Who said anything about full throttles? I'd pull power back toward idle until I began to lose pressurization.
Some of the the F-4 aircrews were issued Moon suits and as part of their syllabus were supposed to zoom up to 70,000 feet. The throttles had to be pulled back to keep from exceeding 100 % RPM, and if you got high enough, you had to shut them down to keep from exceeding RPM limits at altitude. Fortunately for me-they didn't have a moon suit in my rather common size.:}

Below quote in reference to engine operating envelope charts availability.
It seems that the crew somehow was able to keep the engines running for the majority of the descent, at least some of which must have been at very high AOA. Perhaps the fan creates local airflow straightening in front of the engine at high AOA. Perhaps the power settings were low enough to avoid stall.

I don't think that Direct Law would be lost. Just that the crew would not deliberately try to achieve Direct Law as a result of their training. Even with Direct Law, it would not be easy without an Airbus demonstrated procedure. You would probably have to dial in full nose down trim, and then quickly undo it on recovery. Something like what the F-16 has to do to recover.

The F-16 has to rock itself out of a deep stall if it gets stuck there.
Read "Semper Viper" in this link http://www.codeonemagazine.com/image...49318_2157.pdf
to see all the fun you have been missing.:E
Link courtesy of Gums.

infrequentflyer789;
I think you might be right, and possibly, "there is a need...".

32° of rudder would be nice to have - if it was going to get you out of an otherwise "no win" situation that apparently you shouldn't have gotten into.

These theories are getting more & more bizarre the nearer we get to hearing from the CVR and FDR ;)

Back to Cabin VS...
 

grity, quote:
"if the bird is without any control but starts with an extrem climb mayby 60 deg, this (~balistic) path will reach a maximal high of ~12700m (full change the kinetic horizontal energie into potencial energie), while reducing the speed to very low untill it will deap stall and if it then fall with a stable AOA of 50-60 deg. back to 10500m and afterwards down to zero....."

Although this violent pitch-up would be an effective speed-loss mechanism, and looks to be the best (maybe the only) way of greatly exceeding the normal-stall AoA, there is evidence that suggests it couldn't have involved a climb as high as your ~12700m (~FL416).

The cabin altitude and differential pressure in the cruise at FL350 could not have coped with a sudden climb of 6000ft without maximum differential pressure being reached. Shortly after that, the safety valve would operate and the cabin VS (cabin climbing) would increase to something well over the +1800ft/min required to trigger a warning identical to that generated at 02:14z, the subject of the last ACARS message received.

Regret to say that I cannot calculate the flight-level where this would occur, nor the relationship between aircraft VS and cabin VS at the safety-valve differential pressure; but a ball-park figure would be between 2:1 and 2.5:1. So I think an aircraft climb VS of about 4000ft/min would suffice. The passing flight-level where this would start might be about FL370 - FL380. Sorry I cannot be more helpful.

Chris

PS
For new readers, there has been a broad consensus on previous threads that the cabin VS warning at 0214z was the result of the aircraft "catching the cabin" in its last descent, causing the inward-relief valves to open and allowing the cabin to descend at a VS exceeding -1800ft/min. This would have happened at about FL060.

JD-EE
I have never lost an engine during tail slides in multiple types of aircraft. Only once had a compressor stall going to full AB during a high AOA scissors at FL290, and that quickly cleared. So tail slide does not necessarily mean flameout or stall.

takata, quote[FONT=Verdana][SIZE=2]:
AF447 - timed cases for one upset from cruise level:
01.06.2009..35,000..FL350...60s..~5 NM..A-330 (1 mn)
01.06.2009..23,000..FL350...90s..~5 NM..A-330 (1.5 mn)
01.06.2009..17,500..FL350..120s..~5 NM..A-330 (2 mn)
01.06.2009..12,000..FL350..180s..~5 NM..A-330 (3 mn)
01.06.2009...9,000..FL350..240s..~5 NM..A-330 (4 mn)
01.06.2009...7,000..FL350..300s..~5 NM..A-330 (5 mn)
01.06.2009...6,000..FL350..360s..~5 NM..A-330 (6 mn)

Monsieur, can you explain what these figures mean, and from what data they were derived?

The problem of such form of reasoning is obvious to me:

Axiom 1: ACARS sequence (pitots icing) started at 0210; consequently, it would certainly cause an unrecoverable upset at nearly the same time (or even before if you want to make the Normal Law pulling up this zoom climb all by itself).

Axiom 2: This unrecoverable single upset was followed by a massive loss of altitude, but at a very slow rate of descent, from 0210 to crash time; during those four to five minutes, while a very small linear ground distance was nonetheless covered.

Axiom 3: As ACARS sequence ended at 0214:26, this aircraft crashed no later than 0215:14.

While I'm questionning Axioms 1, 2 & 3 altogether, your point will become fully circular if your answer is that those engines could not have stalled at very high AOA because they obviously did not have stalled during all the descent down to sea level, which some part of it should have been at very high AOA.

Don't you think?

As for the "full power", a response (at this time) to low speed stall was supposed to apply (full?) power and to reduce AOA -now revised the other way, if I remember it correctly.

Even if he crew did nothing and the engines stalled they would still keep running.

A stall does not equate to a failed or dead engine. It is only a symptom as a sneeze is to a human.

Takata (Olivier) What if it wasn't pitot icing that caused the ACARS message at 0210 but an actual stall that caused erratic airspeed data in the same way that the Perpignan aircraft experienced bad airspeed data at the stall? No one his been able to discount this possibility yet. Can you?
Thank you for the spirited discourse.

Hi Chris,

Quote:

---

Originally Posted by Chris Scott

Monsieur, can you explain what these figures mean, and from what data they were derived?

---

Sorry if this was not clear.
I formated the columns exactly like the table above including 9 historical cases of LOC in order to compare them easily with 1 to 6 minutes rate of descent from FL350 for AF447 in case of a single upset leading to the crash.
So, the headings are the same like in the first table:
1. Event date
2. Rate of descent (ft/mn)
3. Flight Level (departing)
4. Duration (seconds)
5. Distance covered (Nautic Miles)
6. Type of aircraft
7. Known cause of upset

It is showing that a 1-3 minutes duration event might have fit with the historical table (12,000 to 35,000 ft/mn); that the covered ground distance of 5 NM is also coherent with other historical cases; but that a possible 4-5-6 minutes event (6,000 to 9,000 ft/mn) do not match the precedent cases.

Hi Lomapaseo,
Right!
Nonetheless, such a sneeze would have been reported via ACARS -even loss of thrust, and there is no trace of it in the sequence... Then, I asking what the chance really are that they will keep managing perfectly those throttles all their way down to the sea and avoid it?

A/THR off, FADEC compromised (ADM feeded), heavy tropical rain, airfoil critically turbulent... ice ?

S~
Olivier

takata

howdy. The limits accepted and proposed in your post differ from A330 A/P limits only in ND. The 330 a/p limit for ND is nine degrees. This I recall from data posted by PJ2. I think that given the data supplied by ACARS and the requisite times, it is premature to eliminate other than ICE in Pitot Tubes as the cause of 'upset', a/p drop, or LOC. Something is not right about PF allowing a/p to exceed its limits and decouple involuntarily. He knew what was in store (ITCZ), was alert, and something happened to paint both pilots into a corner. ACARS suggests an attitude locking with the satellite for those four minutes.

At the impact point I think 447 was travelling as slow as she ever did; iow, she was decelerating perhaps continuously from upset and/or loss of control??

Still she could have been travelling at 160knots down, and 80knots forward for an actual airspeed of ~220Knots?? At this speed, with perhaps an AoA of 80 degrees plus, does this fit??

JD-EE I think the debris field we see, from the few photos, suggests that it is composed of mostly "heavy stuff" so lighter objects would be well to the West, elsewhere, or may have left prior to impact. An A330 at touchdown and flaps extended carrying 140 knots with a deck angle of 16 degrees (17 is tailstrike time) has an angle of attack (Flap) of 45 -50 degrees. At 210plus knots (AF447), even with flaps stowed (BEA), and an AoA of 80 degrees, well, one starts to see the stress on the Flap. If the Flap is Prised loose and leaves, the spoilers will follow quickly, having lost their "Prot" from the massive flap. Their plane of stress is opposite stress from beneath.

This is as close as I'll get (at this point, having softened my admitted alarmist views at the outset) to suggesting that the a/c was other than completely intact at impact. My suggestion is that as much as I appreciate and admire your courage in suggesting it, not even bear thinks this airframe had greater than 90 degreees of AoA ('backward' flight)...

"As Close as I'll get..."
 

Bear,
I've been following this and the other threads since June '09. I'm sure many will agree that you've been a lot closer than that to suggesting the A/C was not intact before impact. ;)
Andrew

Hi,

Returning far away in the past .....

http://i54.tinypic.com/2cr7fcn.jpg

I still feel very frustrated when I look at this map ...
It shows the very first search (aerial) and then those incurred by the BEA
The abyssal plain where you found the wreck was known (from a hydrographic point of view)
I always wondered why the BEA has not started his research in this area .. that was easy enough to explore .. compared with the rugged surroundings.
Over the means used during the aerial search (detection signal pingers) was not very efficient.
The negative results of the aerial search in this area (other than find some floating debris) should not be considered conclusive IMHO

Sorry, don't follow - the cabin alt is controlled at max 8000ft equivalent, this will only rise rapidly due to a breach in the pressure vessel - potentially triggering that ACARS message. Surely this variant A330 certified for cruise at more than 40,000ft and no need for cabin air pressure controller to dump cabin air to protect max pressure differential (9psi)?

I do not think it is the engines management system responsibility to report on flight envelope issues, and if the ADM feeds of Temperature, Pressure, Mach Number etc are marked as suspect wouldn't the FADEC be forced to just look at its own data? - can't be FA if critically dependent on external systems! Isn't it possible that from engines own measurements of total pressure inlet, EGT etc the engines themselves were operating within expected performance parameters, and just reporting rotational speeds and effective thrust output back to the higher level systems as normal. 'Rollback' due to ice build up would be different, with the engine able to sense that it wasn't working as expected, so maybe this didn't happen.

There is an interesting story in the online edition of the New York Times (the former International Herald Tribune website) about the search for the aircraft and what has happened since then - here is the link:
What Happened to Air France Flight 447?.

It's not a scholarly article, but it is of a much higher quality than most stories that have been published about the search. Makes for interesting reading.

For flights longer than 2.5 hours the cabin altitude is 7350 ft (*), pressure 11.19 psi. The safety relief valve setting is 8.85 psi (*), so it would open when the ambient pressure passes through 11.19 - 8.85 = 2.34 psi, i.e. FL429.


EDIT:: (*) Reference FCOM 3.01.21 p.1; Operating Limitations; AIR COND/PRESS/VENT; Cabin Pressure.

jcjeant, machaca,

I don't remember anyone at the time thinking that the bodies and recovered debris could have drifted so far from where we now know the wreckage is located. IMO the mistake was to waste the time pursuing the spurious pinger signals "discovered" by Thales. Without this diversion the wreckage might have been found in phase three.
M

Hi,

:confused::confused:

'Morning, sensor validation,
Quote:
"Sorry, don't follow - the cabin alt is controlled at max 8000ft equivalent, this will only rise rapidly due to a breach in the pressure vessel - potentially triggering that ACARS message. Surely this variant A330 certified for cruise at more than 40,000ft and no need for cabin air pressure controller to dump cabin air to protect max pressure differential (9psi)?"

As I implied in my post, I'm suffering from:
(a) not knowing the A330 pressurisation schedule;
(b) lack of expertise in calculating the diff-press for different pairings of cabin-altitude versus aircraft-altitude;
(c) lack of ability to calculate precisely the cabin-VS versus aircraft-VS at a fixed diff-press.

From experience on the A320, A310 and other jets, however, I can tell you that at the medium cruise altitude of FL350, the cabin altitude would not be scheduled as high as the maximum-permitted 8000ft. My GUESS was that it would have been about 6000ft, for passenger comfort. This would keep the diff pressure below the normal maximum of eight-point-something. During even a brisk step-climb in normal operations, that gives the controller a relatively easy task to climb the cabin at the maximum desired rate of 500ft/min without max-diff being reached.

You are quite correct to point out that, in normal (planned) flight at the max cruise alt of FL410, the cabin altitude would be about 8000ft, and the differential pressure would be just below maximum.

Chris

PS
While I've been drafting this, HN39 has kindly provided us with some figures that are significantly different from what I had in mind. As you can see, if he is right (and he normally is!) grity's zoom climb to FL416 would not trigger the cabin VS warning.

Pretty sure the BEA are frustrated too, however there were good reasons for what they did. The area around LKP, including the actual crash site, was searched in previous phases both aerial search for wreckage and undersea search for pingers (not aerial search for pingers - if that is what you meant). The third phase, which that map shows most clearly, was setup based on drift analysis of the wreckage - turned out to be wrong. Frustrating, yes, but understandable given the number of variables and the long delay in finding the floating wreckage.

As to why not look in the easy-to-search areas, the answer is that they had already been searched and no pingers detected. The assumption was that (at least one of) the pingers worked, so if sat on the abyssal plain, they would have been heard. In rugged terrain, pinger signal would be more likely to be blocked, so what do you do if you have not heard the pingers and the drift analysis points to a rugged area - search a different flat area because it's easier ?


The big questions (I think) are:

1 - did two independent pingers both fail (and if so why) or was the pinger search not done right

2 - how did the initial air & sea searches miss the floating wreckage despite apparently searching the right area

Swordfish41;

Thales and the French MOD's part in Phase 3 leaves a lot to be answered. I believe the shutters have come down on that episode, and the true reasons behind what went on will never be disclosed.

You might say that Thales got a little bit out of their depth when deciphering 37kHz pings in the noise. A hearing aid should have been on the shopping list.:}

Hi,

It's not that assumption you wait from professionals.
They know that pingers can fail ... or be destructed for any reasons
In a investigation you must be open at all options and exploit them
If you base an investigation on dogmas and faith .. it's the best way to fail.
So .. this assumption is even more frustrating for me.

jcjeant.
in the graphic of the crash and search phase areas that you posted, there is a seperate box to the south of the hemisphere that was intended to be covered in phase three. The square was searched because Thales reprocessed a sonar search carried out by the French Nuclear sub L'Emeraude, and believed that they could discern amongst the noise, signals from the pingers on the voice and data recorders. Because of this the methodical survey of the phase three area was interupted, and never completed. If it had been, the wreck would have been located. I understand that there was some disagreement amongst the search teams at the time about the wisdom of this diversion, but they were overruled.

Towed Pinger Locaters (TPLs)
 

Undoubtedly the BEA will be examining closely the reasons why the ULBs were not located.

In my mind, there are three reasons:-

The first is that there was only one functioning ULB as the ULB attached to the SSFDR was most likely struck and destroyed by collapsing frames/stringers during the initial impact sequence. Hence the reason it was not attached to the recorder.

Secondly, the SSCVR ULB finished up getting itself half buried and the effective output was attenuated by at least 3dB along with some localized acoustic absorption caused by silt. Overall, the radiated output was down by around 6dB as the polar ends are the least effective radiators.

Thirdly, the grid line that was being towed was north/south from 3°30'00"N to 3°00'00"N and there was the potential for the TPL to drift off line caused by the turn to head back north on the next grid line. If the towing vessel turned too early and gave insufficient time for the TPL to realign, then there was a potential for a nadir (shadow) to appear in the area covered by the tow. The fact that the tow (now deep at around 3,000m) may not have lined up until some time after passing through 3°00'00"N provided the opportunity for missed detection, coupled with the lower than optimal output by the remaining ULB.

The answer lies in the detail.:8

MM43
I too am sure that it will never be revisited, at least in public, although Thales themselves used to have an impressive undersea survey division. I suppose that coming to the final parts of the search area, without any results it was hard not to respond to new evidence, particularly given its source. That however is enough hindsight from me. I would just add that your explanation about the towed locator is why this technology provides far less certainty than AUV's.

Heterodyning ULBs
 

ULB Spec. With two ULBs close to one another, and transmitting, is there not a chance that heterodyning will affect the already very weak signals? Two RF transmitters would certainly be affected - so why not 37Khz?

What do you want then ? Guesswork ?

Some assumptions have to be made or you have an unfeasibly large search area. ACARS can fail too - and that is our only source of LKP. Suppose the a/c lost all power and glided from LKP - it could have been many miles from all the search areas.

The assumptions were (to me) reasonable at every stage, the area searched in phase 1&2 included the crash site - and failed to find the wreckage, but not because of looking in the wrong place.

Phase 3 was specifically targeted based on drift analysis of the floating wreckage.

Phase 4 went back with no assumptions except that the crash was in the vicinity of LKP (ie. no power loss and glide etc.), and no assumption that earlier searches were correct.


In theory yes, but historically they almost always work, and there are two independent that would have to fail. It was a reasonable assumption for the first two phases, after that it was not assumed.

Hi,

I don't agree with this supposition ... if you read the ACARS ..
BEA read also the ACARS ... and maybe more we don't know about.
I can't suppose a glide from them.

The aera phase 1 the very first research .. (with ASSUMPTION of working pingers) was not a large one !

And phase 4 included the crash site.
The difference is they used appropriate tools for this search and they assumed (this time) that the pingers were no more working !
At the first search .. they have two doors to open ....
One with the pingers in the room .. the other with no pingers ...
They opened only one door .

@Chris Scott ; HN39

Although that a violent pitch-up to FL 400+, (or "zoom-boom scenario" like takata it humorously called) was not direkt against one of the ACARS is very surprisingly for my myself, because I just searched the fastest way to reduce the horizontal speed, nothing else

but even the violent pitch-up itself need high aerodynamicel lift and will generate lots of g, how fast can we generate this climb without a wing-brake? 10 sec...15 sec...?

jcjeant;

With some help from takata for the last ACARS message (cabin v/s), the airplane could have started its final 130 nm glide from FL350 at 2:15:14.

grity; The zoom-climb doesn't have to be violent nor involve 'lots of' g. But what caused it?

mm43,
Do you know that for a fact ?

One thing that surprised me was the statement that the pingers could only be detected by TPLs at a bit more than 1700m. Important variables necessary to asses the valididy of this statement are missing: I couldn't find acurate specifications for the TPLs on the net. Another essential variable was the noise intensity spectrum level around the pingers frequency. In this frequency band I think that rainshowers and, to a lesser degree the sea state play a major role. But rainshowers randomness may make detection rely on luck.

Pitching Moment vs AoA
 

Possibly re-regoing over old ground here but Figure 6 from the following NASA report stands out to me ... (thanks to BJ-ENG for posting the link)
http://ntrs.nasa.gov/archive/nasa/ca...2005208658.pdf

So, from the wind tunnel data for this generic-ish twin engined commercial aircraft, we have a family of curves showing the relationship between pitching moment and AoA for elevator angles from -30 deg to +20 deg, all at a zero HS angle and CG mid.

How about we then factor in a typical HS cruise trim angle and aft CG and consider what sort of AoA value might start becoming a nasty place to be?

From page 4 (Section B) of the same document:



(

I think that mm43 is referring to the photo that BEA published of the CVR, showing the pinger half-buried. I am less sure than mm43 about the configuration of the transducer in the pinger case, and what the effect on the sound level would be of the way the pinger landed on the bottom.

I gave an approximate estimation of the the pinger detection range in this post:

http://www.pprune.org/5683946-post951.html

Rain would have some influence at 37kHz. I don't know if it was raining when they searched the now-known debris area. According to the classic text by Robert Urich, there is a 20db difference between the minimum and maximum ambient noise in the deep ocean, between calm and 30knts wind (sea state 0 and 6). I expect that rain is added to that, but light rain does not add much, as near as I can tell. I assumed about sea state 2-3 in my calculation.

Zoom climb initiation
 

HN39

It seemed to me some time ago that the 340 scenario was a prospective candidate to initiate this - and still is the leading possibility IMO. There has been a number of excellent discussions related to it over the last couple of days that have pointed to possible inconsistencies - that were subsequently potentially addressed by others i.e. ACARS, etc. Is this zoom then deep stall scenario within the capabilities of accurate modeling in a simulator?

As my old boss used to say: 'It's surprising what you can find out by flying the sim'.

Unfortunately no. The airframe makers do not test into the region that AF447 apparently found itself, and thus do not collect data for this region of "flight". Any such simulator runs would be purely speculative.

Such testing is inherently dangerous and real flight data sufficient for a simulator is unlikely to be available even for a stray event. Best hope is that Airbus collects data in a wind tunnel and compares it with CFD models (CFD=Computational Fluid Dynamics) to see if the results are rational.

Only up to the point where the manoeuvre gets outside the simulator's envelope of valid data, and that does not include the deep stall, at least not in the simulators intended for training of airline pilots.

May 9 update from BEA
 

Information, 9 May 2011

Simulation during control program development
 

Mahcinbird, HN39

I would hope that simulator programs would extend into regions beyond normal airframe testing limits - based on wind tunnel test data. One can expect aircraft to, on rare occasion, inadvertently experience excursions into extreme realms even with highly trained crews and state of the art instrumentation focused on avoiding it - as nature is always capable of surprising us. Take for instance the temperature fluctuations encountered by the A340 and implications for exceeding mach limit. One would expect such an event could be anticipated during control design but it probably would not be the subject of airframe testing. Thus it seems appropriate to do simulator modeling (during design) - to test the control design for the response to such an event. Perhaps that range of information is only available to A300 Engineering in their simulator.