Jun 3, 2009

AF447 mid-air breakup evidence raises new discussions about the last signals sent to Paris

There is what could be described as technically informed speculation amongst airline professionals in Australia and abroad that AF447 broke apart in flight at or near 35,000 feet and fe

Ben Sandilands — Editor of Plane Talking

Ben Sandilands

Editor of Plane Talking

There is what could be described as technically informed speculation amongst airline professionals in Australia and abroad that AF447 broke apart in flight at or near 35,000 feet and fell into the mid Atlantic in two main parts.

The primary evidence supporting this view is the congregation of the floating debris found in the last day into two areas about 60 kilometres apart.

However there are other supporting clues too.

The A330-200 was known to have flown through severe turbulence for around 30 minutes by some estimations before a series of automatically generated electrical fault reports were transmitted to the airline’s Paris base during the period before and perhaps up to main impact with the sea.

These service alerts begin with what has only so far been described as an electrical short circuit notification.

Followed by cabin depressurisation, after which a ‘flurry’ of electrical systems warnings described as ‘unprecedented’ by Air France are received.

Depressurisation of the jet would be a certain consequence of a mid-air breakup.

This is also where the speculation about such a break-up itself divides into two streams.

One takes the view that the electrical warnings that followed had most likely little or nothing to do with the causes of the break-up and were mainly consequential.

The alternative view is that these warnings precede any break-up, and are evidence of a severe control crisis that may have exceeded the design limits of the aircraft, or led for some reason to a failure in a structural component which so compromised the jet that it broke into at least two major parts.

It has been reported by The Aviation Herald, an online Europe-based journal of aviation incidents and news that the main body of electronic alerts begin with the disengagement of the autopilot and were followed by messages related to the ADIRU or air data and inertial reference units and the PRIM or flight control primary computer which is informed about speed, attitude and other material flight values by the ADIRUs.

Superficially this resembles the onset of the mid air upset that caused Qantas flight QF72 from Singapore to Perth to make an emergency landing in Learmonth last October.

The PRIM will in some flight modes intervene in the flight controls settings of the jet to inhibit pilot inputs which would exceed critical limits which could stall the airliner, or overload parts of the structure or control surfaces on the wings or rudder.

However these limitations can also be in turn locked out by the pilot.

Other reports indicate that these ‘unprecedented’ messages were concentrated in a four minute period, ending with a final advisory message about the vertical speed, that is, the rate at which the jet was falling rather than any speed with which it was also moving forwards.

Faulty ADIRU units in the Qantas A330-300 operating the flight that diverted to Learmonth remain a major focus of an unfinished air accident investigation by the ATSB. That investigation is also looking at other ADIRU related incidents on Qantas A330s.

However the Qantas ADIRU units were made by Northrop Grumman, while those in the Air France jet were made by Honeywell, They are two completely different designs, running totally different sets of programmed logic to serve the same ends.

All of which makes finding and recovering readable flight data and cockpit voice recorders from the crash zone so critical to solving the riddle of AF447 and the deaths of all 228 people who were onboard the flight from Rio de Janeiro to Paris.

(Visited 2 times, 1 visits today)


Leave a comment

19 thoughts on “AF447 mid-air breakup evidence raises new discussions about the last signals sent to Paris

  1. Amos Keeto

    Ben, this is probably a silly question but why is it that the “black boxes” only have power to ping for 30 days?. Seems to me that increasing the capacity to 60-90-356 days just wouldn’t be that hard.

  2. Ben Sandilands

    A number of people are pushing for unlimited capacity to reflect the transfer and storage capabilities available to internet users today. Their suggestion is for flights to download such data in real time, or in a series of back-ups just as current designs, including the AF Airbus was doing in its automated service messages to base.

    That way there would never be any ‘black boxes’ to recover and read, as all of the data would exist, in theory indefinitely, in a master archive of flight recordings at the airline or lessor’s operational base.

    The answers to what happened to AF447 would be yesterday’s news.

  3. dan1

    Thats all well and good, but there may come a point that what ever info available doesn’t actually get the chance to be transmitted to base

    I ask again, why is it that BB only ping for 30days. Theoretically we could make them ping for years

  4. dan1

    dan1 = Amos Keeto

    Different computer, srry

  5. Ben Sandilands


    I don’t know. However the standards for black boxes were set by an ICAO committee years ago sometime after Australian inventor David Warren’s concept was already becoming widely accepted, with Australia being one of the last countries to accept his idea because pilots here resisted the innovation fiercely and our airlines sided with them because of the cost. (A long and embarrassing story about how change resistant aviation is in this country, but I digress.) Those international standards, once fixed, stayed fixed while everything else, including battery technology kept moving forwards. It’s odds on that the standards will be reviewed, but I’d also put money on real time data archiving as the ultimate level of data management.

  6. grappaguy

    Hi Ben:

    How did you confirm that it was the ADIRS from Honeywell and not the LTN-101 from Northrop Grumman on AF447? The Honeywell brochure for their ADIRS excludes A330 aircraft, Northrop Grumman’s brochure for the LTN-101 includes A330’s. Also, there are two existing Airworthness Directives, one issued by the FAA in 2008 and more recently AD#: 2009-0012-E issued by EASA on January 15th, 2009. Both of these are related to Northrop Grumman’s ADIRU’s, so this issue has been around for over a year. Would you agee that if an ADIRU malfunctioned combined with turbulant weather could result in an uncontroloble dive? A question to ponder is, were the pilots aware of and did they follow the required actions stipulated in the EASA directive. The FDR would cleary indicate this one would think? God bless those that perished and the friends and families that are grieving.

  7. Ben Sandilands

    I rang people in Australia and France. I’m in the process of checking for an AD said in another place to apply to the ADIRU units on the AF jet and they are made by Honeywell. There is a string of successive ADs on the Northrop Grumman (formerly Litton) units, only the last one strictly relevant of course. Your question about what the pilots did is very relevant, and we don’t yet have an answer. I should make it clear I have no opinions about complex technical flying issues. As a reporter my role is entirely that of keeping in touch with a range of trusted sources of information whose opinions are informed by their experience and thus worth knowing. I particularly value those who will tell me they don’t know an answer and then come back with a lead. It is a long and tedious process sometimes, especially those who are not instantly contactible.
    I do think it is clear from the ATSB preliminary report on the QF72 incident last October that in the scenario you outlined is remotely possible. There have been a number of ADIRU malfunction incidents in other airliners, some referred to in the ATSB report. But they seem to fall within the general background stream of systems failures day in and day out the pilots have to cope with.

    It is the essential proposition that springs up all the time. Professional airline piloting is not about routine flying of airliners. It is about saving a jet when the dragons appear without warning.

  8. Alison Bailey

    So, if the black box emits a signal, why is it so hard to find the aircraft? Is it because the signal is not strong enough to transmit through several thousand metres of water?

    Why do planes not transmit GPS data continuously to air traffic control, so that their exact location can be pinpointed at any time?

  9. Ben Sandilands


    Continuous transmission of GPS positions are integral to the next generation of air traffic control technology, which is a total redesign and completely different design philosophy compared to current systems. A resource for more information on this is at http://www.faa.gov/about/initiatives/nextgen/

    Current ATC is based on radar returns, or reporting and plotting of positions to track, or dead reckoning in lay terms.

    The GPS dependant system uses live GPS location reports to constantly track everything, which also changes everything. Australia has mandated introduction of the system from as early as 2013 for some aircraft. We may be ahead of ourselves given the formidable resistance to these changes around the world.

    There are several enormous problems with radio signals under water, starting with the need for very long wavelengths, which essentially means large antennas. At more than 3 kilometres down, a crystal clear transmission would involve something the size of a jet, and an order of magnitude or more greater in power, which is also why time is short in picking up the signals that should be emitted now.

    Finally, and alas, dropping the black boxes into the depths where there is a complex submarine mountain range and canyons is like dropping them into a rain forest. They can get buried in sediment, slip down gullies and just die signal unreceived. The emphasis in finding the signals will be divided between powerful surface receivers and others towed by submarine devices.

  10. Global Voices Online » Air Bus AF 447: Sorrow, lack of information and sensationalism

    […] make an emergency landing. Others argue this incident is not relevant to the recent disaster. From Plane Talking, written by Ben Sandilands at Crikey Blogs in Australia. It has been reported by The Aviation […]

  11. JOL

    Usually the signal sent is aucoustic, not radio

  12. Global Voices po polsku » Airbus AF 447: smutek, brak informacji i poszukiwanie sensacji

    […] awaryjnego. Inni twierdzą że tamtem wypadek jest nieistotny w obecnej katastrofie. To ze tekstu Plane Talking, napisane przez Bena Sandilands na Crikey Blogi w Australii [en]. The Aviation Herald, europejski […]

  13. Air7

    Is it possible that the four minute flurry of transmissions reporting electrical faults, was the time taken for the fragmented aircraft to fall to earth, and the fault reporting systems where still functioning during the fall.

  14. Air7

    oh and one more question, why is there no floatation device fitted to black boxes, I suppose a retrofit would prove costly, what percentage of mid air accidents occur over the ocean?

  15. Ben Sandilands

    Air 7,

    Doubts are arising as to whether the four minute transmission interval was necessarily as short as 4 minutes because ACARS depending on how it is used can store some messages for a period and even sent them out of sequence depending on the priority the airline assigns to different categories of maintenance messages. ACARS set ups vary between carriers. Debate about this continues to rage.

    I think based on multiple references to the injuries sustained by the victims that the break up was very sudden and clearly traumatic although following a period of some minutes in which the flight was in crisis. Some of the items photographed on the decks of the recovery ships in the main area are from both the front galley area and the rear of the jet, implying little dispersion prior to impact which is contrary to what would have been expected if the breakup was spread over a number of minutes at varying altitudes.

    On the other hand, this investigation may well take surprising turns. I wouldn’t rule anything out apart from a bomb blast or fire on board, as there are none of the signs of this on the victims nor on the cabin fittings.

    Flotation for data recorders has been suggested over the years. The main issue appears to be the need to have some sort of pressure charge to open a gap through which they can escape. It is quite a dilemma. They need to be located where they won’t be crushed, which is why they are in the tail area, but they also need to be recoverable.

    It now seems inevitable that the next step in data recorders will in fact be to replicate their function, and for longer periods, by parallel real time uploads of critical parameters to an archive.

    This would have meant that the moment this crash was suspected the airline would have known in sequence and in real time most of the data related to critical functions, including exact location, orientation (was it spinning, climbing, falling, G forces, and for how long etc) as well as the performance of the various systems. The bandwidth and satellites to do this aren’t there yet, they but will be in the future.

  16. Paul S.

    I have been reading all the posts on this thread, and I have some comments that may throw some light on certain matters:

    1…Radio waves propagate VERY badly in water…almost to the point of not propagating. The US Navy submarines use underwater radio at CARRIER frequencies of about 10 Khz and wavelengths of about 30,000 meters (30 Km). This requires enormous towed antennas and huge power at shallow depths…o/k in a nuclear powered submarine. Bandwidth is very limited.
    (see: http://en.wikipedia.org/wiki/Very_low_frequency0

    2…One of the most useful crash forensics tools is the Cockpit Voice Recorder (CVR). These were strenuously rejected by the pilots unions at first because they expected that management would listen in to cockpit conversations and penalise pilots for what they heard. It was finally accepted on the condition that it would only record a limited time loop, and would have an erase button. The concept that Fifi will continuously transmit all cockpit conversations to Mother in real time for recording and archiving will probably cause pilots to refuse to fly aircraft with such a system.

    3…The concept of a floating recorder has neen kicked around for years. At first glance, it could be on the upper skin and released by immersion in water, but that would lead to problems if it rains. Airlines are very reluctant to use any device with explosives in it, and numerous suggestions have been put forward for a sort of torpedo tube device. The fate of AF447 may spur some action.

    4…Pingers that ping for a mere 30 days. A simple improvement would be to use a real time clock that causes the pinger to ping about 20 times at very high power on the hour rather than a single low power ping every second. This would enable searchers to get quiet and listen carefully an the hour, and would permit a 22.5 dB increase in ping-power or a nearly 14-fold increase in range without any increase in battery capacity.

    5…In order for the ACARS to transmit via SATCOM, it has to have power, so we know that power was still available. Also, the ADIRU’s must be healthy because the SATCOM has to have aircraft attitude etc (Pitch, roll, heading, lat, long.) in order to point its antenna at the satellite. That indicates that the aircraft was moderately intact during that time. Also, in a vertical dive it is unlikely to be able to point the antenna at the satellite. The antenna is an electronically steered conformal device that wraps around the upper curvature of the fuselage.

    6…Full time uploading of flight data will be difficult due to the bandwidth required. The last few seconds of data before a crash contains the most useful data, and due to the fact that a full time upload would be delayed, there would have to be an onboard recorder that would have to be found and recovered.

    QUESTION: Does anyone khow if large transports like A330 are spun durong flight testing, and do pilots manuals contain spin recovery techniques?? The standard P.A.R.E. may not be best.

  17. Media Channel 2.0 — Blog — Air Bus AF 447: Sorrow, lack of information and sensationalism

    […] make an emergency landing. Others argue this incident is not relevant to the recent disaster. From Plane Talking, written by Ben Sandilands at Crikey Blogs in Australia. It has been reported by The Aviation […]

  18. AF447-world wide ‘pitot’ speed probe replacement to be mandated for all Airbus A330s and A340s – Plane Talking

    […] a Plane Talking reader, Paul S, posted a detailed comment on an earlier item about the crash which includes this simple but very practicable suggestion as to how to make it […]

  19. Airbus AF 447: Tristeza, falta de informação e sensacionalismo · Global Voices em Português

    […] Outros argumentam [En] que esse incidente não é relevante para o recente desastre. De Plane Talking [En], escrito por Ben Sandilands no Crikey Blogs da Austrália: It has been reported by The […]

Share this article with a friend

Just fill out the fields below and we'll send your friend a link to this article along with a message from you.

Your details

Your friend's details