Something else I've just noticed.

The 4/4 release said that the wreckage was found "in the area of the abyssal plain". (http://www.bea.aero/fr/enquetes/vol....4avril2011.pdf)

The original french reads: "Site de l’accident dans une zone de plaine abyssal"), which I read as "in an area of the abyssal plain". Subtle but seems to me to be slightly different.

Regardless of which is more accurate, the abyssal plain appears to lie to the north of LKP (which looks to be right at the southern edge of the plain). This is fairly clear from Machaca's overlay of the image from the 4/4 press release and the search zones.

When you look at the Metron analysis (20/1), you see that after discounting all the areas searched, with a weighting for efficiency of each search conducted, Metron gave 2 likelihood plots for where the wreckage could still be: figures 32 & 33 on p 35.

The 8/4 letter (http://www.bea.aero/en/enquetes/flig...end.phase4.pdf) says:

The interesting aspect is this: figure 32 of the Metron report is the plot assuming that one or two pingers was working, while figure 33 is the plot assuming both failed (and hence Metron discounted searches done looking only for the pinger sound when producing it).

Figure 32 (assuming one or both pingers), shows that the hot spot (red) is almost entirely SOUTH of the LKP. It is a semicircle roughly 6 nm in radius, lopsidedly extending to about 8 nm towards the SW.

Figure 33 (assuming both pincers failed), expands the hot spot north of LKP, into the abyssal plain. It is roughly a circle about 6-7nm in radius. It includes the area in figure 32.

If we assume the 8/4 statement indicates that the wreckage was found in the red area, then:

• either the wreckage was south of LKP, in the red zone of fig 32 but not 33, in which case it is not on the abyssal plain, and BEA's reference is likely to be obfuscation to conceal its location.
  OR
• it was found north of LKP, in the red zone of fig 33 only. In that situation, the wreckage could be on the abyssal plain. But it also suggests the pingers might have been disabled/destroyed on impact, unless there is some other reason they didn't activate.

See this overlay of Machaca's image and fig 33 from the report to try to make it clearer:
http://i.imgur.com/95PTO.jpg

Spiral dive ...
 

Spiral dives generally end in a broken airplane due to excessive G loads during recovery (HS or VS failure in flight) unless you get to FL 0 first that is. The graveyard spiral seems ruled out in this case by the available evidence. Don't we all practice recovery from unusual attitudes on instruments? With loss of one or more primary instruments?

The BEA report indicates the pilots never lost inertial attitude reference information. Power and attitude is how you fly the plane and I see no indication the pilots ever lost power or attitude information on their flight displays even though they did lose airspeed and vertical speed info. I strongly suspect that the pilot and flight computer were working at odds with each other (as someone suggested here due to change in stick feel in direct law) or else the pilots allowed the A/P to fly the plane with two incorrect but agreeing airspeed inputs (note the new AD) trusting the A/P in IMC at night with the computer putting the aircraft outside the flight envelope due to invalid airspeed and altitude data.

Being a software engineer, I am curious about the approach Airbus used for boundary testing of the flight control computer software. It is always the combination of inputs that nobody thought was possible that gets you in the end. The X31 article that was posted here is highly relevant in that regard. I believe that a software design error will be one of multiple contributing factors to this accident. The new AD that was issued suggests that Airbus and the FAA already believe this to be true. You can design better air data sensors but never one that is 100% reliable in all possible situations. Therefore, you better design software that can handle the loss of all air data input without loss of control of the aircraft. And computer systems are not 100% reliable either. You see the basis for my distrust.

Henra, just imagine you could pump the stick while in direct law in time with the aircraft's short period oscillation. Don't you think you could amplify the aircraft's pitching motion until the aircraft stalled in just a few cycles? Particularly with a C.G. almost at the aft limit, it would be easy. That is a dynamic departure from controlled flight, and it can be done in just a few cycles.
Now just imagine that the aircraft is flying in Normal law, on autopilot and with autothrottles and something causes the motion of the control surfaces to slow down so that they never get to the commanded position before the signal reverses. The control surface itself does not begin to reverse direction until its position signal crosses the command signal. This happens significantly later than the command signal's order. The result is like the F-22 prototype that encountered PIO on a go -around.
The something that delays/slows down the control surface is called control rate limiting, and it is a known cause of PIO. Do you see that what you may have learned about the aircraft's control system may not apply in special circumstances like control rate limiting?

This looks a little odd, the Ile de Sein looked like it made a quick stop in Las Palmas de Gran Canaria, then was headed southwest at 11-12 kts, slowed to a knot or less, and did a maneuver roughly shaped like a backward lowercase cursive L, and now is heading west or southwest still under a knot.

I figured they had about a 5-6 day trip to the crash site when they left Palmas de Gran Canaria, where maybe they stopped quickly for provisions, personnel, or equipment, not sure what is up?

Live Ships Map - AIS - Vessel Traffic and Positionshttp://www.marinetraffic.com/ais/def...05:17:39%20AM#

ACLS65;

A little bit of research suggests that the "Ile de Sein" is engaged in carrying out a repair or repeater replacement to the Atlantis-2 submarine cable that links Argentina, Brazil, Canary Is and Portugal. There will be prior commitments to conclude before going on charter to the BEA, and in any case time will be required to position the ROV and all specialist personnel required for the AF447 recovery operation.

ACARS
 

Would you get ACARS for major malfunctions of ailerons, flaps or rudder?

We now have good reason to believe the pingers failed, and the Metron analysis consequently places the wreckage in the area of the Abyssal Plain.

The statistical analysis undertaken by Metron, methodically worked through the permutations, and now after the fact appears to to be valid. Irrespective, the BEA have chosen to fudge the outcome, as they do not want the site to become a tourist mecca; which is understandable. The debris field is in the Abyssal Plain somewhere in the north quadrant relative to the LKP. I don't believe the specific position is relevant at this time, as none of us has a need to plug it into a GPS.:hmm:

sd666, there is another reason I've not touched upon because it is so obvious for the ACARS transmissions to be blocked during some intervals.

Other aircraft are in the air. And they have position report messages to file. Several such could have taken place in the time window specified. I do not know the full ACARS protocol for transmission slots, transmission handshake, and who controls when aircraft send data packets are all unknown. However, two aircraft trying to send a message at the same time would interfere with each other.

Since ACARS is not an emergency sort of channel it is possible the protocol requires some form of wait between transmisson periods. During that wait other planes could request a slot. Then the satellite would designate the next plane to send a data packet. (The strategy after a packet's receipt acknowledgment would call for a small random delay with a very short time slot allocation request message being sent at the end of the delay. Planes heard would be serviced one at a time by the satellite granting them a transmission time slot.)

Regarding a spin maybe driving the satellite out of the antenna beam the satellite is geosynchronous. The antenna has a 40-45 degree cone to its 3dB or half power point judging from its claimed gain. The satellite use was probably the one just about over the coast of Africa. The plane was virtually under it in the N/S plane and well inside the satellites footprint E/W. If the plane spun fast antenna tracking would not work and it would probably go (barely) outside the beam. If the spin was a "leisurely" 2 minutes, as the gaps suggest, the beam COULD be automatically steered to maintain its aim at the satellite. (I was involved with a completely different use for Inmarsat satellites. And our antenna easily tracked something as leisurely as 2 minutes. I heard that some Arabs mounted antennas on top of their cars in Saudi Arabia and drove around using Inmarsat as their "cell phone" before there were cellphones. Cars turn. But the antenna tracked it.)

That's the basis for my doubting the transmission gaps are caused by the antenna aim being off the satellites.

On the other hand, the antenna is a patch antenna. So it's steering ability is somewhat limited to a cone about 75 degrees from vertical. A heavy bank could position the antenna so that communications was very marginal. A really severe bank could literally hide the satellite from the antenna.

One last observation I believe I noticed in the BEA report (English version) was that there were no indications of attempted ACARS transmissions that did not make it through. What I don't know is whether my mind is playing tricks on me or not. Bad aim would lead to partial messages with high error rates. That would cause them to be tossed. But the system would likely record the event as a lost message.

(Time for me to shut up. I'm getting into stuff to obscure to try to detail for probably no great purpose.)

Here's a brief addendum inspired by a trip to Wikipedia (Inmarsat antenna locations) and PJ2's drawings.

Apparently the satellite likely in use is at about 15.5 degrees West. That means the antenna aim was on the order of 15 degrees off vertical. That means with its beamwidth if the antenna was aimed 15 degrees off vertical the opposite direction the satellite would still be well within the lobe perhaps at its 1dB or 1.5dB points ( 80% to 62% power points). Any transmissions would have been noticed.

(Visits to the Inmarsat site dug up pictures that were rather vague and no actual parking slot location. So I had thought the antenna was closer to 0 degrees via a really conservative guess.)

JD-EE;

I seem to remember, but can't find the source right now, that:

- Somewhere in report #1 or #2, BEA says that all messages passed through the same satellite, and names that satellite. The Inmarsat site locates that satellite to the West of the aircraft.

- Somewhere else BEA says that during one of the 'breaks', the connection to the satellite was lost. Some considerable time ago, a poster on this thread disputed that, saying that BEA had misunderstood information provided by the satellite operator, and that in fact it was not established that communication had been lost in that period.

IIRC, ...

Sure, ACARS are triggered for any malfunction (minor, major) and cockpit alerts, including ACARS being inop... but nobody would receive it.

Thanks TAkata,

If this is the case, is it safe to assume there were no malfunctions of rudder, flaps and ailerons at least until the last ACARS received (cabin vertical pressure i.e. +/- 6000ft)? If true, then no VS separation at high altitude?

Drift below surface
 

Salute,
Well, a close look at the currents below surface will tell that it would be almost negligible for most of the heavy parts sinking right after impact (see, for example, this SHOM study about this spot hydrography available on the BEA site).

Consequently, the wreckage field should be pretty close from the actual crash site. Most wreckage would sink immediately with very little drift and only some of the parts with any remaining buoyancy would drift further away on the surface before sinking when its buoyancy will become negative.

Salute,
Well, it is safe to assume that what fault was not detected until 0214:26 was very unlikely.

But... this last advisory (cockpit ECAM signal) is about Cabin V/S (vertical speed) being +/- 1,800 ft/mn during 5 seconds.
It call for a reset/switch of the CPC (Cabin Pressure Controler).

From A330 FCOM:
To force a CPC changeover:
- Mode SEL.............MAN
. AFTER 3 seconds:
- Mode SEL............AUTO
The inactive CPC may then be reset
(check the CAB PRESS ECAM page).

What it means, then, is that the cabin pressure control was still in AUTO mode up to this point. As the CPC is also feeded by those ADIRUS (which faulted), I really don't know if any altitude may be safely deduced from this advisory alone without any following informations.

Nowhere did I read in the BEA report that this "fact" was established: that the aircraft altitude at this point was 7,350 ft (its cabin alt at cruise) or below. If it was a "fact", in my opinion, it would have been already mentioned as an evidence.

Auraflyer

It will be interesting to see where the wreckage really is when the location is finally disclosed. I have long believed it is to the S of LKP based on drift analysis - although I am the first to admit that there are a lot of assumptions inherent in this. In the end however, the location really doesn't matter. Speculation as to the location of the wreckage was no more than a means to an end - examination of the wreckage and hopefully recovery of the boxes. The location alone can't tell us what is important - the events that led to this disaster.

One thing I am a bit puzzled about. The vessels involved in stage 3 had AIS (I remember mm43 giving us frequent updates). If phase 3 had been successful, we would learned of the location - unless of course they turned AIS off at that point. This time there is no AIS and they have not disclosed the location. Is it just coincidence, or did they make a conscious decision to deny the public this information in phase 4. If a conscious decision, why the change in policy?

I guess I can understand the reasons not to disclose the location. But then why is it they apparently didn't care in the earlier searches?

Although not searching for a conspiracy (and I really don't believe there is one), there are a few too many unusual things about the entire case.

Yes, Olivier,

we discussed this at length some time ago. The options are:

- if the negative relief valve opened, it is irrelevant what the cabin pressure control was doing.

- if it did not open, then please come up with a mode of the CPC that would command the cabin to climb or descend at a rate greater than 1800 fpm. I believe that to be so improbable that it does not merit discussion on this thread.

PS:: In the earlier discussion you mentioned that the ATA code of the message points to the CPC. I believe the reason for that to be that the cabin pressure sensor is located in the CPC.

That's probably true for the initial pitch down when you intentionally let the speed to increase to VMO with idle thrust but soon the ROD will decrease and you certainly won't lose 36000 in 6 minutes that way.

The 330 is more known as a great glider that just wants to fly.

mm43
in any case time will be required to position the ROV and all specialist personnel required for the AF447 recovery operation.

Yes, and in less than two months, the hurricane season starts and it will also be two years since the demise of AF447. I hope that they can recover everything they want before having to contend with the ITCZ seasonal weather. The French certainly like to takes things in a laid back manner.

JD_EE; I agree with you in that I think there is evidence in the BEA report that the antenna was not blocked immediately prior to impact.

Therefore there is no evidence that there was a loss of hydraulics or VS prior to impact.

Takata; I don't disagree with you. My point is that there is no strong reason to suppose that an impact point close to the LKP is unlikely given that the 3D path of the aircraft between these points is unknown until the FDR is recovered and successfully read.

Some colleagues in a 330 simulator, I say again, in the simulator, tried to exit from a full stall situation ... Very harsh, some managed, but most didn't. Slats out helped and one even used max thrust on one engine only ... Is the simulator representative of real life ?

For fuel economy, the 330 adopt an aft CG in cruise, and this does not help in case of stall.

Don't forget that a 330 is supposed to be protected from a stall, the most you usually do is to train up to stall entry and recover, or if in normal law, you watch how the protections work to avoid a stall and maintain alpha max.

Cabin V/S message
 

Quote from HN39:
"...if it [the negative relief valve] did not open, then please come up with a mode of the CPC that would command the cabin to climb or descend at a rate greater than 1800 fpm. I believe that to be so improbable that it does not merit discussion on this thread."

takata and I have argued about this before, but I think there is one very remote possibility.

If at 0214z the aeroplane was still at medium-to-high altitude, a failure of both packs (or a double engine-failure) would cause a positive cabin VS (loss of cabin pressure) until the outflow valves had a chance to close. This might exceed +1800ft/min, falling (hopefully) to +500ft/min or less with the outflow valves closed (depending on the state of the door seals). Outflow valves closure might be automatic by a serviceable CPC, or by crew action if not (see the drill on takata's post).

However, I would have expected double pack failure and double engine-failure both to be ahead of cabin VS in the ACARS message hierarchy - they would be in ECAM hierarchy.

Maybe I should elaborate on my earlier post on ACARS.

"MAINTENANCE STATUS ADR2" was transmitted at 2:14:14.

This should have been followed up by a Class 2 fault message between 2:15:00 and 2:15:14 (referring to BEA interim report). That Class 2 Fault message was not received. Since the Aircraft's orientation should have permitted sattellite communication prior to impact - that makes it likely that impact was within a minute post 2:14:14 - making it unlikely that the aircraft was at "medium/high" altitude at 2:14:14.

Salute,
If, if, if, hourah !
:-)
My point is that everything unproved by BEA analysis is worth mentioning. And so far, no analysis of this Cabin V/S adivisory will point at: there is only one explanation to this ACARS: the aircraft altitude had catched the cabin altitude at 0214:26 (less transmission protocole, less 5 seconds).

a) When the CPC is working in auto mode, some FMGC and ADIRUs imputs are needed (in this case, there is none and an upset would switch it to descent mode without any recovery altitude entered). Then, it is left in semi-auto mode (no FMGC/ADIRUs imput). Then, up to what speed may work the CPC in case of brutal and massive loss of altitude? (those working limits are for the full-auto mode).

b) In case of CPC problem (I'm mostly thinking about power supply due to on-going engine flameout issue); the CPC should switch to manual mode and use its built-in backup sensor (the one that you are refering at). In this case, the difference in sensitivity alone could produce an immediate altitude difference in cabin altitude of +/- 1,000 ft. While 5 seconds is enough to trigger such an advisory (+/- 150 ft difference = 30 ft/sec = 1,800 ft/mn).

So, I'm not as sure as you about its meaning, considering also that it is the last in the sequence and much more could have followed if more ACARS could be sent after this point in case, of course, that some power supply was still available for the SATCOM (or that the aircraft was not crashed).

high sink rate
 

On what ground are experts still thinking in terms of a normal descent from 36.000 feet down to the ocean?

There must have been a upset scenario pretty much from the beginning to loose that amount of height in the probable time frame. In itself its nothing more then loosing the potential energy. It´s traded for speed and or distance in the normal flight regime, and its used up by a lot of drag and loss of altitude in a deep stall scenario nd airspeed below flying speed.

To regain the speed, the drag has to be reduced by flight control input, normaly the stabilator to reduce the AOA. For that you need a favorable CG and full stabilator effectiveness (no out of trim situation like full nose up trim).
That simplified basic probably applies to most aircraft designs. Can´t think of AB being totally different when it comes back to basic aerodynamics.

I think it is pretty sure that an upset occurred leading to low speed, high sink rate and that continued to prevail until the final end. Aircraft had been lost before for that reason and a descent rate of 6.000 feet / minute does nor look unbelievable in such a situation.

Do we know anything about the position of stabilator trim and the position of the CG? Could´nt find it in the amount of posts.

ACARS
 

Quote:

---

Originally Posted by sd666

Maybe I should elaborate on my earlier post on ACARS.

---

Salute,

Two other ACARS were "in the pipe" after PRIM1 & SEC1 faults which should have triggered something also before 0215.

But look at the BEA explanation for the end of ACARS:
a) the aircraft crashed. (possible)
b) the ACARS system was inoperative (possible)
c) the was no other ACARS to be transmitted. (false)

b) A power supply failure after 0214:28 could have stopped the ACARS transmission. Consequently, the crash time canot be acertained from the end of ACARS alone.

Hi Olivier,

Thanks for your informed and elaborate reply. I've just one question. As far as I could find in the available documentation, the CPC has only one source for cabin pressure, i.e. its built-in sensor. Am I missing something?

The system diagrams are also quite obscure to me but look at the FCOM note "Control indicators":
Note:
The pilot may notice a variation (up +/- 1000 feet) in CAB ALT indication on the ECAM PRESS page, when the system switches from the cabin pressure control AUTO mode to MAN mode, due to the reduced resolution of the backup pressure sensor.
Obviously, the pressure sensors in AUTO and MAN modes are definitively different ones.

HN39;

Just digging around while we anticipate further once the Ile de Sein arrives and begins work.

The following is some more information about the cabin pressure and the advisory message. Just trying to get information out there to think with:

The negative pressure relief valve (located between STA 24 & STA 25 on the port side of the fuselage), is rectangular, has 4 spring rods which force it closed until a pressure differential opens it. From what I can determine, the test equipment "vacuum" pressure for a test of the relief system is about 0.3psi. There is also a physical test...push the valve inwards for freedom of motion and re-seating after returning to the closed position. I don't know the valve's dimensions.

We know that the ADVISORY CABIN VERTICAL SPEED message was sent at 0214+.

Regarding system parameters monitoring and the ADVISORY function....Some system parameters are monitored throughout the whole flight and automatically displayed on the relevant system page when their value drifts out of normal range but well before the warning level is reached.

This is performed by the ADVISORY function of the EIS. In this case, the related key system light comes on on the ECAM control panel. The relevant page is displayed as long as the advisory is present and no other system page is called (manually or automatically).

On the right part of the lower ECAM (STS section) are displayed (in priority order):
- inoperative systems.
- maintenance items (class 2 maintenance status).

NOTE : The class 2 maintenance statuses are not displayed in flight but they are sent in real time to the CMC and they can be transmitted to ground via the ACARS when installed.

I don't think the Advisory message is a Class 2 maintenance message - as described, an advisory is an indication of a parameter exceedence and is not advice of a failure which requires maintenance action at the destination station. But the fact that Class 2 messages are sent in real time to the Central Monitoring Computer is interesting.

The last message received was the Cabin Advisory message. We know that such an advisory indicates a system parameter exceedence, in this case, >1800fpm or <1800fpm cabin vertical rate. (The size of the negative pressure relief valve opening could be used to determine flow rates and therefore rates of descent of the cabin but the exercise may not be worth the trouble).

As discussed and posited, we have concluded that cabin pressurization would become negative below the cabin altitude, whatever that was, and we have assumed around 7500 to 8000 ft barometric altitude, perhaps slightly lower depending upon data available to the CPC, (ie, flight phase will not have changed from CRUISE to DESCENT because the FMGEC and therefore the CPC will not have been in the descent mode and may not have begun depressurizing the cabin as per the usual schedule).

Then we have this information:

Low Delta P/High Descent
This shows that the aircraft will have a negative differential pressure if the descent rate is continued and that a negative pressure relief will occur. The signal is sent from the automatic part of the cabin pressure controller 1.5 minutes before the external pressure is the same as the pressure in the fuselage.

and,

Low Differential Pressure
If the cabin pressure controllers detect that a negative differential pressure situation may occur in 1.5 minutes: - a single chime is heard, - the amber MASTER CAUT lights come on, - on the EWD of the EIS, CAB PR LO DIFF PR and the necessary steps are shown.

Pondering the question...depending upon altitude, descent rate and the incoming data to the CPC, we might have expected this caution to trigger if the conditions which triggered the CAB VERT SPD Advisory were as we are assuming, (essentially, high aircraft descent rate "caught the cabin", as we say) a high descent rate passing through 8000', approximately. But the CAB PR LO DIFF PR, which, (now theorizing), according to the above description would have occurred either before (a minute and a half to be exact) or simultaneously with the Advisory, is not in the ACARS listing. We don't know if the message depends on the flight phase, (some warnings and cautions are inhibited in various flight modes).

The risk is always in over-thinking something but "obvious" is only available afterwards! These are some random thoughts which may or may not be useful in others' thinking.

A chilling conversation
 

Our discussion on the choice of initial, and subsequent, search area has been made with the assumption that the pilots on the flight deck were behaving as rational actors in most respects. From this, we attempt to solve the mystery regarding how far and for what amount of time they remained "on original track." This leads to the perplexing question of

"How can they not have been further than position "X" based on time, distance, velocity, and eventual arrival at the ocean after descent at "Y" rate?"

There's a chilling possiblity that had not occurred to me until I spoke with some old colleagues (we had a twenty-five year reunion from our old squadron) all of whom are now in either the Passenger Carrying or Freight Carrying profession with such companies as AA / Delta / Southwest or Fedex / UPS.

That possibility: a delayed course change (to evade towering cells) that was in progress at the time of the upset, the course change delayed due to physical fatigue. I asked a few of the guys if they'd been watching this investigation (after we'd had a chat on the Cargo Crash in UAE), and Capt "H" very quickly opined the following:

"I think you'll find out the crew were asleep when they find the CVR."

I wasn't prepared to consider that.

I asked how the heck he knows that?

I pointed out that I understood the Captain to be in rest, per a typical long haul practice in many airlines, with two FOs alive and well on the flight deck.

Capt "H" nodded, and said quite simply that he finds the Occam's Razor explanation to be that the two on the flight deck nodded off. He'd had it happen to himself more than once when he was an FO. Typical happening was the Captain on a long haul would advise him "I'll be getting 40 winks, make sure I am awake in an hour" and before an hour is up, "H" with all great intention to stay awake, nodded off himself, to be awakened by said Captain a bit later ... but nothing bad happened in any of those cases ...

He then helped me understand what could happen.

NOTE: this is a bit of reasoning not intended to injure the rep of the gents on the flight deck that evening, but a way to address

"How did they not go further down intended track based on what is available to know from Wx and ACARS to date?"
(Idea being that they should have been farther down track from LKP).

With a scheduling variation that doesn't allow both of the co-pilots to quite adjust their circadian rhythms, they are by rule "crew rested" but in mid-cycle (circadian rhythms) on the day of this event.

So, per usual, Captain does take off, and a bit later goes off to rest bunk to be sure he's fresh for approach and landing later on in Paris. Two perfectly good and reasonably rested pilots/FO's are on the flight deck, but due to their Circadian rhythm being a bit out of synch, are on a not-quite-settled rest cycle. (I ran into this problelm in the fleet years ago when we began a few weeks of midnight launches, after a few months of mostly day and evening ops. It takes a few days to reset the clock ... we had a number of close calls with eyelids on pilots, and in one case a crewman who made the "wake up" call in time ... )

One nods off, then the other ... and by bad luck, none of the FA's or other Cabin Crew had occasion to communicate with the flight deck for any of the usual reasons. When the big cells appear on Wx radar, neither cued the other to avoid storm (stay with me here) since neither saw the display.

At little bit later (perhaps an FA knocks on the door with coffee?) one of the two awakens much closer to the storm system, and with a peak at the radar realizes

"Whoa, must deviate/change course, too close to storm system."

Being human, and not wanting the Captain to be angry for getting behind the situation, he has awakened his mate and is in process of a course change (both being in the "must sort this out and not get the Captain mad at us" mind frame) when they encounter the severe weather's effects on the airplane. (Note, this sequence of events also points to flying as a priority, aviate navigate communicate, and thus they don't get the message out to the pax "strap in, rough weather ahead" nor to the FA's. Playing catch up, and all that goes with it. Evidence points to quite a few folks not being not strapped in ... )

With (5 or 10 or 15? degree) AoB of turn, you lower your stall margin, right? (At those altitudes, and at that mach number, per previous posts in this thread, it's pretty obvious that one handles any turn with care.)

That aerodynamic reality suggests that if this course change were in progress, the stall occurred in turning flight rather than in level flight ... with the idea that the stall was triggered by hitting unstable air with less stall margin than one normally carries.

This might explain the upset in the first place. (Reduced stall margin).

To compound that, pitot tubes all iced up won't help them after upset occurs, and they have bad to no airspeed information (at least initially) in their scan as they attempt to recover from the upset.

With this unexpected event confronting them, crew for one reason or another (many such reasons were discussed in massive depth in this thread for the past year and a half) were unable to recover.

While attempting to, they were not heading in the general direction of the original flight planned route, but in a turn (one way or the other) away from flight planned route. This changes/reduced the time and distance clicking away. This proposed scenario may resolve the puzzle over why they were this close to LKP.

The other chilling idea based on Capt "H" and his Occam's Razor -- nod off and just fly into the cells -- doesn't answer the problem of "so close to LKP" and isn't something to discuss in the context of

"How do you decide where to look and why, based on a datum of LKP and an estimate of upset based on ACARS?"

BEA had to make some assumptions on where to place datum and where to begin their search. If the assumption included "generally heading along original course from this flight plan" it could explain why they didn't look closer to LKP.

If I remember rightly, early on in the discussions here on PPRUNE forums, some folks had raised that idea of a nod off. It is not my intention to resurrect that zombie, but to apply that idea to the possible effects of track and thus location of the wreckage based on LKP.

Apologies if I was not clear about that earlier.

I am not sure if Capt "H" is right. His remark got me thinking that AF447 migh have been making a course correction/deviation around the cells when an upset occurred, with consequences that I had not yet seen in our lengthy discussion here.

EDIT: I had to revise this a bit to make it clearer.

Update on 2010/04/11 meeting between BEA and family representatives
 

Tour Hebdo - site officiel - Transport : AF 447 : les éléments de l'avion rapatriés mi-juin

Google translation:

Ah the great french judicial system, the wreckage will probably spend another 2 years in a shed near paris waiting for a "justice" to decide whether or not to allow the BEA access to it.

And before this will start, the President himself, Mr. Sarkozy, will do a "test flight" on an A330 Simulator.

More on 2010/04/11 BEA meeting
 

Nothing new on the recovery, but more insight into the relations between France and the families:

Frana nega representante de famlias no resgate do voo 447

Google translation (too long to paste):

Google Translate

Hi,

The president Sarkosy don't need a A330 simulator ... as the presidential plane "SARKO 1" :) is a A330 !

CPC
 

PJ2;

Thank you for collecting and sharing so much fascinating information!

One question comes to mind. In order to generate the 1.5 minute advance warning, the CPC's need aircraft altitude (or static pressure). Would that information coming from the ADIRU's in ADR DISAGREE mode be considered valid by the CPC's?

Takata:

"b) A power supply failure after 0214:28 could have stopped the ACARS transmission."

What evidence is there of a power supply failure?

The only evidence is the absence of a Class 2 status message. This would require the loss of both power sources within a 60 second window. It's possible (so BEA rightfully list it) but unlikely - there's redundancy in the power supply.

Hi SD666,

Quote:
What evidence is there of a power supply failure?

(I shall assume that is not a rhetorical question.) Definitely at the end of flight; otherwise it reqiures the roughly-simultaneous loss of all AC generation. Assuming the APU was not running, this would involve the loss of both engine-driven AC generators (one per engine). In each case, that could be caused by a failure of either the generator itself or the engine. As you could tell me: to prevent an ACARS message being sent for the first failure, the second failure has to be very soon after the first. By definition, therefore, there can be no ACARS evidence of such a failure sequence.

Thanks for your informative and well-argued posts.

Bleeding speed fast
 

Gentlemen,

there is one possibility for a very fast deceleration from cruising speed to stall speed (low speed, high AOA stall), and that possibility was only brushed by a few people here : pulling back hard on the elevator controls.

I leave to your imagination why such an action would be taken. I have my own idea. And remember, the humans in the cockpit are not the only ones with access to flight controls. The aircraft itself usually has the final say over these controls.

Regards

Chris:

Good point, though I'd still argue you've got auto re-light and RAT as protection layers. All-in, I think dual-flameout is not a high probability.

sd666,

All agreed, with the proviso that RAT in itself would not, I think, restore enough AC power to enable ACARS messages.