeed actions fix what stall actions seem to bring back.

bearfoil wants and "easy" switch. It appears AirBus has patented such an idea. Way back in the June 2009 "era" of this discussion the idea of using the INS and GPS as a backup to the pitot tubes was broached. It was denied. Everyone figured long term, for which it cannot work. For short term, say a couple minutes, it should work just fine as an extrapolation tool. ABI seems to be planning to use this or is using it, likely on the A380. That's as "easy" as it gets. Warn the pilots to be on their toes and more attentive than usual. But fly through the mess, AoA, Thust, and GPS/INS. If at least one of the pitot tube readings does not become "sane" within 2 or 3 minutes or of altitude hold is too difficult then let the pilot ease into control. He's had a couple minutes to become well aware of what has been going on.

(By the way, the "easy switch" is the path towards a fully robotic plane, you know.)

BOAC seems to still desipise the glass cockpit and FBW. Right or wrong it is not going to go away no matter how much he rails about it. On a per million miles basis the glass cockpit in the A330 line shows half the accident rate and hull loss rate of the Boeing competitor he loves so much. Those are VERY loud words in mute bare statistics. I'd have more sympathy for him if the statistics were reversed. I suspect he will see a two person "executive" cockpit before he sees a return to cables or direct control. (The executive makes executive decisions about where the plane goes, never "files it", and is there rather than on the end of a long string of satellite relays away from Nevada. The other person is the relief executive. Both are trained with three special skills, giving the executive direction to the aircraft if something odd happens, telling the world about these changes, and reading the radars. He might even be trained to push a button to switch from one set of computers to another if something extraordinary were to happen.) It seems to me that is the direction it is all moving. And when the electonics can handle a hole blown in a plane with the plane still marginally flyable it will happen. "Get over it."

Confiture is remarking about the pilots seeming unawareness of what each was doing. If the quiet cockpit the BEA release conjures as a vision is real, at least one new feature must be trained into pilots: thinking out loud. Announce what you see and what you are doing about it. And acknowledge what the other has said. If that oddball scenario I mentioned above with PF seeing something quite different from PNF and recorded left seat only data then such an announcement of observation and intended action could have saved the plane.

PJ2 reposted (this time), for the Nth time, the schematic of where the various sensor signals go. I guess after RR_NDB's comments about redundancy it's time to wonder what makes the pitots redundant with only one going to any given ADIRU? I hope somebody did the numbers and decided this was better than feeding all three to all three ADIRUs. Or else there are data feeds between ADIRUs that are not shown. Anyway, I'll run an idea up a flagpole. Display all three to give speed with an error band with special displays for lost channels. Then there would be no chance PF would see something different from PNF. (I'd add a fourth channel to the picture, the consolidated GPS/INS report corrected for the last known calculated air speed.)

And we have the bozos who setup the cirriculum for AB330 pilots teaching them to keep their grubby paws off the trim wheels at all times and the designers who told the software people to never resume autotrim once the pilots mess with the trim wheels. Oy vey!

Detail regarding Captain seeing over shoulders of the PF and PNF: Note that he was found floating. The others were found belted in. He was probably standing observing closely.

As a hint for people who have a new thought it would be smart to revisit the BEA release and work it into the last 6 minutes and 21 seconds of the plane's flight. Ask, "How does this fit?" It might prevent your looking like a bonehead. (bearfoil continues his attempts at this proof.)

bearfoil, "Certification is mainly a mystery to me; I trust the system to certify a/c to be safe and reliable."
Sir, there is no such thing if you mean "never fails". The only way to achive "never fails" is to "never try." Stay on the ground and never fly. Then something other than an airplane will kill you.

Certainly not intentionally from me, bf. That is why I concentrate so hard on why they performed the damnfool stunt looking in training for the answer.

It's clear "what" killed the plane and the people aboard it. What is not at all clear is "why?"

Mr. O and bf mentioned "protections". They are NOT protections. They are limits. Calling then protections gives a false sense of security. I sense this all through this discussion, by the way. There is too much, "It can't fail." Can't is a very very big short contracted pair of words.

I remember asking WAY WAY WAY back in the June 4 2009 thread about some of this. If there is a standard drill for handling a plane in an unusual condition why does the plane not do it itself rather than expect the under informed (for instance no AoA) humans to take the load? There should be a better way.

Please forgive me here if I "bitch" about this a littlte. You have a TEAM in the cockpit. They are not telepathic. As you aviate tell your partner what and why. Think out loud, as I said above.

Do you read more than you post or do you post more than you read? (As it happens I rementioned this above.)

Quoth ChristianJ, "Personally just looking forward to the next BEA report... and the way that will put the fox among the chickens, once again."

Amen.

For those interested, The Shadow's post is on pg 61, #1218 R&N thread. We have discussed similar issues on the Tech Log threads, although not in such a concentrated manner.
It is pretty obvious that no one on the AF447 flight deck saw the trim indication in time. If they had, there should have been an understanding of what the problem was.

JD-EE

Thanks for the gentle treatment, ma'am. Guilty. I think you referred to BUSS (BackUpSpeedSystem), not fitted to 447. I do not advocate a robot of any description, it is (will always be), a potential lethal comeuppance for the Hubris of those who have more education than brains.

Machinbird. If that alone is the upshot of this tragic accident, everyone on several continents who had a hand in 447's demise deserves a flogging.

(ie: Not "noticing" Trim). For the record, may I opine that calling that THS a Trim device is abundantly fantastic? Like comparing a haircut to a decapitation.

The following graphic is a compilation of work done by HazelNuts39 and updated to include modifications suggested by A33Zab. It covers a period of 2 minutes from just prior to AP/ATHR disconnect to about one minute after the aircraft actually stalled.

http://oi51.tinypic.com/dbr05u.jpg

The time the stall actually occurred is shown as 2:10:59 and observation will show that the aircraft gained a small amount of height by ballistic momentum over the next 6 seconds.

I'm sure that HN39 will assist with explanations should they arise.

The graphic above is 785 x 500 pixels so as not to stretch the page when being viewed on displays of 1024 pixels horizontal resolution. Clicking on the graphic will link to a full size image (1100 x 700) displayed in a separate window.

Hi Machinbird,
I've read it and, quite frankly, it is full of obvious misunderstanding of the very basic systems. Hence, beside very plausible description of some likely crew reactions, the theory developped is completely confusing with such a mix about flight control laws including their related sub-systems monitoring.

I wonder how he may think that the initial PF pitch-up described by the BEA was certainly commanded by the "flight system" and was not due to any manual side stick imput?
If the flight system needs to move a control surface, if possible within the current flight law, it will do it directly without making some side-stick imputs! But if the recorded side-stick channel is showing an imput, this is evidently a PF action... Why would one have a single doubt about that?

Hence, what such post is mostly showing, is that many people simply do not understand the basic logic behind the flight controls but still are furiously patching up some theories that will contradict most of the investigators work... (see Bearfoil's sustained daily efforts to continuously stall this thread into the deep-deep!)

What would be the problem with the trim, beside the initial roll tendency after AP disconnection?

JD-EE, we were temporarily lead astray by forum members who apparently believed that autotrim would not resume control once you made a manual trim input. PJ-2 has kindly set us straight (with technical references) and explained that autotrim will resume once you take your paws off the trim wheel. I strongly suspect there is some misinformation adrift in the Airbus community that needs correction. I wonder how BEA will address that?

mayby the different positions of the static ports also generate a different sensibility in case of icing.....

dolphins have their sensors at the forehead under the blowhole!
http://onderwijs1.amc.nl/medfysica/d...s/image004.jpg

Takata,
The problem with the trim is that it moved to a high aircraft nose up setting without crew awareness.
It appears that a FBW aircraft requires the pilot to know exactly what mode the aircraft is operating in or else the question arises, "What's it doing now?" In the case of the AF447 crew, they had no attention to spare to consider what the trim might be doing, thus the nose up demands caused the trim to run silently to a high setting without the knowledge or anticipation of the crew.

Once at a high setting, the trim acted to stabilize the aircraft in the deep stall that they eventually achieved. When the Captain arrived on the flight deck, he had to puzzle out what he was seeing. From the jump seat, the indications were essentially hidden by the wheels themselves.

The THS autotrim system has virtually eliminated the problem of trim-runaway, but it has replaced that problem with a new set of problems.

With regard to The Shadow's post, I merely pointed to its location. It was a pretty good post considering the date of posting and the breadth of coverage, particularly the human factors. I did not agree with all his points however.

Hi,

JD_EE
Another Pprune scoop ...
Better than Le Figaro there ...
JD_EE .. you also been a "person close to the investigation" .. a "reliable source" ?
Can you show your reliable source confirming the text in bold ?