OK, so this is largely Google Translate, with a bit of what I've picked up from here, but this seems to be the meat of it:
Translated version of cvr.doc
2 - HISTORY OF FLIGHT: ANALYSIS POINTS
The flight was divided into three phases:
Phase 1: the beginning of the CVR recording to the autopilot disconnect.
Phase 2: disconnecting the autopilot until the outbreak of the stall alarm.
Phase 3: the onset of stall warning until the end of the flight.
• Phase 1
At the beginning of the CVR recording shortly after midnight, the aircraft was cruising at flight level 350. 2 The autopilot and auto-thrust incurred. Automatic transfer of fuel in the "trim tank" was performed during the climb, and the center of gravity is then 27.5%, to a mass of 218 tonnes. The flight is calm. The crew, consisting of the captain and co-pilot, VHF is in contact with the control center of Recife.
The crew discusses the high temperature (standard plus eleven) and reported the estimated amount of fuel remaining on arrival, which is evolving. The crew sees the city of Natal and finds that the weather has not been a problem. He is concerned about the closure of a land support ETOPS, Sal-Cape Verde Amilcar. It requires a change to the CCO, who said that the field is open in an emergency.
The captain offers the first officer to take a rest because of the length of his vacation. The latter replied that he did not want to sleep.
35 to 1 pm, the plane gets to the point INTOL and crew leave the frequency of Recife to switch HF communication with the control center Oceanic Atlantico. A trial is SELCAL effectué5 successfully, but attempts to connect with ADS ocean Dakar fail.
Soon after, the first officer change the scale of the ND 320 NM 160 NM and found that "there is something ahead," which presumably refers to echoes detected by weather radar. The captain and the crew discusses confirms again that the high temperature does not allow them to ride at 370.
A 1: 45 pm, the aircraft enters a turbulent zone slightly, just before the point SALPU.
Note: the crew had received a 30 to 0 h information from the OCC on the presence of a convective zone associated with the ITCZ between SALPU and TASIL.
The crew reduced the lighting in flight deck and turns on the headlights "to see outside." The first officer noted that they will "return to the layer" and it would have been nice to ride. A few minutes later, the turbulence slightly and reinforce the co-pilot suggested asking up in 360 non-standard as it thinks is "really on the edge" of the layer. The captain says they will wait. It reduces the scale of the ND 40 NM, the weather radar will go into mode + weather turbulence. Soon after, he reported the appearance of St. Elmo's fire and said that "rotor as it goes" when he goes to rest.
Shortly after 1 h 52, the turbulence ceases. The first officer again draws the attention of the captain on the value MAX REC, which then reaches the FL 375. The captain did not comment and a few moments later, he wakes up in the second co-pilot, said he will take his place, and asks the copilot in the right seat if it has a pilot license online. Thus it ensures that it is entitled to the supplement and refers implicitly as reserve driver. This question probably means that the first officer issues the driver alternate the captain had not been addressed during the briefing before the flight.
On his return to the cockpit, the second co-pilot said he dozed. He sits on the left seat and the copilot in the right seat makes it a briefing, pointing out that "little bit of turbulence that you come to see you go ... we should find the same before it is actually in the layer Unfortunately we can not get over for now because the temperature decreases more slowly than expected. " Before leaving the cockpit, the captain recalled the HF frequencies to contact Atlantico and Dakar ocean. In fact, after the departure of captain, first officer on the left is PNF and co-pilot on the right is PF.
The two co-pilots are still discussing the temperature and REC MAX. The turbulence increases slightly and they decide to prevent cabin crew from entering an area more turbulent. The first officer on the right says they are "apparently in the boundary layer," before adding that he would have preferred to climb to FL 360. Go to the next level is a constant concern of the crew. The drivers clearly want change outside of the layer, presumably to minimize turbulence.
08 to 2 pm, the first officer on the left, probably after seeing the echoes detected by radar weather, proposes to "alter a little to the left." The HDG mode is activated and select the course dropped 12 degrees to the road. The first officer left pass the gain setting up the weather radar, after observing that it was calibrated mode. The discussion is interrupted by a sensation of temperature increase and the appearance of an odor about which drivers exchange for more than a minute. The first officer on the left identifies this as the smell of ozone.
Conversations in the cockpit did not reveal any malfunction of the weather radar and show that it has submitted a usable image.
The background noise changes rapidly to 2 h 09 min 46. This change in background noise has been identified as being characteristic of the presence of ice crystals but does not give rise to any particular crew, the phenomenon was little known drivers at the time. The PNF then takes the initiative to reduce the Mach 0.8 and anti-icing engines are engaged.
The departure of captain has been without a clear operational guidelines, particularly on the role of each of the two co-pilots. The absence of a formal framework of operation of a crew consisting of two co-pilots have led to the division of labor is not optimal seen them.
• Phase 2
A 2 h 10 min 05, the sharp drop in measured velocities, presumably due to blockage of the Pitot probes by ice crystals, causing the disengagement of the autopilot and auto-thrust (thrust is then fixed) and the change of control law of flying to alternate normal. The presence of turbulence, as evidenced by the activity of the PA to control body roll in the seconds before, led to a disconnection from the plane to roll right up to about 8 °.
The co-pilot PF says "I have control" and has fast action and high amplitude side, almost lock to lock. He was also an action that increases the pitch attitude of the aircraft to 11 ° in ten seconds. Flight directors are not disengaged by the crew, but the trend bars disappear.
A 2 h 10 min 10, the PF are increasing the incidence and the stall warning activated twice unexpectedly. Probably in response to this alarm, the PNF exclaims, "What's that?". The PF then said "there was not a good one ... you do not speed ... good news" and the PNF "we lost the speed." The reported incidence is about 5 °, a value theoretical threshold for triggering the alarm just over 4 °.
The crew identifies the loss anemometer indications but none of the two co-pilots then calls the associated procedure. The emergency maneuver "IAS dubious" requires first to disconnect and disengage the automatic flight directors. The two co-pilots had been trained at the emergency maneuver in low layer, at which the plate to adopt is 10 ° or 15 °.
However, a note describing the problem OSV loss anemometer indications previously encountered on the cruise and A330/A340 fleet reminiscent of the procedures to be applied. The note had been circulated to all NTP division A330/A340.
Between 2 and 18 h 10 min 2 h 10 min 25, the PNF read the ECAM messages in a disorderly manner, but mention the loss of self-driven and alternate passage into law. Thrust lock function is disabled. PNF calls and triggers anti-icing of the wings.
The PNF then attracts the attention of the PF on speed. At that time, the two speeds recorded (the one displayed on the PFD of the left and ISIS) are less than 100 knots and vertical speed reaches a maximum of 7000 ft / min. Longitudinal movements of the aircraft are the result of the actions of PF, which also continues to perform actions of high amplitude side to control the roll, less than 10 ° right and left.
A reading of the three instruments (both PFD and ISIS), the SOP notes that the aircraft climbs and asked repeatedly to go back down to the PF. The latter then makes a number of actions that have to bite the effect of reducing the tax base and the vertical velocity, whose values ??are nevertheless still excessive and the aircraft is about 37,000 ft and continues to rise, without the intervention of PNF. Although the REC MAX was a constant concern before disconnecting the PA, none of the two co-pilots there refers.
To 2 h 10 min 34, the speed displayed on the left side becomes valid and then to 215 kt, the speed of the ISIS is always wrong. The aircraft then lost about 60 kt from disconnecting the autopilot and the start of the climb, which is consistent with increasing altitude of about 2000 ft.
A 2 h 10 min 47, the thrust levers are slightly remote to 2 / 3 of the beach IDLE / CLB (85% N1). Two seconds later, the plate came back a little below 6 degrees, roll is controlled, the impact is slightly less than 5 ° and the PHR is 3 ° nose. The vertical velocity is always high at a 100 ft / min. From a few seconds, the PNF tries to recall the captain.
In this phase, at any time one of the two co-pilots has announced value of speed, attitude, vertical speed and altitude. The flight directors were not disengaged, the bars tend to have disappeared and reappeared several times it is not possible at this stage of the investigation to know what orders they were able to identify or determine whether these orders may have influenced the actions of the PF.
At this point, after a quick trim and altitude, which is the consequence of the actions of PF, the trajectory of the aircraft appears to be controlled. The increased pitch angle and the initial vertical velocity resulting were excessive for the altitude and should have been advertised immediate difference from the PNF. The lack of specific training manual flying at high altitudes likely contributed to the actions of steering and monitoring inadequate.
The low synergy observed between the two co-pilots can originate on one hand the lack of clear division of roles by the captain, and also the lack of CRM training between two co-pilots in a situation substitute the captain.
Note: no regulation requiring such training or criteria such as experience or ability in the decision to designate an alternate captain during flights by crews strengthened.
• Phase 3
A 2 h 10 min 51, the stall alarm is activated again, with an incidence of approximately 6 °, which corresponds to the theoretical threshold for triggering the alarm for the Mach 0.68 worth while. PF continues to hold considerable nose-up input: the plate increases by 6 to 13 and the incidence of 6 ° to 10 °. The vertical acceleration recorded reflects the rapid onset of vibration can be the buffet. Five seconds later, probably in response to the stall warning, the PF advance thrust levers to the notch TO / GA and the announcement. That's about at that moment that the plane out of its flight.
Despite some actions to sting, the PF maintains a global action to pitch up. The plate is between 11 ° and 18 ° and the incidence between 11 ° and 23 °. The THS begins a movement consistent with the actions of PF and will reach the value of 13 ° nose about a minute later. It should be recalled that in alternate law, auto trim is still active. By cons, it is difficult for the crew to know the position of the trim and no alarm does not notify the crew that place.
A 2 h 11 min 06, after several attempts to call, the PNF is concerned about the lack of new captain. This concern probably increases the stress of the PNF in front of a situation he does not understand.
A second later, the speed of the ISIS becomes valid. ADR 3 is selected on the right side PFD, the speed of the PF becomes equally valid. It is 183 kt and then the three speeds shown are consistent. This requires no comment from the crew.
Within 30 seconds after the start of the stall warning, the airspeed decreased from 205 to about 160 kt. The vertical velocity is zero and then gradually became strongly negative (4000 ft / min). The maximum altitude of about 38,000 ft was reached at 2 h 11 min 10.
Shortly after 2 h 11 min 30, the OP said twice that he lost control of the aircraft. This may reflect the difficulty of controlling a roll, the airplane banked to the right when he maintains his grip on the left stop. However, his few actions to sting every time cause a reduction of assessment.
A 2 h 11 min 37, the PNF called "controls on the left," takes precedence and gives a short-acting to a stop on the left, the PF takes priority almost immediately without any announcement and continue to drive.
To 2 h 11 min 42, the captain enters the cockpit shortly before the stall warning stops. The parameters of the plane are then: altitude of about 35,800 ft, vertical speed of 9100 ft / min, calibrated airspeed of 100 kt down, base 12 ° and N1 engine 102%. Neither of the two co-pilots will not result in a detailed account of problems or actions taken, except that they have lost control of the aircraft and they have tried everything. In response, the CDB said many times "take it", probably referring to the FPV. The parameters show that the stall warning stopped because the three values ??of incidence have become invalid.
Note: from 2 h 11 min 45, the speeds are no longer displayed continuously on the PFD.
A 2 h 12 min 04, the OP said he thinks they are overspeed condition, perhaps because of high aerodynamic noise prevailing in the cockpit. None of the other two drivers parse this hypothesis when it is inconsistent with the pitch attitude and vertical speed high descent.
Until the end of the flight, the values ??revert incidence successively valid and invalid. Every time that at least one value becomes valid, the stall alarm is reactivated, and each time that the effects are disabled, the alarm stops. Several actions to sting causes a decrease in the impact assessment and whose values ??revert while valid, so that a free action to sting results in the activation of stall warning. It seems that the drivers then react with a nose-up action, the consequences are an increased incidence, a decrease in measured velocities and therefore stopping the stall alarm. Until the end of the flight, no value will be valid incidence below 35 °.
Neither of the two co-pilots has formally identified the dropout situation in which was the plane or through the alarm either by recognition of the buffet, or by the interpretation of high values ??of vertical velocity and the base. It should be noted that the buffet is the only indication of approaching the stall at high altitude on other airplanes of the alarm threshold stall does not vary with the Mach.
In the absence of relevant information from the co-pilots, the only reading of information available on the screens (plate, roll, push, vertical speed, altitude, etc ...) did not allow the CBD to quickly realize the situation of the aircraft. He did not then ask questions that might help them understand the sequence of events.
The stall warning lasted 54 seconds continuously, during which none of the co-pilots are referred to. It is likely that the CBD has heard the alarm a few moments before entering the office, but it is also likely that its many stops and reactivations were added to the confusion and disrupted his diagnosis of the situation.
Despite several references to the altitude, which decreased, none of the three crew members did not seem able to determine what information to trust: the values of trim, roll and thrust might have seemed inconsistent with the values vertical speed and altitude.
3 - CONCLUSIONS
3.1 New Findings
the composition of the crew was in accordance with the procedures of the operator,
the weight and center of gravity were within operational limits,
at the time of autopilot disconnect, the captain was at rest,
the departure of captain has been without a clear operational guidelines, particularly on the role of each co-pilots,
the crew had identified the echoes on radar weather,
the crew made ??a change in direction of 12 ° to the left of the road
The PA has withdrawn while the aircraft was operating in an upper cloud layer of slightly turbulent
co-pilots had not received training at high altitude, the procedure "IAS questionable" and flight manual
there was inconsistency between the measured velocities, presumably as a result of obstruction of Pitot probes in environment of ice crystals,
although identified and announced the loss of airspeed indications, none of the two co-pilots has called the procedure "IAS questionable"
the invalidity of the speed displayed on the PFD lasted 29 seconds left, the speed of the ISIS 54 seconds,
in less than a minute after the disengagement of the autopilot, the plane went out of its flight as a result of manual control actions predominantly to pitch,
The captain returned to the cockpit about 30 minutes after a disengagement of the autopilot,
throughout the flight, movements of the elevators and PHR were consistent with the actions of the pilot
to the exit of the flight, the longitudinal movements of the aircraft were consistent with the position of the control,
there was no explicit allocation of tasks between the two co-pilots,
there is no CRM training for crew consists of two co-pilots in a position of substitution of the captain,
no announcement regarding the standard deviations of attitude and vertical speed has been made,
the impact of the plane is not directly presented to pilots,
approaching the stall was characterized by the activation of the alarm and the onset of buffet,
neither pilot refers to the stall warning,
neither pilot formally identified the situation stall
the stall warning was triggered continuously for 54 seconds,
shortly after activation of stall warning, the PF applied thrust TO / GA and had a nose-up action,
the incidence is the parameter to activate the stall warning and if the values of angle-of-attack become disabled, the alarm stops, by design, when the speed values were measured below 60 kt, the three values of incidence/AoA have become invalid
whenever the stall warning was activated, the incidence exceeded the value of the theoretical threshold of activation,
engines worked and always responded to the commands of the crew,
no announcement has been made to passengers.
4 - SAFETY RECOMMENDATIONS
Reminder: in accordance with Article 17.3 of Regulation No 996/2010 of the European Parliament and the Council of 20 October 2010 on the investigation and prevention of accidents and incidents in civil aviation, a safety recommendation does by no means a presumption of fault or liability in an accident, serious incident or incident. The recipients of safety recommendations report to the authority responsible for security investigations that issued the actions taken or planned to ensure their implementation, as provided by Article 18 of the Regulation.
4.1 Recommendations on operation
Flight Training Manual
The survey has highlighted the weaknesses of the two co-drivers: the inappropriate actions of PF on the flight controls at high altitude have not been identified by the PNF by lack of effective monitoring of the trajectory. The stall warning buffet and have not been identified. This is probably due to a lack of specific training despite compliance with regulatory programs. Manual control can not be improvised and requires precision and measured actions on the controls. There are other opportunities for disengagement of the autopilot for which only a specific and regular training can provide the necessary expertise to ensure flight safety. A reading of their latest training and checking, it was shown that co-pilots were not trained in manual control, approach and recovery stall at high altitude.
Consequently, the BEA recommends:
• that the EFSA review the content of training programs and to impose such control and the introduction of specific exercises and regular dedicated to manual control, the approach and stall recovery, including high altitude.
Replacement of the command
The survey revealed that lack of education and training a crew of two co-pilots do not guarantee a performance level equivalent to that of a crew of one captain and a co-pilot in a situation worse. The lack of hierarchy and efficient allocation of tasks in the cockpit have greatly contributed to the low synergy. The concern arose from the absence of captain in the cockpit shows that the two co-pilots lacked the capacity to address this emergency. This can be explained both by the lack of adequate training and lack of decision-making practice of the two co-pilots. Many events have had a favorable outcome due to the presence of the captain whose training and experience have enabled a more robust and more serene reactions to the situation.
Consequently, the BEA recommends:
• that the EFSA defines additional criteria for access to the function
Deputy of the captain to ensure a better distribution of
work crews in case of reinforced
and
• that, temporarily, the DGAC defines additional criteria for
access to the function of deputy commander to ensure
better division of labor in cases of crews strengthened.
4.2. Recommendation for certification
Measuring impact
The crew was never formally identified the dropout situation. The angle-of-attack information is not directly accessible to the pilots. The angle-of-attack of cruise flight is close to the angle-of-attack of the onset of stall warning in a law other than the normal. Under these conditions, manual control can cause the aircraft to high incidences such as those encountered during the event. It is essential to ensure the safety of the flight to reduce the angle-of-attack when the stall is imminent. Only a direct reading of the impact could allow crews to quickly identify the location of the aircraft aerodynamic and have any necessary actions.
Consequently, the BEA recommends:
• that the EASA and the FAA evaluate the merits of integrating the presence of an impact indicator directly accessible by the pilots on board aircraft.
4.3. Recommendations for flight recorders
The operating parameters of the FDR and CVR listening to provide information essential to understanding the event. However it is difficult to reconstruct the information available to the crew on their flight instruments, including the orders given by the bars in the trend of flight directors when they reappear. We can not see if there have been attempts to re-engagement of the autopilot. A view of the dashboard complement the information given by the FDR and CVR and would confirm the information available to the crew and the actions it has taken. Many recommendations have already been issued on the subject over the last ten years without a breakthrough has occurred.
Consequently, the BEA recommends that again:
• that ICAO imposes on aircraft performing public transport passengers an image recording for viewing the entire panel,
and
That, at the same time, ICAO establish very strict rules of operation of such records to ensure confidentiality of stored data.
Today, regulations require the registration of conduct of the flight parameters displayed on the left. Some key parameters for the analysis of flight control are lacking, particularly those presented in the right seat driver: speed, altitude, attitude, position bars tend flight director, etc.. In addition, aircraft are equipped with complex systems whose analysis of the operation is limited and hampered by the lack of registration of the parameters of all the data sources they use.
Consequently, the BEA recommends:
• that the EASA and the FAA mandate the recording:
o the position of the bars tend flight directors,
o the conduct of the flight parameters displayed in the right seat, in addition to those shown on the left,
and
• that the EASA and the FAA are studying the merits of mandatory recording anemometer and inertial parameters of all sources used by the systems.
4.4 Recommendations on the transmission of flight data
In the Progress Report No. 2, BEA has issued safety recommendations regarding the increase in the duration and scope of the ULB, sending regular data recorders and carriage of ejection. These recommendations are based on the findings of an international working group led by government-industrie12 the BEA as part of the safety investigation into the crash of flight AF-447, which has since studied the feasibility of transmission triggered flight data. The concept is to analyze real-time flight data on aircraft to detect emergency situations. In these cases, the transmission of flight data is triggered to facilitate the location of an aircraft in emergency situations. The working group's results show that it is technically feasible to define reliable criteria based on flight parameters to detect emergency situations, while limiting the false detection. The group also concluded that it is technically feasible to obtain an impact position.