AF447

[NotPilotAtALL]

Hello,

There is way too much blame being assigned here to a single entity be it a manufacturer or a single system.

The investigation is on-going and damn few facts are available to establish a causal chain. This thread is begining to look amateurish in its over simplification of what went wrong and who is to blame.

The idea that anybody involved with the investigation would not like to find the critical evidence (FDR/CVR etc.) is outlandish and demeans a supposedly professional internet site.

This is not about blame, lawyers or insurrance carriers, it's about finding enough information to prevent a similar accident.

Root cause is hogwash. Nobody can effectively eliminate any system malfunction whether it be pitot tube interactions with computers or a weather radar outage.

Our efforts need to be focused on minimization (of an occurrence) and mitigation (of the resulting hazard) To do this you need data on all the interactions that were at play in this accident including the pilots.

In today's machine you can not take the pilot out of the accident causal chain, yet nobody is even considering the what-ifs in this discussion. Remember for every system failure including the pilots there is a cause effect and that is ultimately how the corrective action will be accomodated.

Very interesting and informative post about different matters ..

Now .. if we give a raw sight on the aviation accidents statistics .. the figures show .. in almost 100 accidents .. pilot error is responsible for 80 .. so a average of 80% (nothing so far proof this new accident will inflate this 80%)
This average was different in the past .. but cause the progress of technologies .. technical failures are reduced at best engineering can.
So it's seem's that the corrections for avoid accidents work mainly when applied to technical parts and not so best on humans.
So we can conclude .. the technology progress (sometime with big jumps) but unfortunately the human stay the same .....
It's possible the strong response for avoid at max the possibilities of accident will be to reduce the most possible the human actions in a plane .. and ultimately (after some more big technology jumps) remove humans pilots from plane ??

Regards.

[JD-EE]

Pontius, I think you misunderstand me.

I am envisioning a joystick that is simply sitting up in the air. Pick one for your favorite game. Hold the base. Now move it rapidly left and right. What is the effect as read on your computer screen? Does it stay put? Does it move as if somebody had their hand on the stick and was exerting control?

Experience suggests you will find that your computer gets input that it does not distinguish from that you might have given yourself. Heck, you can see the stick flopping around.

Now imagine it runs your automobile. Forward means go forward, the further forward the faster. And so forth. Now suppose you hit an obstacle, a speed bump. Your body jerks forward. Your car lurches forward into a crash. This is not a problem if you're sitting comfortably with no sudden accelerations over a small fraction of a G. But add in jerks from turbulence and you might find your arm involuntarily inputting a control surface change that was out of the range the plane could tolerate.

Meanwhile, what are your feet doing as you bounce six inches up and down? Might such a bounce have led to the pilot overcontrolling in several of the accidents attributed to "stupid pilot over controlled the plane?"

(Maybe a high-G accelerometer set needs to be mounted in the cockpit so that the data recorders can note that it's not the pilot's fault he could not control the plane, nothing could.)

[ClippedCub]

and ultimately (after some more big technology jumps) remove humans pilots from plane ??

You don't want that in passenger planes. The software specification writers can't think of all combination and sequence of failures. Besides, computers gives up when they reach, 'that does not compute', the pilot never gives up.

Some of those pilot error causes are due to poor training, and should be labeled training requirement error - like Colgan 3407 for instance.

[h3dxb]

@JD-EE

I believe it's been established there is no dish. They use Inmarsat as their carrier. And the antenna is omnidirectional for this mode

.

Believe what U want

For antenna control steering and computed doppler correction, the following ARINC 429 octal labels are transmitted from the ADIRU to the SDU:
310 Present position - Latitude
311 Present position - Longitude
312 Ground speed
313 Track angle
314 True heading
324 Pitch angle
325 Roll angle
361 Inertial altitude

This is required for:

The high gain antenna is an electronically steerable phased array. Simultaneous transmission and reception of satellite signals (full duplex operation) provides two bands of operation: receive (Rx) band, and transmit (Tx) band.
Beam steering of the antenna is performed via serial transmission of phase shifter data from the BSU. RF signals are transmitted to and received from the D/LNA unit.
The high gain antenna provides +12dBic nominal gain with near hemispherical coverage and enables transmission of high rate data and voice communications.

But U can ask me every time

[Graybeard]

TCAS Fail Unrelated to Other Faults

Thanks for the translation of the TCAS Fault message, h3dxb. Now, let's see if your manuals can confirm or refute the following:

Ref: "There are five official, damning reports on A330 and A340 aircraft operated by major airlines suffering complete failure of all flight instruments in flight. Pitot probes iced up in meteorological conditions at night, by day, in cloud and in clear air."

The early failures reported by AF447 ACARS all seem to be explainable by iced pitot tubes, save for the TCAS Fail. TCAS does not use airspeed, just altitude via the transponder.

Air data Altitude failure will cause the transponder to revert to Mode A. I'm not sure if it would report failure of its altitude input to the CMS/ACARS, but it didn't report a Fault.

Transponder reversion to Mode A will not cause a TCAS Fail condition, but a TCAS OFF condition.

Therefore, we have a TCAS Fault that occurred nearly simultaneously with the airspeed sourced faults, but unrelated.
-------

GB

[Smilin_Ed]

Unrelated? Maybe.

Therefore, we have a TCAS Fault that occurred nearly simultaneously with the airspeed sourced faults, but unrelated.

I don't think we can conclude that they are absolutely unrelated. They may well be unrelated from a systems viewpoint, but what if the failures were both the result of catastrophic airframe failure or lightning strike?

[Mike-Bracknell]

I am envisioning a joystick that is simply sitting up in the air. Pick one for your favorite game. Hold the base. Now move it rapidly left and right. What is the effect as read on your computer screen? Does it stay put? Does it move as if somebody had their hand on the stick and was exerting control?

One would hope that the sprung self-centring of the joystick versus the movable part's weight with no hand on it would far exceed the transient g-forces available in flight to move it on it's own?

I have no experience of the A330 stick, but I assume it's got to have a certain amount of sprung(?) self-centring otherwise it's one hell of a choppy ride when the pilot wants to scratch his chin.

[Gringobr]

What the bodies say

Some of this info is new.. Passengers unconscious due to decompression.
Aircraft hit the sea probably in a near horizontal attitude.
I am taking this information from a local magzine called Veja dated 17 June (but available for sale today), page 69.. It is a top quality magazine and has the best analyses etc I have seen on this accident.
Summarizing because I ( and others) have already posted some details here
1 There are no burns etc (no bomb/explosion/fire) and the bone structures are basicaly preserved.. means the plane did not fully disintegrate in the air and/or at least the part where they were sititng arrived at the ocean relatively intact
2 What the doctors here call "4 fractures" upper parts of arms and legs ( 4 parts) broken.. This happens with accidents in which aircraft hit the sea with passengers sitting in their seats
3 Rosy (coloured) teeth. this symptom means the passengers were subjected to decompression , sudden exposure to rarified air and -56 degrees, they lost consciousness quickly before dying (all bodies rescued to date)
4 Clothes and documents, many of the bodies (but not all) had their clothes relatively intact, with documents and boarding passes. This means a goo part of the part of the cabin where they were seated hit the sea relatively intact

[bravoxray]

I have been reading this very informative thread with great interest and as a one time private single engine VFR pilot learned an awful lot . More a passenger now but not a terrified one. Air travel is still statistically very safe IMO compared to other mediums.
Cannot understand the obsession of this low tech ram air pitot tubes as an explanation for this accident. They are basically all the same in principle regardless who manufactured it. How the various computers process the input data is a different ball game.
The accident was surely a sum of various problems, intrinsic and extrinsic.
I do take offense though for even suggesting/insinuating pilot errors in this accident without any shred of evidence. They would not have flown into the center of a CB as very qualified experienced pilots and am sure were aware of past (temporary) icing problems on subject pitot tubes but was not considered a nogo.
I do have a question though for prof. com. pilots. Is there any reason why there are not multiple FDR's and CVR's located in the aeroplane, excluding the cost factor? The weight factor appears to be minimal.
The FDR info, if certain flight/monitoring devices failed, would that not diminish the outputs recorded on the FDR at critical moments?
I apologize if I do sound ignorant.
On the other hand, some of the comments I have seen are so ludricous that would not be surprised to see a recommendation for parachutes for all passengers. Kind of funny at 35 fl. , probably better shape upon impact but still dead.

[HarryMann]

Will a stick with no hand on it stay perfectly still poking up there in the air above the mounting point as the plane accelerates with rapid jerks in random directions?

Mike, you beat me to it! It's breakout force would hopefully exceed any likely in-flight accelerations.

This is independent of whether the stick's output is controlling the plane or not. It's a question about whether under very heavy turbulence the input to the stick from the effects of the turbulence is stable enough to fly the plane. If not and the pilot HAS control isn't it possible that the bouncing around could lead to improper or damaging imputs to the flight control system?

It is possible though, that under extreme turbulence, consistently accurate pilot flight control inputs in Alternate or Direct Law may be less easy than with a larger, less sensitive device having larger amplitude stop-to-stop movements e.g. a conventional control column - judged without having any experience with an AB style joystick.

[Jetdoc]

I have a drawing of the A330 side stick controller showing the guts of it. Its a pretty beefy unit. I don't think it will flop around too much. I don't have anyway of posting it but if anyone wants to do so, I'll send it to them.

Here's something about ACARS messages. I was sitting on an A320 the other day. It was on overhaul and some testing was being done on it. The system under test failed and within seconds we got a message that the failure had been transmitted.

I would bet that alot of people here probably just learned that fault messages are transmitted on ACARS. Remember that the primary reason for doing so is to give advance warning of system faults so that maintenance personnel can prepare for troubleshooting and repair. This is especially useful during short turnarounds. There is no need for Lat/Long info and there is no information that would be hidden by Air France or Airbus.

[Jetdoc]

JD-EE

(Maybe a high-G accelerometer set needs to be mounted in the cockpit so that the data recorders can note that it's not the pilot's fault he could not control the plane, nothing could.)

There is one installed on the aircraft specifically reporting to the Flight Data Recorder.

[Jetdoc]

A330 Airbus Side Stick Controller

[WNcommuter]

ChrisVJ wrote:

There is little evidence so far in statistical terms that FBW is any less safe than the old physical connection controls.

Suppose hypothetically that FBW reduces errors and improves safety overall but suppose that physical connection reduces pilot error in conditions exceeding the abilities of the FBW software. Wouldn't tort lawyers have a field day blaming FBW design for its shortcomings, or blaming non-FBW for its shortcomings, depending on what type of accident happened? This could be a no-win situation for aircraft designers.

[WhyIsThereAir]

Jetdoc -

Thanks for the image, that looks like a nice beefy stick with redundant linkages. I will be willing to believe that the Airbus engineers considered turbulance and saw to it that the stick is dynamically balanced against side thrusts, *when sitting there untouched*.

Now: Is there a procedure for dynamically balancing the sidestick *taking into account the mass of the PF's hand and arm*? Arms and hands of PFs (and PNFs) are also subject to lateral accelerations in turbulance. If the hand is coupled to the stick, is it not possible that the moment arm of the hand and arm might couple unintended movement into the stick?

If there is a procedure to balance the stick *with hand and arm attached* against spurious accelerations, then this is not a problem. Otherwise perhaps it could be in some cases.

[JD-EE]

h3dxb

For the Satellite Telephone system and other high speed connections I well believe there is an aimed antenna. ACARS could be using something like, but more modern than, Inmarsat-C which had a low data rate, small messages, and an omnidirectional antenna. It "sort of" makes sense to send ACARS through such a connection. That is why I believed it when someone else mentioned that ACARS used an omnidirectional antenna. Telephones (and Internet) would need a high gain steered patch antenna to be effective. For ACARS that might not be an issue.

[Mad (Flt) Scientist]

Originally Posted by Graybeard

The early failures reported by AF447 ACARS all seem to be explainable by iced pitot tubes, save for the TCAS Fail. TCAS does not use airspeed, just altitude via the transponder.

Air data Altitude failure will cause the transponder to revert to Mode A. I'm not sure if it would report failure of its altitude input to the CMS/ACARS, but it didn't report a Fault.

Transponder reversion to Mode A will not cause a TCAS Fail condition, but a TCAS OFF condition.

Therefore, we have a TCAS Fault that occurred nearly simultaneously with the airspeed sourced faults, but unrelated.

Question: does the entire ADR section of an ADIRU "fail" as one unit? That is, having detected that it is producing "unreliable" airspeed data (through the ADR DISAGREE method) is the entire ADR output regarded as suspect by other systems?

If so then you don't need any fault on the altitude to trigger the TCAS message - the airspeed problems cause the other aircraft systems - including TCAS - to regard all ADR info from the identified "unreliable" ADIRU(s) as suspect.

From a quick scan of such AB docs as are available easily online, I can only find references that say things like:

Each ADIRU is divided in two parts, either of witch [sic] can work separately in case of failure of the other.

Which would tend to suggest that the whole of the ADR part is considered to fail as a unit.

I can see logic for either approach - once a source of data is suspect, you may wish to disregard it entirely, but there's also an argument for trying to use what may still be valid data streams from a partial;ly invalid source. Hopefully someone know which applies.

[Jetdoc]

WhyIsThereAir

There are about 73 pages of maintenance manual procedures regarding the various parts of the side stick and 8 pages of testing. There is some 'rigging' involved so to speak regarding the transducers but no balancing.
I do know that the pilots arm rest adjustment is important when using the side stick but thats best explained by an airbus pilot.

[md80fanatic]

What about the cheese?

I would like to advise caution when chasing a single goose. It's still possible the pitots were not frozen over and the plane never entered extreme turbulence. We simply do not know anything yet. We have perhaps one, maybe two holes in the cheese, with certainly many to follow. We should keep this discussion academic IMO ..... trying hard not to cast, or giving the impression of casting, blame in any particular direction. I'm finding this thread very educational, and I would like to thank everyone for their contributions.

[h3dxb]

@JD-EE

Unfortunanly it's called SATCOM Antenna not ACARS Antenna, and this one is steerable.

@Graybeard

Anyhow, going through the downlinked ACARS messages like published in "Der Spiegel" 24/2009 page 129, all warnings happened from 0210Z.

There are 2 failures downlinked with the same time tag:
341115 and 279334 .
Checking the TSM for the operational MSG F/CTL ALTN LAW with failure in 341115 it points to PROBE PITOT 1+2/2+3/1+3 detected by EFCS1 or 2.

279334 does not refer to ALTN LAW, its going direction RUD TRVL LIM FAULT

Referring TSM
TASK 27-91-00-810-822
Disagree of the Pitot Probe Data in the FCPCs



The rest on you guys , I'm out now, let my name out of the game