Tech Log The very best in practical technical discussion on the web


Old 7th Jul 2009, 05:56
  #3161 (permalink)  
Join Date: Jun 2009
Location: Paris
Posts: 691
Likes: 0
Received 0 Likes on 0 Posts
At 420 knots the plane is going 7 miles a minute.
At 468 knots (Mach 0.82) and FL350, the plane is going 7.8 miles a minute, and this was F-GZCP configuration at 0210.

At 0210 we have a position. Presuming the plane was still aimed "roughly" along the planned flightpath and presuming perhaps some pushes by storm winds where could the plane have gotten to by 0214 plus maybe a little Kentucky Windage for surviving a bit past 0214?
Why would the plane be only "roughly" aimed along the planned flight path at 0210? By all calculations possible about her position, she was maybe slightly deviating 3 NM to the West, but not much. The "pushes from storm wind" is something you know for sure or is it your imagination?

This hard flying condition are not documented anywhere and all we can verify by sattelite IR imagery is that she was pretty close to the coldest part of the convective CB cells around her, but we are left without a single clue about how bad or good the real situation was. LH507 crossed 20 mn before with a 10 NM deviation West and reported only moderate turbulences. There is a strong evidence by F-GZCP thrust setting that she was not in severe turbulences at 0210. Please, give a little credit to the pilots and consider first that they will do what they were supposed to do, weather avoidance and turbulence procedures, until we got something like a hard proof that they failed to do so. The next Air France flight, 37 mn later, deviated 80 NM as it is not in the company policy to fly strait into bad weather for saving fuel. If they picked this way, it would have been painted safe (right or wrong) on their radar and they are supposed to have been trained well enough to use it, or we need evidence they were not.

"Where could the plane possibly be?" might be a better search criterion than "where do the currents extrapolate backwards to at the time of the crash?"
And, of course, I am fastened on to the apparent fact that the plane was either deceased or was no longer even approximately horizontal after 0214.
Again, it is mostly your imagination: what clue do we have for considering the aircraft 'deceased' at 0214? Unreliable air speed indications? Is it enough without adding a little bit more imagination: updraft, downdraft, storms, lightnings, stall, spin, spiral, dive, upsets, tailfin wiped out, computer glitches... crash?
All the facts we know for sure about the real weather situation is resumed in one word: ice. How we know it? The ACARS are telling us that the probes are freezing, that's all and we should stick with it as the primary weather issue encountered.

Aiming of the satellite antenna is somewhere between a TV yagi on a rotator and a pair of rabbit ears in a high signal area for reference. You can be a whole lot off and it still works. Plus or minus maybe 60 degrees of vertical for a 3dB (half power) loss. Within about 35 degrees you'd have trouble picking the optimum direction without automation.

This is why "what is the region it could have reached in 5 minutes or less is my search criterion. Then match that with the currents to try to refine it all. And I have an untutored hunch that turning an A330 or anything else at FL350 is not "sharp" by any stretch of the imagination unless you are purposely entering a dive as for a bombing run.
Then, all your 'search criterion' are mostly based on... imagination, but not on verified facts. Here is a relevant example where, if this incident would have ended in catastrophe a few miles before the runaway, without any communications and radar coverage, your imagination would have placed the crash site 85 nautical miles off the coast. This flight would have ceased to transmit any ACARS at 0626 @ FL345, after sending an engine failure at 0613 and when flying at FL390:

Air Transat Flight TS236, was en route at FL390 when at 05:36 UTC, the crew became aware of a fuel imbalance between the left and right-wing main fuel tanks. Five minutes later the crew concerned about the lower-that-expected fuel quantity indication, decided to divert to Lajes Airport in the Azores. At 05:48 UTC, when the crew ascertained that a fuel leak could be the reason for the possible fuel loss, an emergency was declared to Santa Maria Oceanic Control. At 06:13, at a calculated distance of 135 miles from Lajes, the right engine (Rolls-Royce Trent 772B) flamed out. At 06:26, when the aircraft was about 85 nm from Lajes and at an altitude of about FL345, the left engine flamed out. At 06:39 the aircraft was at 13,000 feet and 8 miles from the threshold of runway 33. An engines-out visual approach was carried out and the aircraft landed on runway 33. Eight of the plane's ten tyres burst during the landing.
Investigation has determined that a low-pressure fuel line on the right engine, had failed probably as the result of its coming into contact with an adjacent hydraulic line.


Last edited by takata; 7th Jul 2009 at 06:10.
takata is offline  
Old 7th Jul 2009, 06:00
  #3162 (permalink)  
Join Date: Jun 2009
Location: Colorado
Posts: 41
Likes: 0
Received 0 Likes on 0 Posts
I do enjoy reading what you big plane flyers write.

The nature of a flat spin is that recovery via unsustained flight control inputs is highly unlikely and, because of that high centrifugal force, anything not tied down in the cabin or holds would've migrated aft, further exacerbating the aft CofG complication for a flat spin.
On this particular point, I do believe that it's a standard spin where the centrifugal forces would migrate everything aft. In a flat-spin they go to the nearest extremity, what's in the cockpit gets forced forward.

Again, it is mostly your imagination: what clue do we have for considering the aircraft 'deceased' at 0214?
On this point I'm with JD given: the preliminary report of essentially vertical impact on the belly, the projected positions at 0214 being in the neighborhood of the initial debris finds, and the most likely reason a satellite connection was lost. They tend to pair up. All other possibilities get more complex.

Last edited by ttcse; 7th Jul 2009 at 06:12.
ttcse is offline  
Old 7th Jul 2009, 06:11
  #3163 (permalink)  
Join Date: Jan 2008
Location: Australia
Posts: 50
Received 0 Likes on 0 Posts
I think that some information that can be gathered from the raw BEA data might put paid to some of the speculation, especially if viewed chronologically. Here's my amateur attempt. There may well be errors in what follows (especially towards the end when I got tired), and I'm happy to be corrected.


If we look at the BEA report, we see that the aircraft was transmitting AOC position messages approximately every ten minutes, as it was set to do (sec 1.6.8 and

The map on page 11 has plots of positions at 00:29, 00:39, 00:50, 1:00, 1:10, 1:20, 1:30, 1:40, 1:50, 2:00 and 2:10. I assume these are all taken from those messages. Nothing thereafter.

The transmission sequence at (p 46) shows the position according to the message for 2:10, which was received at 2:10:34. That was about 24 secs after the first relevant ACARS message (AP off) was received.

Let's examine the lead up to this.

Radio and ATC

The last successful radio conversation was with Brazilian ATC at 1:35:43 (sec 1.9.1). At this point, the aircraft was at the "edge of radar range of the Brazilian control centres" (findings, p 68).

At 1:35:15, AF447 contacted Atlantico. (fwiw, transcript is: "AIR FRANCE FOUR FOUR SEVEN, by checking INTOL zero one three three, level three five zero, SALPU zero one four eight, next ORARO zero two zero zero, selcall check Charlie Papa Hotel Quebec." (annexure 3 p 99))

They thanked the controller at 1:35:43, but did not respond to requests made by him from 1:35:46 to 1:36:14.

Hypothesis 1: The crew passed from radio coverage with Brazil at this point, all still being in order.

AF447 made 3 attempts to connect to Dakar ADS-C: at 1:33, 1:35 and 2:01. These were rejected because of a flight plan absence or mismatch (sec

The conversation with ATLANTICO seemed to follow the first rejection attempt: the crew contacted ATLANTICO at 1:33:25 (sec 1.9.1; annexure 3 p 99). [I'm assuming all times in the report have been synchronised - see sec 1.15].

At this point they said they were at FL350 (as cleared).

Hypothesis 2: The crew was working normally at all three rejected connection attempts, as nothing was mentioned at 1:35, and I assume they would not have attempted to connect the third time if an emergency had already arisen. If correct, this suggests that they tried to connect to Dakar ADS-C about the time radio would be lost with Brazil, then contacted Atlantico presumably when they could not get on. They then tried to connect again around about the time they finished talking, and again 25 mins later.

Hypothesis 3: We know that there was never any radio contact between AF447 and Dakar (sec 1.9.2). In addition to ADS being rejected, the crew was simply unable to establish radio contact with Dakar. This was consistent with the experience of IB6024 ("difficulties") and AF459 (contact established around 3:45) (sec 1.18.3). LH507 (20 mins ahead) reports hearing nothing from AF447 on 121.5 throughout the flight: 1.18.3.

Observation 1: We are not told if anyone has attempted to identify who from AF447 was talking at that point - could put to rest some of the speculation about who was where on the flight deck.


At 2:01, based on the map at the bottom of p 30, the aircraft would have been very close to the southern edge of the CB activity. [Assuming for this purpose that the data in the title bar correctly shows that this was the plot for 2:00. Query extent of any averaging with earlier data done to produce this image.]

At 2:07, there is an image of the weather, apparently as at this instant: see sec 2.2.2 of annexure 1 (French report). [It refers to "l’image prise à 2 h 07", ie "the image taken at 2:07"].

If we estimate the position of the aircraft as 70% of the way between its position at 2:00 and 2:10, and align it over the image at sec 2.2.2 of annexure 1, we are into bad weather. The aircraft appears to be within the area of -70C, shown in figures 9 and 10 (which only show the position of the aircraft at 2:10), but not yet into the -75C weather shown in Figure 11.

By 2:10, the aircraft has either just touched on, or is just leaving, the area of -75C in Figure 11, assuming that area remains about the same within the 3 minutes since the image was taken. That is about the time (within 24 sec) of the first ACARS messages - ie AP disconnect -- and about the time of loss of windshear reaction, alternate law, autothrottle off, TCAS fault, rudder travel limit warning (p 49).

Note that these all occurred in a cascade at 2:10, even though some were not received until later: see the sequence at p 49.

Also note: "ECAM messages are transmitted in real time as soon as they are acquired. Flag or advisory messages are transmitted as soon as they have been confirmed. Fault messages are transmitted as soon as the corresponding correlation window is closed." (p 25)

Third note: There is an assumption that the position transmitted for 2:10 (+2.98, -30.59) is actually correct - see sec "The position transmitted was the aircraft’s FM position which, in normal conditions, is close to the GPS position." I haven't seen this commented on, but it means we should keep in mind the possibility that the actual position could have differed, if something had already gone wrong.

At 2:11, we have a sequence of ACARs messages received, showing that the ACARS system was still connected.

Hypothesis 4: The aircraft did not move more than 90 degrees in any direction from normal flight up to 2:13:14, as this would have put the antenna out of contact with the satellite (I assume a 180 hemispherical envelope).

At 2:12:51, the "ADR disagree" message is received at base. The time *logged* is 2:12, so it appears that no more messages are banked up at this point. [Note: I may be wrong, my understanding of ACARS is very imperfect]

Between 2:13:14 and 2:13:45, ACARS communication is lost, for whole or part of the period; "the gap observed between the message sent at 2 h 13 min 14 s and the one sent at 2 h 13 min 45 s is due, at least in part, to a temporary interruption in the communication link between the aircraft and the satellite" (, p 47)

Some time at 2:13, we have the Prim 1 and Sec 1 faults logged. They are received at 2:13:45 and 2:13:51 respectively. We also have the "FMGEC1 (1CA1) (2 h 13)" message, which "has not been fully explained at this stage of the investigation. The fact that it was “INTERMITTENT” means that the fault was detected for less than 2.5 seconds" (p 54). This is received at 2:14:20.

Some time between 2:14:00 and 2:14:14, the "maintenance status ADR2" is logged, and then received at 2:14:14. Some time between 2:14:00 and 2:14:26, the cabin vertical speed advisory is logged, and received at 2:14:26.

The last signal successfully received by the aircraft was transmitted at 2 h 14 min 28 s: page 47.

ACARS & Questions

It is at this point that data runs out, and it is not prudent to go any further without answers on known-good information.

According to PJ2, ACARS uses VHF if possible, otherwise Satcom [post 791, ]. The BEA report states that all were sent by SatCom. Corollary: no VHF, consistent with no comms to Dakar.

ACARS is also powered by the AC1 bus.

Question 1: to what backup level is ACARS powered. Say we lose AC1. Does ACARS lose power & ability to transmit? If not, at what point does it lose power? Cessation at 2:14 might only signal loss of power up to the point at which ACARS ceases to run, but there may still be power to other systems. If made out, this would direct attention towards partial/total loss of power as a factor.

Question 2: what is the behaviour of the aircraft on loss of Bus 1 - eg by lightning. How long to recovery of any power, and what happens? This will affect hypotheses why there was a loss between 2:13:14 and 2:13:45 -- could it have been electrical, rather than attitude (or both)?

Observation 2: the permanent cessation of ACARS messages after 2:14 could also be due to attitude of the aircraft. But if BEA is correct, and the aircraft at least struck the ocean flat and vertical, then (assuming the plane was intact at that point) query whether attitude was an issue -- based on geometry, the antenna could have gotten a lock even at this point. This suggests there must be some other cause for why it didn't.

Observation 3: I don't recall people having canvassed the potential of power loss on FDR/CVR recording -- query if we may be disappointed even if they are found, as in SwissAir 111 or TWA 800.

Hypothesis 5 (speculation): Could we be dealing with weather induced disconnects and/or unreliable airspeed up to 2:13, followed by a power problem. Possibly unrelated - perhaps lightning after all? Engines out? First an interruption (2:13:14 to 2:13:45) then permanent loss up to some threshold that is not sufficient to power ACARS (after 2:14:28). Or could it be weather induced problems up to 2:13, then an attitude-induced loss of ACARS from 2:13:14 to 2:13:45, followed by a power-induced loss of ACARS (and again, query what level of loss).

Hypothesis 6 (pure speculation): Faced with this kind of sequence, assuming you had some control, what would you do? Head to Fernando de Noronhas (nearest airport) or Natal? Descent and U-turn? Might we even be reduced to a glide? It is here that Takata's excellent work ought to fit in.
auraflyer is offline  
Old 7th Jul 2009, 06:43
  #3164 (permalink)  
Join Date: Jun 2009
Location: Paris
Posts: 691
Likes: 0
Received 0 Likes on 0 Posts
Again, it is mostly your imagination: what clue do we have for considering the aircraft 'deceased' at 0214?
On this point I'm with JD given: the preliminary report of essentially vertical impact on the belly, the projected positions at 0214 being in the neighborhood of the initial debris finds, and the most likely reason a satellite connection was lost. They tend to pair up. All other possibilities get more complex.
1. What clue, on this impact description, is revealing to you the location of the crash? Why would she had to impact, "level with strong vertical acceleration" only exactly at the spot (or very close to it) where the contact by ACARS was lost?
2. The bodies drifted 100 NM (to the North) in five days after the first recovery but would have drifted previously 10 NM during the same time, and in the wrong direction (to the South-West), in an area where the drift is much stronger?
3. Why are they not able to detect the flight Recorders on this primary search location?
4. Why should we consider only the more extraordinary circumstances to fit this theory when we already have a sufficient one (icing) which may explain what happened?

I'm sorry, but the complexity of the postulat is not on the side of those I've made. I fully agree that my postulat is perhaps less obvious at the first look, but it is certainly fairly simple when you look at the whole situation much closer.

takata is offline  
Old 7th Jul 2009, 07:09
  #3165 (permalink)  
Join Date: Jun 2009
Location: LA
Age: 73
Posts: 50
Likes: 0
Received 0 Likes on 0 Posts
Tekata, unless I'm misreading something (quite possible) I think you misinterpreted the question in JD-EE's post. I think he was saying something like
  • assuming they were on course at 0210 and then decided they were in a bad place and for some reason decided to turn, what is the turning circle/radius at FL350 and 468kt?
  • Assuming they turned starting at 0410, and further assuming they hit sometime near 0414 or a little after, where could they have hit?
Its my opinion that the impact point was some place they could physically reach in a descent from FL350 to FL0, with the descent and any associated direction change starting no sooner than 0210.

I'd like to know how sharply they could have turned in Alternate 2, starting at FL350 at 0210, and possibly descending thereafter.

In other words, can we match the airplane characteristics to the drift data and end up at the same point in the water?
WhyIsThereAir is offline  
Old 7th Jul 2009, 07:42
  #3166 (permalink)  
Join Date: Jun 2009
Location: I am where I am and that's all where I am.
Posts: 660
Likes: 0
Received 0 Likes on 0 Posts
Remember, the 0214 time is the last received time for any datalink from AF447. Reversion to anything other than 'Normal AC Power' would have turned the ACARS system off. The plane could have been airbourne for some time after that, expanding the search area requirements.
Very good point. I still think it would help to define a minimum search area.

JD-EE is offline  
Old 7th Jul 2009, 07:56
  #3167 (permalink)  
Join Date: Jun 2009
Location: Paris
Posts: 691
Likes: 0
Received 0 Likes on 0 Posts
Tekata, unless I'm misreading something (quite possible) I think you misinterpreted the question in JD-EE's post. I think he was saying something like
  • assuming they were on course at 0210 and then decided they were in a bad place and for some reason decided to turn, what is the turning circle/radius at FL350 and 468kt?
  • Assuming they turned starting at 0410, and further assuming they hit sometime near 0414 or a little after, where could they have hit?
My bad if I misunderstood something, but have a look at the situation which could develop and tell me what you would do in the seat of F-GZCP's driver?

0209.. heading to TASIL, cruise, slightly deviating to the west in order to avoid a suspected dangerous cell.
0210.. AP/A/THR kicking OFF
.. ALT2, etc, warnings
.. thrust and AOA set to adequate values (N1=91, 3.5 deg?)
.. de-icing ON
.. troobleshooting start
0214.. engines rollback (ice)
.. pitch down to avoid stall
.. start emergency descent
.. cabin vertical speed advisory
.. engines flame out (after 15 mn of flight in CBs, mostly in Anvil)
.. EMER ELEC => ACARS fails, only HF1 & VHF1 left

What to do now?
1. aviate: avoiding stall?
2. navigate: where we go?
.. Cap Verde: ahead but too far
.. Brazil (Fernando Noronha) turn needed without colliding an active cell
3. communicate: Mayday? we'll see later, hands full now

What could be the turn radius in this case? Certainly the best possible in order to keep the aircraft under human control and not to lose too much speed/altitude? The turn-East option seems locked because of the bad cells, so turn to the West?


Last edited by takata; 7th Jul 2009 at 08:06.
takata is offline  
Old 7th Jul 2009, 08:14
  #3168 (permalink)  
Join Date: Jun 2009
Location: I am where I am and that's all where I am.
Posts: 660
Likes: 0
Received 0 Likes on 0 Posts
takata - used 420 as a bogey value. Fill in your real value.

singpilot came up with a good alternative. Another is that there were no more ACARS messages to send. The former is a very real possibility. The latter seems to my untutored intuition to be far fetched.

If the plane was within 20 degrees of horizontal it could communicate any ACARS messages that might be generated. Loss of power could do that. The fact that the plane hit near horizontal with, as I read it, a strong downward component to its velocity and some unknown speed along its heading.

I am not trying to identify why. I am simply trying to identify an approximate range of when. And absent an ability to continue on with battery power, somehow, there are not many things that would have it still in one piece much after 0214. Given a position report of 0210 and a presumption that the plane MAY have initiated a turn slightly before the report was generated where could it have gotten to? First guess is about 40 nautical miles from the position report's position. Within that radius lies the plane if the sessation of ACARS messages means it had no more problems it could report or did report before conact with the water.

If it came apart at FL350, that fits the criterion. If it went onto battery power, that fits the criterion and expands the search area. If it hit the water at 0214+, that fits the criterion. That it was under power and under controlled flight after 0214 does not fit the criterion.

How it got there is another question I am NOT trying to answer. I'm simply wondering "where are the most fruitful places to initiate search for the remains?" I would contend that the most probably circle for searching would be wthin 40 nautical miles of the 02:10:34 position report. Can that be refined? I'd presume yes, with a slightly smaller number, perhaps. I pushed the time range by allowing some delays on generating the reports.

Do remember, a long time ago I asked whether there might have been other ACARS messages cued after 0214. Nobody saw fit to answer that. Can you?
I am open to answers. I'm not open to screaming fits.

JD-EE is offline  
Old 7th Jul 2009, 08:16
  #3169 (permalink)  
Join Date: Mar 2003
Location: BC
Age: 75
Posts: 2,482
Received 0 Likes on 0 Posts
BTW, neither one of you answered my question with respect to the pilot’s ability to deliberately take the aircraft out of Alternate Law and force it into Direct Law if he so desires and I don’t know the answer myself. Although someone posted a fine diagram indicating which protections were retained in Alternate Law and which were lost, I do not recall them exactly. Safetypee you said you believe that stall protection is lost in ALTN Law but what I want to know is: Can the pilot go to Direct Law and remove all “protections” if he chooses to do so? From what I've read, stall protection is lost but stall warning is available from the AOA source.
The ability to take the Airbus into Alternate or Direct Law resides within the design of the autoflight system but is a profoundly non-standard approach with this design and is entirely within test-pilot territory.

I state this strongly because it is not even in the realm of an "ad-hoc, emergency response" to a badly degraded aircraft or flight control system. I could not see this kind of system intervention condoned or even considered by any Airbus pilot nor can I see it as a legitimate response in the present discussion.

This is because the flight control system, in Normal Law, does not prevent the crew from doing what is necessary to fly the aircraft within (and slightly outside of) expected maneuvers.

I think what you're asking is, is there a switch like there is in the 777 that connects the control column/wheel "directly" with the flight controls such that the pilot can exercise as much control input as is deemed required to intervene with abnormal regimes and attitudes. The answer, for both types, is, "no" in the sense that there is no more authority than is already granted by the flight control design when the aircraft is in such abnormal states, (system failures or abnormal attitudes). I have previously posted information on this twice.

The 777 switch bypasses the (equivalent of the) primary flight control computers but still sends electrical signals to the flight control servos in "direct control" (another name for direct law). In other words, control is achieved by wire, not cables/pulleys. (I assure you that I am familiar with and have flown the DC8, DC9 and B727 and have flown "manual reversion" in all these types in simulator exercises. The A320 can be flown on engine thrust and stab trim alone, to a successful landing). This is exactly the same as the Airbus 330/340 system except that the FCPC and FCSC - primary and secondary flight control computers execute the direct flight control orders from the stick. The available authority from the sidestick in alternate 2 law and direct law is, with minor variations in alternate 2 law, the same as would be available to a 777 crew in the same circumstances.

The essential decision by the design and engineering people is, because the autoflight system is beyond it's design capabilities either during a serious system failure (hydraulic, electrical or data-loss) or a "jet upset", (> 10deg ND, 30NU, 50deg roll, approximately), it can no longer reference and interpret the situation the aircraft is in and necessarily hands control over the flight crew.

This design is not a matter of "handing the aircraft over at the worst possible time" even though that is what it looks like - it is a matter of design practicality and software capabilities in rescuing a serious loss of control. In the Airbus, loss of hydraulics or electrics means that the autoflight system may not be capable of exercising all the control that is required of it when in Normal Law and so it is designed to degrade to Alternate Law in the 320, and Alternate 1 and 2 Laws, then Direct Laws in the 330/340. There is nothing about this that either intervenes in pilot direct control or prevents a pilot from flying the aircraft as s/he will. Any one of five flight control computers on the 330 will provide full use of all flight controls without restrictions. I have posted many schematics and diagrams illustrating both these laws, the degradation of same and the aircraft attitudes and/or system faults which will cause Normal Law (C* law as it is known) to revert to Alternate and Direct laws. The 777 degrades in approximately the same way.

I have to emphasize that there is nothing in the A330 design in terms of pilot interventions that were not available in the B707/DC8 design. The flight control computers do not mysteriously "modify" pilot input to do what the engineers and designers really want but haven't told the flight crew.

To be clear because there is a question, in Alternate law 1 & 2, pitch law is referenced to 'g' loading, (same as Normal law). The AOM does not specifiy that Pitch Alternate Law restricts 'g' loading to "2.5" but I suspect they do. In Direct Law, the aircraft is a "DC8".

The essential question is, "can the pilot get whatever 'g' s/he can pull above certification limits (2.5 positive, 2.0 negative, flaps up)? The answer is, yes in Direct Law and likely no, in pitch alternate laws 1 & 2, (1 is the same as 2 in pitch).

I hope this is of some help - I'm being a bit "direct" only to save space and not to dismiss concerns. This is the way the system works - there is only complexity, but not mystery. To explain the entire 330 autoflight system here would take substantially more space and would probably not be necessary.
I ask this question because the entire concept of the automated control system seems to be that it will prevent undesirable positions/attitudes, speeds, etc., as long as we are not in Direct Law. If we carry that concept to its limits - upsets, stalls and overspeeds simply can’t occur – so there is no need to waste time or money training for recovery from them. I do not believe that to be truth.
Not entirely true, and for the reasons described above. Otherwise, we have a fully-automated airplane in which the pilot is just so muich "spam in a can", as John Glenn said to NASA engineers when they refused to give the astronaut some control over his Mercury capsule. In the end, it was Glenn who saved a mission via skill and manual control when the computers could not deal with a stuck thruster...

I therefore agree completely with your "If we carry that concept to its limits...etc". As much as it seems reasonable on the surface, we can't have it both ways.
Every effort and tons of money have been made and spent to keep the Space Shuttle safe yet the vehicle has been lost twice. We build things that we say are ‘fail safe” but nothing is truly fail proof.
The latter part of the statement is absolutely true. Some have mentioned Murphy's Law and variations on the theme.

Even without a creeping hubris focussed on "the mission" rather than the traps, a characteristic which can infect and even convince the most skeptical to ignore the "nay-sayers", and under the very best of managerial and engineering intentions and comprehension of the risks at hand, we can still fail extremely badly.

It is not that failures occur of course; It is what we have done with the inevitable but frequently ignored precursors to a major accident that we must surely reflect upon. For me, that is the case here, but truly that is an old aviation story indeed. What happened to AF447 after 0215Z is immaterial in terms of systems design, aircraft response and crew handling.

I note that the shuttle accidents are referenced in comparison to automation accidents regarding "tons of money" etc. I disagree with this comparison and the reasons stated, and although it is perhaps a side-bar, I would suggest, if I may, a close reading of Diane Vaughan's book, "The Challenger Launch Decision", and Moshe Farjoun's and William Starbuck's edited book, "Organization at the Limit" for a full understanding of why the Challenger and Columbia disasters occured. In almost every aspect, they bear very little relationship to this accident.

Last edited by PJ2; 7th Jul 2009 at 09:22.
PJ2 is offline  
Old 7th Jul 2009, 08:26
  #3170 (permalink)  
Join Date: Jun 2009
Location: I am where I am and that's all where I am.
Posts: 660
Likes: 0
Received 0 Likes on 0 Posts
takata, sorry if I'm grinning ruefully here, sir. I don't know. Essentially I am asking regardless of the pilot's decision how far away from his position report at 02:10:34 could he have gotten in the time between 0210 and 0215.

It has been pointed out he could be further away on what, if I understand the power situation correctly, would have been an unpowered glide. That might extend the search for the second level of probability.

When we can draw those circles we learn something. If we can map drift backwards into one of those circles that means something important. If we can't, well, I notice the positions and the drift rates diagram simply do not seem to fit at all well. That suggests a problem with the data or our interpretation of it. Scrap the hypothesis and work on another. Your work is good thinking. It just arrived at a bad conclusion, likely because of presumptions about the currents that were not true. It's not your fault or anything. It's our groping for understanding with inadequate data.

edit - change an It's to be more specific.
JD-EE is offline  
Old 7th Jul 2009, 10:15
  #3171 (permalink)  
Join Date: Jul 2009
Location: France - mostly
Age: 84
Posts: 1,688
Likes: 0
Received 0 Likes on 0 Posts

tanaka (post#3170)

quote: "What could be the turn radius in this case? "

Turning at 468 kt TAS at various angles of bank:

Bank angle (deg) 15 30 40
Loadfactor 1.035 1.155 1.305
Turnradius (NM) 11.9 5.5 3.8
Turnrate (deg/s) 0.625 1.348 1.96

Does this help?

Last edited by HazelNuts39; 7th Jul 2009 at 10:19. Reason: improve readability
HazelNuts39 is offline  
Old 7th Jul 2009, 10:53
  #3172 (permalink)  
Join Date: Mar 2001
Location: us
Posts: 694
Likes: 0
Received 0 Likes on 0 Posts
One comment, one note, two images.

I assume the turbulence penetration speed for an A-330 is not mach 0.82.

I re-found the lat & long for the VS recovery point on June 7 (unofficial source): 3.41N, 30.42W. The first body(ies) recovered were June 6 at 3.34'05"N 30.27'18"W. There is a progressive northward recovery of bodies and wreckage over time, although recoveries in the latter days of the search are significantly more north and west of Tasil. (I assume the searching vessels were on courses from south to north as they recovered. If they had been on a north to south course, they would have intercepted the drifting wreckage sooner.)

This is the French bathymetric profile. No coordinates are given, but if the grid is the same as the chart prepared for the bottom sediment, it is roughly centered on 3.0N and from 30.0 to 30.65W

Unfortunately, the western sectors of the search area are the more mountainous and rugged, and likely to impede locating and/or retrieving the FDR and CDR.
SaturnV is offline  
Old 7th Jul 2009, 12:14
  #3173 (permalink)  
Join Date: Sep 2007
Location: Germany
Posts: 152
Likes: 0
Received 0 Likes on 0 Posts
Originally Posted by takata
What to do now?
1. aviate: avoiding stall?
2. navigate: where we go?
.. Cap Verde: ahead but too far
.. Brazil (Fernando Noronha) turn needed without colliding an active cell
3. communicate: Mayday? we'll see later, hands full now
I see the logic in your conclusion, but how long would it take to glide down from FL350 over a distance of ~100 nm ? (15-20 min?). I think the closer they would get to the surface the higher their priority to communicate would become. Taking into acount that in this scenario they would have had sufficent time (and a crew of three) one would expect distress calls to be made on 121.5 and to be picked up by nearby a/c.

Last edited by OleOle; 7th Jul 2009 at 13:12.
OleOle is offline  
Old 7th Jul 2009, 12:21
  #3174 (permalink)  
Join Date: Aug 2001
Location: England
Posts: 17
Likes: 0
Received 0 Likes on 0 Posts
Just had this email sent through...

Dear All, Bonjour.

I would highly appreciate if you could answer me asap regarding our question below; we would need the answer this afternoon by latest.

many thanks for your precious and much appreciated cooperation.



Request from SNPL Air France ALPA - M. Louis JOBARD, President of SNPL Air France ALPA.


Could you please tell us which airlines in your country, whom your union represents the pilots,

have already retrofitted (or decided to do so) their A320/A330/A340 with the system :

BACK UP SPEED AND ALTITUDE DISPLAY suggested by Airbus since 2007.

Many thanks for your diligent and precious cooperation,

With my best regards,

International & Technical Coordinator SNPL - France ALPA
Email: [email protected]
Direct land line: +33.(0) / Office: +33.(0)
Mobile: +33.(0)686.709.702
dougydog is offline  
Old 7th Jul 2009, 12:54
  #3175 (permalink)  
Join Date: Jan 2008
Age: 76
Posts: 53
Likes: 0
Received 0 Likes on 0 Posts
PJ2, it would be difficult to do better than your exhaustive answer. But to help (maybe) answering the question "switching deliberately from alternate to direct law" and as a former airbus driver I knew that at least, what would come to mind when you are not a test pilot, when there is no test flight engineer sitting back in the cabin with all test equipment would be ...lower the landing gear. You are in direct law and you put your hand on the trim wheel and you understand, now, what your trim balance is ...sorry if this seems very simple-minded and even stupid, but when you are lost with out-of-trim airplane and inefficient stick inputs, at least you go back to basics. It would have to be a very frightening situation to go to such extremes (apart from landing gear speed and altitude limitations which you maybe would not care about anymore) but it might save the day, only, of course, considering that the alternate law might be going mad at the absent airspeed(s) informations.
NARVAL is offline  
Old 7th Jul 2009, 13:39
  #3176 (permalink)  
Join Date: Dec 2003
Location: UK
Posts: 24
Likes: 0
Received 0 Likes on 0 Posts
Flat Spins

Thank you, Surplus1; for your post # 3061 and post #3144.
And thank you, Woodvale; for your post 3155 too.

It seems to me we are pretty close to saying much the same thing, certainly there are many big ifs involved at this stage and maybe always will be. My point was and is to take what seems to be judged as “known” or believed by competent authority and see if it fits anything known in the world of aerodynamics. If there is a fit, and it seems to me that there might be, surely it is interesting and should lead one to wonder about less sure assumptions.

Here, for all you spin skeptics, is a little food for thought and although I may, in the large part, be preaching to the converted, it seems it still needs to be said.

A modern swept wing airliner has somewhat different stall characteristics from straight-wing aircraft that many of us may have stalled when training in the past. The difference is that although straight wings tend and are encouraged to stall at the root first, aft-swept wings tend to want to stall at the wing-tips first, rather than at the root. With a swept wing, the Center-of-Lift may move ahead of the Center-of-Gravity (CG) at the stall causing a pitch-up moment. And so the aircraft may begin descending in a nose up attitude unless positive forward pressure on the controls is applied (Straight-wing aircraft will usually pitch down on their own accord when the stall occurs, as long as they are within aft CG limits.). That is why swept-wing aircraft have stick shakers that give an artificial warning of impending wing stall. Many also have stick pushers to force the aircraft to a lower angle-of-attack before the stall progresses too far. Many types of aircraft will tend to roll or yaw if recovery from the stall is delayed, and swept wing aircraft are particularly prone to becoming laterally unstable as the stall progresses. A spin can be thus inadvertently entered. An aircraft might be very stable in “normal” conditions, but once you find yourself at stall speeds; the rules change.

To further complicate the problem, those aircraft with engines mounted underneath the wings can experience a further pitch up, since the engines' thrust lines are below the aircraft's CG. The pitch-up associated with adding power can cause the stall to worsen, if the controls are not moved forward to counter this tendency, or if emphasis is not placed on lowering angle-of-attack first with forward pressure on the controls.

Swept-wing aircraft do not normally have the luxury of engine or propeller wash over the horizontal tail and elevator, to aid in pitch control. Therefore, if a pilot encounters an impending stall in such an aircraft, he has been taught to hold the pitch attitude and apply maximum power to minimize altitude loss and to "fly" out of the stall. The success of this recovery lies in the fact that a stall has not yet occurred (the stick shaker will typically activate at an airspeed 5-10% above the stall speed.). Thus it is not really a stall recovery at all, since a stall has not occurred.

It also seems that there are a variety of ideas regarding just what a flat spin is, clearly, a number of them are quite wrong. Let me just say that regardless of how it was entered, in a flat spin one wing will have a large angle of attack and the other a smaller angle of attack, in comparison to the relative wind, which is CB activity could be coming from almost anywhere. Both wings might be stalled, but not necessarily so and one will be stalled more than the other. Thus, you can have a doubly stalled flat spin and also a singly stalled flat spin.

In a non-spinning airplane, if one wing were producing more lift than the other, that wing would rise. So the question is; why is a flat spin stable? Or, why doesn’t the outside wing continue to roll to ever-higher bank angles? The answer is centrifugal force. In an airplane spinning about a vertical axis, the high (outside) wing will be centrifuged outward and downward (toward the horizontal), while the low (inside) wing will be centrifuged outward and upward (again toward the horizontal). In a steady flat spin, these centrifugal forces cancel the rolling moment that results from one wing producing a lot more lift than the other.

As has been mentioned previously in this thread, in the 1970s, NASA conducted a series of experiments on spin behavior. They noted that there was “considerable confusion” surrounding the definition of steep versus flat spin modes, and offered the classification scheme shown here.

SPIN MODE:............STEEP;.....Mod’ly Steep;..Mod’ly Flat;...FLAT;
ANGLE OF ATTACK...20 to 30... 30 to 45....... 45 to 65........65 to 90
NOSE ATTITUDE......extreme nose-down....... less nose-down
RATE OF DESCENT...very rapid.....................less rapid
RATE OF ROLLK.......extreme....................... moderate
RATE OF YAW.........moderate..................... extreme
of ATTACK.............modest........................ large
nose-to-tail DIFFERENCE
in SLIP..................large........................... large

The angle of attack that appears in this table is measured in the aircraft’s plane of symmetry; the actual angle of attack at other positions along the span will depend on position. Notice also that in a flat spin the rate of yaw is extreme and angles of attack are quite large. indeed.

In all cases NASA studied, the flat spin had a faster rate of rotation (and a slower rate of descent) than the steep spin. Meanwhile, there are reports of experiments in which the flatter pitch attitudes were associated with slower rates of rotation. This is not a contradiction, because the latter dealt with an unsteady spin (with frequent changes in pitch attitude), rather than a fully stabilized flat spin. A sudden change to a flatter pitch attitude will cause a temporary reduction in spin rate, for the following reason.

In any system where angular momentum is not changing, the system will spin faster when the mass is more concentrated near the axis of rotation (i.e. lesser moment of inertia). By the same token, if the mass of a spinning object is redistributed farther from the axis, the rotation will slow down.

When the spinning airplane pitches up into a flatter attitude, whatever mass is in the nose and tail will move farther from the axis of rotation. Angular momentum doesn’t change in the short run, so the rotation will slow down in the short run.

In the longer run — in a steady flat spin — the aerodynamics of the spin will pump more angular momentum into the system, and the rotation rate will increase quite a lot. The rotation rate of the established flat spin is typically twice that of the steep spin.

Recovering from an established flat spin requires forcing the nose down. This brings the mass in the nose and tail closer to the axis of rotation. Once again using the principle of conservation of angular momentum, you can see that the rotation rate will increase (at least in the short run) as you do so.

As for the point of rotation, it is likely found some way from the inner wingtip following the axis of rotation from outer wingtip via c of g to inner wingtip to point of rotation – the aircraft does not “spin like a top” as some seem to think.

Lastly, in turbulence with a high angle of attack and a stalled wing, you do not need a rudder to induce a spin; adverse yaw brought on by aileron drag in an attempt to level the wings will do just fine. Which is likely why, when learning, we were always taught to use rudder only; at high angles of attack leading up to Insipient Spins. Power will hold the aircraft in the spin; as long as the engines produce thrust and I suspect there would be little if anything the crew could do to recover, regardless of skill, especially on instruments.
Do you think you could recognize a spin in IMC?

The mentioned Russian aircraft that crashed following a “flat spin” entered at altitude, in turbulence, had a Second Officer who was a qualified aerobatics pilot - not that it helped.

So far, although there are those who scoff, no one has yet suggested a solution to better fit what little is known of the end result; damage observed from recovered debris or any of the other evidence, scant though it may be, and I suspect if taken a little more seriously it would be found that other things fit a spin solution too.

Yes, there are many unknowns and much we don’t know but IF AF447 DID fly into moderate to severe turbulence and IF that did cause and inadvertent loss of control and IF that is what the ACARS messages started to try to tell home base and IF yaw and high angle of attack caused loss of power and loss of satellite communication. Would the aircraft likely end up as it did? The idea is to suggest a solution that allows the evidence to fit, not manipulate the evidence to fit a solution.
AstraMike is offline  
Old 7th Jul 2009, 13:50
  #3177 (permalink)  
Join Date: Jun 2009
Location: Colorado
Posts: 41
Likes: 0
Received 0 Likes on 0 Posts

As for the likely sequence from 02:10 to the surface, several times others have posts which is essentially the camp I'm in. (control difficulty starting around 02:10, upset at some time, essentially stall/spin, flight ending at 02:14 or soon after, not much distance to be traveled after 02:10, fin-off at impact.) Not that it makes me credible but I don't want to repeat their work, I'll let their posts do the talking.

but this part
2. The bodies drifted 100 NM (to the North) in five days after the first recovery but would have drifted previously 10 NM during the same time, and in the wrong direction (to the South-West), in an area where the drift is much stronger?
With all the official, credible and yet conflicting sources of current and drift information, I'm not so certain myself what the drift at impact site was at the time of crash. One of these official sources, posted here a couple times by SaturnV using this OSCAR - Ocean Surface Currents Analyses - Realtime, causes trouble for your drift back-track projections. While it agrees with all the northish drifts at the middle and northern side of the 0210-0214 area, on the south side of this it shows an eastward flow. The specific place it begins, direction and speed on the date/time of the crash could very well been to the east for all we know. I can only argue for 'uncertainty' and some question of the back-track projections. I know you're aware of wind & weather conditions at the specific time of the crash that could potentially throw a wrench into the projections.

Last edited by ttcse; 7th Jul 2009 at 14:22.
ttcse is offline  
Old 7th Jul 2009, 13:52
  #3178 (permalink)  
Join Date: Jun 2009
Location: SPAIN
Age: 64
Posts: 50
Likes: 0
Received 0 Likes on 0 Posts

Do remember, a long time ago I asked whether there might have been other ACARS messages cued after 0214. Nobody saw fit to answer that. Can you?
I am open to answers. I'm not open to screaming fits.

As I commented when it was first time posted .. it's a very interesting point ... and from all the noise here I wait also some answer to this question
If I speculate ... the plane after 0214 was completely out of control (no more contact between antenna and satellite due to aircraft movements) or .. the plane was no more in one piece ......

LeandroSecundo is offline  
Old 7th Jul 2009, 14:00
  #3179 (permalink)  
Join Date: Jun 2002
Location: Orlando, FL, USA
Posts: 108
Likes: 0
Received 0 Likes on 0 Posts
Dougyboy, re your #3177

Most likely too late by now or already done but just in case: Your friend Carole might be able to use an additional source of information. I believe NWA (US) is the world's largest operator of the A-330.

She should contact ALPA-US

ATTN: Capt. Rory Kay
Executive Air Safety Chairman
email: [email protected]
US Tel: 703-689-4200

I'm certain Capt. Kay would be happy to assist in obataing that information or any other that might be helpful from US operators of ABI aircraft.

surplus1 is offline  
Old 7th Jul 2009, 14:07
  #3180 (permalink)  
Join Date: Dec 2003
Location: UK
Posts: 24
Likes: 0
Received 0 Likes on 0 Posts
I thought it was well established that it took a while, days in fact, depending on surrounding temperature, before a human body would float. Equally, the conditions below the surface are different than those on the surface.

Therefore you cannot base constant drift assumptions on the bodies, until they surface.

Strangely, it seems to me that the remains are largely co-located with other debris, along a drift trail, although I stand to be corrected. If I am right - this has to be wrong, unless the debris started to surface and float at about the same time that the bodies did, give ot take a bit for wind effect.

Either way, you surely can't extend the drift line back assuming a constant drift rate? And then, what was the sea state at the time of the accident and what account has been taken of this?

There has to be an answer that precludes a 180 turn in inclaement weather where a stall or overspeed are real threats as I doubt very much that this would have been achievable and once in the throws of disaster the winds at the time, let alone wind shear, would have taken control of the direction of the aircraft, likely moving it off track.


Last edited by AstraMike; 7th Jul 2009 at 14:12. Reason: spelling edit
AstraMike is offline  

Thread Tools
Search this Thread

Contact Us - Archive - Advertising - Cookie Policy - Privacy Statement - Terms of Service - Do Not Sell or Share My Personal Information

Copyright © 2023 MH Sub I, LLC dba Internet Brands. All rights reserved. Use of this site indicates your consent to the Terms of Use.