Fdr/cvr data recovery
 

@ Noske
Thanks for that.

Interesting that the BEA had problems with memory cards and Honeywell found shorts on the board. Sounds like it could have been shock vibration damage to the die connection wires rather than water ingress.

The recovered memory unit housing we see from yesterday's recovery looks in remarkably good shape to me - the indent does not seem to extend beyond the weld seam, and although there is a bit of rust, the seam appears to be intact. If the seal is compromised, then the integrated memory circuit packages will eventually get exposed to 400atmos pressure, and it's possible that the void inside the chip package will collapse onto the micro circuit die. In the past, Aerospace or Mil parts have used ceramic packaging for integrated circuits, with plastic, which is a lot cheaper to fabricate, the norm for commercial applications. There has been some use of plastic in mil aerospace apps in recent years, but one of the problems with plastic packages is that they are porous and allow water to penetrate the void. So, for either package, the void could collapse and crack the die. Or, on the other hand, a plastic package could be subject to a slow increase in pressure gradient (leak through a micro crack in the cylindrical metal containment housing), and it could fill with water and expose the die to corrosion. There are a few manufacturers around who will put die into ceramic packages, but not that many, and usually the cost is hugely expensive compared with plastic. If there is any moisture present around the die, then any application of voltage will cause an electrolyses effect and irreversible damage. So, If there is any suspicion that water has penetrated, the most like course of action is for a trip back to the manufacturer for the top of the package to be removed under controlled conditions and a session of drying and cleaning of corrosive deposits in the hope that maybe something useful can be recovered. Of course, It's quite possible that the CVR/FDR manufacturer has either obtained a special package for the FLASH ram, or has encapsulated the devices in some form of epoxy resin to cover the contingency of water ingress.

It's also interesting that the BEA allowed Honeywell to dig into the electronics. In section 2.4.1 of the EUROCAE document: MINIMUM OPERATIONAL PERFORMANCE SPECIFICATION FOR CRASH PROTECTED AIRBORNE RECORDER SYSTEMS, it states:

NOTE 1: Microscopic examination of the surface characteristics of memory
devices is not acceptable as means of data recovery.

NOTE 2: The repair of individual memory devices is not permitted.......


Seems a rather restrictive policy considering that all measures need to be applied to discover the truth - maybe it's a legal issue...

On a more hopeful note, we know from the interim report that the G forces experienced for the majority of passengers (front to rear of aircraft) was in the range 100 to 200G (from pathology), so it is therefore reasonable to say that the FDR and CVR experience something similar or less, and was then subjected to 400atmos pressure for 2 years. Given that these devices must pass certification at 3400 G (6.5ms) and also be able to resist 600 atmos (up to 30 days to pass certification test), I think it reasonable to assume that the contents have survived even after all this time. However, I'm not so confident about the FADEC's and the other 12 or 13 subsystems with non-volitile memory devices which would have minimal, if at all, protection from pressure.

BEA and the Yemenair A310-300 accident at Moroni, Comores
 

noske and rotor12,

As you know, the BEA was only one of the participants in the Yemenair accident investigation, which protocols dictate was the responsibility of the tiny state where the A310 crashed: Comores. Their agency, ANACM (Agence Nationale de l’Aviation Civile et de la Météorologie) has not closed its deliberation.

The logistical and political difficulties faced by expert accident investigators working abroad in third-world countries were the subject of a recent programme on BBC Radio4:

BBC - BBC Radio 4 Programmes - File on 4, Air Crashes

BJ-ENG

Given that this is a technical forum I wonder if one could explain the following:

How does 100-200gs on a persons body equate to G loadings on a CVR/FDR? innards

How does 200 atmos pressure at the bottom of the sea cause damage to a memory chip in a module if it is exposed to water intrusion?

Recorded parameters
 

RR_DNB,
I fear you are being far too optimistic. Although I expect all of the data to be recoverable, the recorded capacity is limited. The Honeywell 4X Unit has a max recording rate of 256 words/sec, probably only enough to record a couple of hundred or so parameters at a reasonable rate, out of the many thousands of data items flashing between systems. The data recorded will typically be validated data, not necessary reality, not diagnostic data and not 'raw' data; it should be sufficient to identify the system(s) that are contributory, but not why a particular system 'went wrong'. A typical example in point is QF72; both recorders are available, as well as a bunch of expert witnesses, but "everything" is still far from understood.

Information gold mine
 

BJ-ENG,

Considering the careful way the whole issue is being handled we could expect they would avoid any risk in loosing the valuable information we hope is still present at both CSMU´s.

The non tangible (and tangible) value of the info after it´s analysis is "immensurable".

"Everything"
 

VicMel,

In the meaning of "what is really important" to understand what went wrong so dramatically to make a "complex System" fail in that way.

The ultimate goal is to learn how to avoid similar cases, addressing all issues involved.

The synergy between FDR and CVR info normally allow a very good UNDERSTANDING.

But i agree with you that we will be trying to understand the "reaction" of a complex System (in the sense an A/B is a "flying System").

We hope the 4X recorder with it´s (uncompressed) 37 MB could shown what we need.

Indeed a fascinating issue to many of us.

PS

IMO we will understand. Bu i´m not so confident that A/B will be able to learn everything to make their Systems perfect.

Reagan´s SDI (complex system) lack "Testability". Complex Systems are intrinsically unpredictable.

I prefer K.I.S.S. solutions with more "human touch" and involvement.

This trend makes me feel "too much alert" to situations i could not even understand when they arise.

A/B flying Systems has an history of "complex behavior" impossible to be solved IN TIME by the crew...

Noske, if memory serves me well, BEA gave up after a couple of attempts. The "whole month" was spent on negotiating with the US authorities on how to maintain French judicial custody of the unit while shipping it to Honeywell during the Christmas/New Year time period.

RR NDB
 

Forgive me, please, but this statement from you:

"A/B flying Systems has an history of "complex behavior" impossible to be solved IN TIME by the crew..."

...only rings true to me if you add "... "Impossible" because the crews didn't have an adequate understanding of the rather complex system they were piloting."

Or am I missing some incidents/accidents where more knowledge wouldn't have helped the crew solve their particular complex behavior issue?

Is there any information on when we may expect to hear if the data on the FDR is good or has been lost?

Effects of water pressure
 

It might not be so much pressure as the rate of increase of water pressure.

I am thinking "water hammer" if the leak is sudden.

THL,

Knowledge always is necessary. But remember the crews that faced and unsolved "different behavior" (i replaced complex by different) were trained in SIM, by all the required Standards and they failed.

I don´t think what happened in many cases was just an issue related to complex systems they piloted. It involved much more, like issues later improved by software updates, hardware improvements, etc.

Even in a perfect complex System the "real world" presents new situations.

And our capabilities are much more dependable on what we learned before a "completely unusual situation". When they arise in complex Systems our capability to solve it (with the several steps since the start of the issue to the adequate solution) IN TIME as you know, is LIMITED.

As a Designer i am not contrary to new solutions, automation, etc.

But i understand the limitations of Systems in the hands of "normal professionals" in "real world environment".

And i respect the K.I.S.S. design approach and threat´s of Murphy´s law.

The crew never of Northwest Airlines 727-25 N26 4US near Thiells Dec 1, 1974 never understood in time what happened to their "simple" a/c in the ferry flight.

Where a simple interlock could have prevented the unreliable air speed that triggered the accident.

Mac
 

Thanks for the response RR NDB, I believe I understand your point. I guess what I'm trying to say is that while the a/c become more complex, the knowledge required to successfully operate them takes more study and training than it used to.... Training departments, and Pilots MUST keep up!

What RR_NDB is referring to is N274US, a Northwest 727-25 on a ferry to Buffalo, that crashed after stalling following near simultaneous freezing of all three independent pitot systems. The pitot heating was not selected because the 3 crew members got tangled up in the pre-flight checklist.

info update BEA
 

Information, 2 May 2011


"2nd May 2011 briefing

The Remora 6000 made a further dive yesterday evening. The search operations are continuing in order to localize the Cockpit Voice Recorder (CVR).
After the Flight Data Recorder (FDR) was raised, the BEA, as agreed, asked the French Navy to send a patrol boat to transport the recorders to Cayenne (French Guyana), from where they will be transferred by airplane to the BEA in Le Bourget.
The BEA Investigator-in-Charge, an officer from the French judicial police and an investigator from CENIPA (the Brazilian equivalent of the BEA) will be present during the transfer from the Ile de Sein to the BEA, which should take around ten days.
At present, there are 69 people on board the ship, an officer from the Brazilian Navy having joined last Friday (29 April). "

BEA, with their dubious track record, will doubtless try to put a 'spin' on the FDR evidence so that Air France and AI will not be unduly embarrassed. Difficult to imagine how they will get away with this one, but try they shall.

Frozen pitots
 

Quote from deSitter:
What RR_NDB is referring to is N274US, a Northwest 727-25 on a ferry to Buffalo, that crashed after stalling following near simultaneous freezing of all three independent pitot systems. The pitot heating was not selected because the 3 crew members got tangled up in the pre-flight checklist.

Hope you will not mind me adding some words of explanation for those not familiar with pitot-probes or the above accident (even I was young then), since freezing of pitot tubes is presumed to be material to AF447.

The tubes froze completely at some stage in the climb, trapping the current dynamic pressure on the dynamic side of each (mechanical) pressure capsule. Because the aircraft was climbing, the static pressure continued to fall steadily, transmitted to the static side of each aneroid (pressure) capsule from the (unfrozen) static ports. The increasing differential pressure across each capsule led to a steady increase in the IAS reading on the cockpit ASIs, even to the point of an overspeed warning. The PF allegedly flew indicated speed in preference to attitude, with a stall being the inevitable result. (Unlike its British T-tail counterparts, the B727 did not have a stall-ident/stick-pusher system, although it did have a stick shaker.)

It appears that the scenario was dependent on the whole of each pitot head being frozen, including the small water drain-hole which is normally placed near the aft end of a pitot tube. If that hole remains open, the dynamic pressure is able to dissipate; leading to an under-reading, or zero-reading, of airspeed.

In the case of AF447, history of previous anomalies on A330s and A340s suggests that the drain holes may have remained open. Also, when maintaining a pressure altitude, the static pressure remains constant by definition (although an automatic altimeter correction for speed-related errors at the static ports may complicate the issue slightly). The result may have lead to an under-reading of airspeed. However, I'm not aware that the BEA has discussed publicly the issue and/or effects of drain holes being frozen or remaining unfrozen. It would certainly be crucial during climb or descent.


PS
In those days (the early 1970s), US jet transports like the B727 and B707/720 were still being certificated with scratch-foil flight-recorders capable of handling fewer than 10 parameters. Their British counterparts (BAC 1-11, VC10 and Trident) had analogue, wire-recording FDRs with many times the number of parameters. They paved the way for what became the SESMA programme, which monitors crew/aircraft performance on all flights. I think the Americans can lay claim to having caught up with the Brits by now...

Neptunus Rex posted :

"BEA, with their dubious track record, will doubtless try to put a 'spin' on the FDR evidence so that Air France and AI will not be unduly embarrassed. Difficult to imagine how they will get away with this one, but try they shall."

I understand that some forum members are sceptical about the independence of the BEA and that they believe that there is a conspiracy of some kind afoot .

Can we now take that viewpoint as a given rather than have it tiresomely repeated every few pages?

Additionally I noticed a reference by a poster to this thread being 'fun'. It is certainly intellectually challenging and of huge interest - not least to me as a frequent flier to Brasil and a past user of AF447 (the last time on a Friday the 13th!).

With reference to the circumstances however, and numbers reading posts, might I suggest that use of the word 'fun' may not be wholly appropriate?

I'm truly sure that nothing inappropriate was intended, just an unfortunate use of the word.

Back to SLF lurking!

Hi,

French politician at work ...........

Google Vertaling

:confused:

Hi,

Standards in force BEFORE the accident
AF reviewed the standards of training for such "different behavior" some months after the AF447 tragedy and all AF pilots were send again to the sim !

Neptunus Rex

Its interesting to note that even if the BEA is not sucessful in their nefearious deed that they will have been deemed to at least tried :E

dufc

I agree that it is stipulated that there is a contingent of BEA sceptics aboard. I think the repetition may come from anxiety that it will be overlooked, as BEA's history is well defended here also. Perhaps a truce until we have more results in the form of releases of prose and pictures.

More importantly, I apologize to anyone who was offended by my choice of the word "fun" in describing my experiences here. It was selfish, and unthinking. In my defense I will say that I have spent far more time in some anxiety and gloom, thinking of how the people may have felt in their last moments. A better word(s) would by rewarding, and fascinating.

Chris;
I really wonder if they are. I know Canada isn't. Canada has no CARS governing flight data analysis programs. The programs are entirely voluntary with no specifications on parameters or data handling. I believe the JARs are well ahead of this.

The FAA Mandatory parameters are:

- For FDAU-equipped aircraft, (Flight Data Aquisition Unit) manufactured before October 11, 1991, 22 parameters are mandatory;
- After that date, 34 parameters are mandatory
- After August 20, 2000, 57 parameters are required;
- After August 19, 2002 88 parameters are required.

Canada's requirement is essentially the same.

Eighty-eight parameters is barely adequate in terms of investigative capability.

Functional checks of the mandatory parameters are required once per year.

EASA has proposed rules requiring internal self-monitoring of mandatory parameters so that failures, (even the silent kind where a recorder just quits) can be addressed immediately instead of the next check, which could theoretically be 364 days away.

QARs for FOQA programs routinely record up to 2000 parameters many times at rates higher than required. The requirement for SSFDRs remains minimal while voluntary programs are exceeding minimums by a substantial amount. However, the FOQA QAR is mounted in or under the cockpit in the same area as the FDAU/DFDAU and as such is not crash-proof.

Upgrade and rule changes are driven by cost, not need.



The Honeywell 980-4700-42 was capable of 1300 parameters.

Anyone with a rudimentary understanding of French may be interested in the downloadable video clips on the BEA.aero website, in particular the one relating to the 'first dive'. This screengrab shows a labelled debris field...

http://i91.photobucket.com/albums/k2...screengrab.png

Did anyone notice that the clear container with a CSMU inside in the lower left hand corner of this picture: http://www.bea.aero/fr/enquetes/vol....mages/fdr3.jpg is the same one that the recovered CSMU is inside in the last picture in this sequence: Information, 2 May 2011 . I saw a quick video clip of the CMSU being put into the clear box this morning and noted that it was the same as the one that slf99 pointed out in the first picture supposedly taken before the CSMU was recovered.

PJ2,

Yes, I might have mentioned that all the SESMA and maintenance-monitoring data is acquired from the QAR in the avionics bay near the cockpit. On BA short-haul, the QAR cassette is changed every 24 hrs, normally at base; not sure what happens on long-haul.

It does seem unfortunate when crash-resistant DFDRs record only the legal minimum of data, but I don't know how the JARs compare with FARs. As for AF447, 1300 parameters sounds impressive, but one also hopes all the differing sample rates will prove to be enough for this investigation.

If AF447 impacted nose-up, I wonder if there could still be a chance of finding an intact QAR cassette...

High energy impact
 

The remains of the radome enables us estimate the forces in the a/c nose section.

Chris;

The SESMA program is excellent from what I know about it. Understand the download process.

I'm led to believe that the QAR is an optical disc rather than a PCMCIA card. Either way, I have doubts about survival/readability of the medium but we won't know until the unit is located and recovered. That may be a priority after the SSCVR is located and recovered and both sent to France, (I wonder if they'll send them on separate aircraft?).

QAR media
 

PJ2,

At pg 6 the vendor sez both cards (Optical disk or PC card):

http://www.teledyne-controls.com/pdf/FDMS-Brochure.pdf

But here:

http://www.teledyne-controls.com/pdf/OQAR_brochure.pdf

If the unit was not crushed there are chances to recover the cassete info.

On LHR 777 i wrote (Pt):

Soft landing, Hard landing ou Crash landing? | insiderNews

With an analogy between World Economy hit by US financial Tsunami and the BA 777 at LHR. A short article for a Stock Market website.

Vidcaps from BEA video "Premiere Plongee":

http://i337.photobucket.com/albums/n...Avidcap010.jpg


http://i337.photobucket.com/albums/n...Avidcap011.jpg


http://i337.photobucket.com/albums/n...Avidcap012.jpg


http://i337.photobucket.com/albums/n...Avidcap013.jpg


http://i337.photobucket.com/albums/n...Avidcap015.jpg


Composite of 3 vidcaps to reveal a bit more:

http://i337.photobucket.com/albums/n...Avidcap111.jpg

I don't think so.

The soft frangible parts are extremely difficult to decipher vectors. Far easier to look at the strongest parts of the aircraft like the cockpit or engines or at least something with wide breadth to it.

This is a mockup to train for the recovery...

Then once they bring it up, it has to be sealed in fresh water for an amount of time to leach out any salt intrusions, and perhaps pressure treatment.

(if you look closely, you will see that the 2 images are not of the same containment vessel)

Falling tail first and "aluminum shower" speed
 

lomapaseo,

What i thought:

1) a/c hit surface at pitch 5~15
2) nose drops
3) lower part of radome breaks in several parts
4) higher part (the one found) shows coherence to the model at 1) and 2)

I agree that for evaluate g´s better to look not to light composites.

The radome was my "reference" of the impact of a/c nose cone "falling"

In a low energy impact would just break and detach. It shown like crushed from vertical deceleration despite it´s lightweight.

Question:

Supposing first impact at tail and second at a/c nose. (falling some degrees pitch up) how we compare the g´s. My feeling is the second (nose) is more intense. Is it a valid assumption? The difference may be negligible, i guess.

PS

We can imagine the "aluminum shower" in the water to the bottom.

Question:

Roughly how many minutes estimated? Supposing no big currents.

Mac,

What are you trying to figure out?

There are many factors.....

Mind models
 

FlightPathOBN,

In order to imagine some models:

After a/c hit water how many feet (horizontal) the debris probably traveled?

They publicized the xy coordinates of the wreckage?

What´s the wreckage field distance to LKP?

The question of "dive time" is because they mentioned "sea currents" for defining search area priorities.

And considering the debris concentration this appears to me meaningless.

And it seems revealing "gross errors". Why? Just delayed the investigation.

RR,

Not sure what delayed the investigation means.

I have significant experience in underwater recovery.

There are many factors, when even if you know the surface impact point, determine the search grid and potential bottom location.

I understand the depth of the 447 was substantial. This in itself, leads to quite a bit of research to ascertain the bottom location. All of the materials will ave a relative density as in relates to the seawater. The seawater also has bands of density in relation to currents, temperatures, salinity, and MET factors. Surface conditions at the time of impacts need to be considered as well.

The oceans have different describable levels, as levels of dense, cold water, will migrate, much as the currents in the air. Pilots are aware of all of the different layers of air, temperature, and density, and the gulf stream affects, the same is true of the underwater 'currents', nothing is ever static.

You mentioned the 'aluminum shower', well the debris field can be theoretically calculated, but as we have all noted in the Titanic, that mass took quite the effort to find. The pieces of the aircraft would have a relative density to the seawater on descent, not to mention the buoyancy effects of the shapes in flutter. Given that, the denser, less aerodynamic pieces, would tend to descend rapidly, hence, more directly straight down, and due to pressure at depth, would potentially accelerate at the same rate. I believe what we are seeing in the underwater images, is all of the 'pieces' of the near the same relative aerodynamic density with regards to the descent.
To me, this means that the impact of the vessel had little forward momentum, as the density of impact, had there been significant forward motion, would have left assemblies of relative density along a trail.

How can 100-200g relate to the relatively undamaged galley unit (sorry I forget the correct name but I’m sure the experts know what I mean)?

Chris,

Can you explain further....

A 'unit' may have characteristics that 'fly' or have specific properties underwater.

Forensic Taphonomy is a pseudo science, and there is certainly an artform associated with the forensic evaluations.....

ChrisN

Are you talking about decell rates?? and where did they come from??

With a car, your bumper, grill and fenders do a great job of taking an instantaneous high G load and spread it over a couple of seconds before you feel it. In that period of time much energy is being absorbed to lessen the load on your innards.

OTOH if you manage to get a uncrushable object like an engine block between you and the immovable brick wall your G loads are gonna be a lot higher.

Many of the box structures a couple of feet above the airplane keel beam are not gonna feel the same load as the belly structure simply due to the crushing and speading out in area of the load as well as in time.

All I'm doing in this explanation is explaining why your eye sees things the way they are in a photo.

In real life I've seen occupants that have survived sitting within feet of those that haven't all due to variances in G load and their relative nearness to an unbuckled part of the airframe in a given accident.

Shape of debris field an orientation
 

FlightPathOBN,

I am speaking on the first searches far from LKP. They spent time and effort doing that and delaying the current phase.

Amazingly wreckage is very near LKP. How near? Any info. on the distance?

The shape of debris field and orientation are related to the direction the plane hit sea surface? They informed the direction?

In a previous post i made some questions on travel time of debris. How much time you estimate? In your opinion the debris did fall near vertical?

What surprised me was the concentration of them. And the flat terrain.

Last question: What´s the estimated temperature at bottom?

chrisN

Howdy. Lomapaseo's proposed g values are astronomic, and frankly, it is hard to disagree with him. I have offered my experience of dealing with similar structures in falls and warehouse damage that exhibited far more damage than the Galley stack, and some other recovered bits. It is a monstrous disconnect, one I cannot get my arms around, at all. If she hit tail on, in a flat aspect, the cockpit became the crack on the whip, perhaps even accelerating as the tail stopped. The forward part of the a/c hit with certainly the same velocity as the tail, if not more. The crew rest, in the hold, suffered a severe display of flattening, not just from stopping on the hold floor, (and the sea's surface), but then suffering the collapse of two more diaphragms on top, the cabin floor, and the dorsal skin of the fuse.

Initially, I thought a hull rupture at altitude had allowed some contents to spill out, which would have had a far less emphatic impact with the sea, being in free fall, etc. I don't think that is wrong, but I continually adjust the altitude at which the contents separated from the a/c. Beyond the impact, the potential damage to be done as the a/c spreads out in a surface field on a stormy sea, and the potential for damage beyond that which pieces may have endured on impact, the galley cabinet with its intact food boxes is either a miracle or a hint at separation from the hull prior to impact. In any case, if photos of the cockpit are allowed, the disconnect will be affirmed as other than simultaneous and co-located sea entry, imo.