AF447

[greenspinner]

To MoateAir
QUOTE
FROM : AIRBUS FLIGHT SAFETY DEPARTMENT TOULOUSE
ACCIDENT INFORMATION TELEX - ACCIDENT INFORMATION TELEX
SUBJECT: AF447 ACCIDENT INTO THE ATLANTIC OCEAN
OUR REF: AF447 AIT 2 June 4th 2009
PREVIOUS REF:
- Ref 1: AF447 AIT 1 dated June 1st 2009

This AIT is an update of the previous AIT n°1 concerning the AF447
accident into the Atlantic ocean on June 1st, 2009.
In line with the ICAO Annex 13 recommendations, the French
investigation Board - BEA (Bureau d'Enquêtes et d'Analyses) is
leading the technical investigation, with accredited representatives
from the Brazilian Investigation Board and US NTSB, with Airbus
providing technical support.
The following data have been approved for release by the French BEA.
The route of the aircraft was crossing a tropical multicell
convective area at the time of the accident.
Failure/ maintenance messages have been transmitted automatically
from the aircraft to the airline maintenance center.
The above mentionned messages indicate that there was inconsistency
between the different measured airspeeds. Therefore and without
prejudging the final outcome of the investigation, the data available
leads Airbus to remind operators what are the applicable operational
recommendations in case of unreliable airspeed indication.
The following operational procedures are available for the Airbus
Aircraft Type :
-A300: QRH 13.01 thru 13.03, FCOM 8.05.10;
-A310: QRH 13.01 thru 13.03, FCOM 2.05.80;
-A300-600: QRH 13.01 thru 13.03, FCOM 2.05.80;
-A318/A318/A320/A321 family: QRH 2.15 thru 2.18A, FCOM 3.02.34;
-A330/A340 Family: QRH 2.21 thru 2.23B , FCOM 3.02.34;
-A380: ECAM not-sensed procedures, FCOM - Procedures / ECAM
Abnormal and Emergency Procedures / 34 Navigation.
An update on the accident data will be provided as soon as further
valuable information is approved for release by the Investigation
Board.

UNQUOTE

[Flight Safety]

Many thanks to all for the ACARS data and interpretation. Sounds like probe icing or perhaps hail damage to the probes to me.

[MR8]

Gentlemen,

The FACTS we got from the ACARS messages clearly point to one possible cause: Unreliable Airspeed Indication/ADR check procedure. This is one of the most difficult failures you can have in an A330/A340 since you loose ALL air data. No speed or altitude info, not even on the stand-by system until you recover the 'right' ADR if any. The checklist asks to switch off the faulty ADR if known.
Anyway, the procedure is to fly pitch/power. Then troubleshooting can be done by using e.g. GPS speed against predicted wind (the wind on the ND is wrong), GPS altitude, etc... This is extremely difficult to accomplish, so I imagine barely impossible in the reported heavy turbulence at high altitudes with changing winds because of TS's.

Rests the question how such a failure can occur: given the flight conditions: either the pitots froze up because of severe icing, or the radome disintegrated due to hail impact (remember the pictures of that easyJet B737 in GVA that entered a CB? And that was just a 'European' CB, not a tropical one.)

How is it possible that this aircraft with experienced pilots fly through a dangerous CB? (This is pure speculation from my side!!)

First of all, not all CB's go with associated lightning, so if the radar is not on, they might be quite hard to spot on a dark night. (At the time of the events, the moon at present position was setting in the west while the aircraft was flying north east bound. No moonlight must be considered)
Why might the radar not be on? Ask ANY A330/A340 pilot and they will all agree that at some stage in their career they were caught out with the previous crew dimming the radar. For the non-Airbus guys: the ND 'control' knob is twofold. The inner knob is the dim for the ND display itself, while the outer ring is the dim function for the WX radar/Terrain information. If the outer knob is turned in the dim position, no WX/Terrain information will be visible on the ND. The WX radar with the tilt info (not the actual WX info) will be in bright since that is ND info, not WX info.
Looking at the SIGMET chart for that day, it seems that the route was basically clear of weather until they reached that wall of CB's around the ITCZ. So maybe they didn't have a chance to pick up that the radar was dimmed until it was too late.

Again, the last part is complete speculation and I would like to have input on my thoughts by experiences Airbus pilots.

MR8


Note: I do not believe the crew lost pressurization. Unreliable Airspeed Indication/ADR Check Procedure Checklist asks to switch off 2 ADR's if the affected ADR's can not be identified or if all ADR's are affected. If the ADR's that feed the active Cabin Pressure Controller are switched off, this will give a 'soft' warning (Amber Caution with no 'ding' - sorry, forgot the actual name of such a warning). Hence in my opinion it is just an ADVISORY on the ACARS and not a failure. Can someone please confirm the ACARS message regarding ATA21?

[Lemurian]

From greespinner, post #73 :

341234 – ADIRU
279334 – EFCS
341115 – PROBE – PITOT

If one looks at all the ATA, we basically have a majority of 34xxxxxx and 22xxxxxx.
The maintenance status 341115 refers to pitot probe (-s ?).
Then what is common to ADIRUS, ISIS, 34xx flags on PFDs, rudder travel limiter ?
Bar a major electrics break-down, I can only see air data inputs.
Then the ADRs Disagree point towards the same origin : differing input data to the ADRs, which again puts the finger in the direction of faulty air data sensors.
So what can cause these faults ?
Icing.
Pitot probe icing.
Although unlikely and far-fetched, that's my only explanation, fwiw.

14:18
the Airbus telex greenspinner quotes confirms "unreliable airspeeds", and he posted it while I was typing.
Any thoughts ?

[grebllaw123d]

greenspinner,

Thank you very much for your valuable information in your last post!

- Do you know which failure flag/flags might have been displayed in the PFDs with the mentioned failure codes (341200106/341201106)?

- Which information would be missing on the PFDs in case of this failure - do you have an idea? Airspeed? Altitude? Vertical speed etc....

Thks

[Wossname]

Cap'n Crunch

Forgive me, a couple of dumb questions for my own edification. Firstly, if we don't know what caused AF to go down until the report comes out, what steps should crew take to avoid the same fate? Secondly, is it really the case that there are so many commercial pilots so unused to high-FL hand flying that doing so in less than optimal conditions might result in an upset?

Regards

[md-100]

Does anyone know the upper and lower speed limit for FL350 and the "weight" at that time of flight?

just wondering... a total pitot-static failure (ice) then all data went off then a hand flying in the middle of CBs?? and too close to coffin corner.. then stall or maybe exceed the max speed and overstressed the aircraft..

just wondering..

[Feathers McGraw]

Reading about the A330 pitot icing comments that have been made, is it being suggested that the problem is caused by setting probe heat to 'auto' and the ice detection then not being sensitive enough to prevent rapid icing when conditions make that likely?

I got an impression from one post that even with probe heat set to 'on' rapid icing could happen, which clearly would raise questions about the certification tests done on this system.

It seems that there is some concern about the A330 in this regard, is it justified?

[clivewatson]

I have no intention of speculating of the possible cause/s of this tragic accident, but I would like to ask a question that relates to the Airbus radar.

Several posters have mentioned that it is possible to have the radar "dimmed" and that this may not be noticed by the crew.

What I don't understand is how it is possible to select, test, tilt and operate the radar if it is dimmed to an extent that it prevents seeing any image or returns. Surely, in order to start any kind of scan using the radar, the image must be visible so that tilt and intensity scans can be seen and set up for the phase of flight.

Radar is not simply something that you switch on in the hope that significant returns will pop up - the display must be seen to even test and start using it - unless Airbus have some fancy system that I have not yet come accross.

Can Airbus pilots please explain the "dimming" problem that several have mentioned

[p51guy]

Wossname, I think most professional pilots could fly at altitude using attitude and power to hand fly through turbulence. All airliners have charts for unreliable airspeed showing attitude and power settings for the altitude and weight.

[evergreek]

"just one point as many seem to believe that the thread is now becoming factual.

The released acars information appears not to be the actual information from the flight but from a search function performed on the ground. Unless my eyes are deceiving me the released pages are 28 + 29 from over 250 produced after performing a search from the 12/05 - 01/06.

That means information could be missing and that also means that this thread is actually no different from the original which has been closed. We may be witnessing a more professional response but the basic underlying information is not complete and that means we are speculating."

Note that on the right hand side of the ACARS where it says NOVO, they all came from AF447 Thats how we know its from that flight, not others.

[Captain-Crunch]

Cap'n Crunch
Forgive me, a couple of dumb questions for my own edification. Firstly, if we don't know what caused AF to go down until the report comes out, what steps should crew take to avoid the same fate? Secondly, is it really the case that there are so many commercial pilots so unused to high-FL hand flying that doing so in less than optimal conditions might result in an upset?

Regards

Wossname,

Clearly airbus has some idea of what might of happened. Our big clue, is that they have already released an accident alert to all operators with the focus on "Flight with unreliable airspeed". As my astute colleague has pointed out, this emergency is one of the worst. When it happened to me, I just set 90% power, held about x degrees pitch and waited to come into the clear. The ice was loud. The airplane felt loaded. After we regained everything minutes later, even MCT (abv climb power) was barely holding altitude. We picked up a buttload of ice. It temporarily took out all the pitots and all the static ports. The autopilot let go right in the buildups.

Luckily, I had an earlier job on junk jets with unreliable autopilots for months. I did a lot of hand flying for hours at altitude in severe weather. And at the time we iced up on the bus, I had been hand flying up and down from FL180 despite the SOP's to always stay coupled.

Yeah, it's true. Most airline pilots have never hand flown at altitude all night. But they should do it. Even if it scares everybody in the back.

The simulator is just not the same thing imho.

CC

[Rananim]

MR8,
Excellent.Either they had dual ADR failure and didnt fly the correct IAS on the standby ASI(or whatever Airbus call it) or they had all 3 go down and had to fly pitch/N1 using GPS-generated GS(which as you say would be suicidal given the conditions).
How does an Airbus pilot set the thrust for best turb-penetration speed?On a Boeing with moving thrust levers,we simply disengage and set FMC-generated TURB N1 and leave it there.Would the overspeed warning be on continuously in the assumed scenario?
The question though is how/why they flew into such weather.Shouldnt this be the subject of the first advisory from Airbus?

[parabellum]

Wossname - I assume the first part of your question is rhetorical? If we don't know what caused the incident then there is nothing we can do to avoid it happening again, all we can do is to try and avoid similar known conditions.

As the air is very thin at altitude the auto pilot does a much better job of anticipating and correcting deviations from a required flight path, it works in milli second whereas the human is much slower. Hand flying at altitude is usually demonstrated in training just to show how clumsy it really is and how uncomfortable for the passengers too, (the pilots up front often have no idea how uncomfortable hand flying is for the people down the back where the resultant effect of every move of the controls is magnified many times over in the passenger cabin), I have been on flights when the auto pilot has failed and even between the two of us we could not deliver a really smooth flight at altitude.

[Bridge Builder]

Difference between too low and too high speed at high alt is very narrow

Is it possible the plane went a bit too high, in order to to fly over the storm? I remember seeing a post last year from a fighter pilot, who said that at maximum ceiling, the difference between a slow speed stall on the one hand. and going too fast for the airframe to cope, can be very tiny.

[MR8]

p51guy - I think you are mistaken. The pitch/power tables are very rudimentary. When flying close to the coffin corner at altitude and in HEAVY turbulence, you don't stand a chance, not even if you're a stick & rudder ace.

clivewatson - as I mentioned before, the ND control know has an inner knob and an outer ring. The outer ring is to dim the Weather RETURN. The inner knob is to display the rest of the ND info, INCLUDING the WX radar mode and tilt angle.

It is Airbus procedure to check the WX radar return on taxi out when switching on the WX radar. However - and I am guilty of that as well - this is almost never done on a clear day at the departure field since there is no immediate weather threat. For TO the WX radar is set up at about 5 degrees and adjusted during climb. The adjustment is quite easy since we have a very accurate tilt indication on the ND (which - again - is NOT dimmed as it is part of the ND). After level off, the WX radar is set to -0.25 degrees to have a good look ahead and slightly below. At this setting, you won't get those annoying ground returns. So, if you forget to check for returns on the ground, you might not get returns at all until it's too late.

I personally found out the 'hard' way as a young FO when I questioned the Capt. why my Radar didn't work. The only thing I lost was my pride on that day and a beer to the Capt. and I gained a lifesaving experience.

Maybe both guys had their radar dimmed and since there was no weather out of Rio, they just didn't notice before it was too late. It is not that hard to do...

MR8

[greenspinner]

Clivewatson,
The dim of the Radar display is located (for Airbus of course) on the button of the In/Off ND display (double button) on the left hand side of the Pilot (right for the F/o), whereas the radar Ctl panel is located on the pedestal, with no command for the dim function on the display at all.
I do have one time a captain reported to me that is WR was not working at all.... I just shown him how to adjust the brighness of WR ( hopfully F/o was working, means not dimmed). To be meditate .....

[Swedish Steve]

people who routinely deal with ACARS stuff have suggested that the number of messages on these two pages is already large for a single flight.

Yes the average is Zero or One. More than 3 would be very unusual.

Reading about the A330 pitot icing comments that have been made, is it being suggested that the problem is caused by setting probe heat to 'auto' and the ice detection then not being sensitive enough to prevent rapid icing when conditions make that likely?

The setting of probe heat AUTO means that it comes on during the start procedure and is always ON in flight. All Auto does it turn it off automatically when you are parked at the gate.

[Safety Concerns]

Note that on the right hand side of the ACARS where it says NOVO, they all came from AF447 Thats how we know its from that flight, not others.

I am not disputing that the info relates to AF447. I am just raising the possibility that the information is not complete because the supplied report is NOT the actual aircraft report. The report is a ground report produced by someone performing a search. We do not know what filters were applied, if any at all or what was on page 27 or page 30.

Also interesting for example would be to see the information produced for the leg into Rio.

[Southernboy]

Airspeed

The posts from Airbus pilots regarding conflicting airspeed indications is very interesting. The Australian A330 upset in October 2008 (VH-QPA) also featured differing airspeed data between Capt & FO.

I've not flown one of these electronic marvels but would welcome comments by Airbus pilots in trying to understand this. Is this - shall we say - not uncommon? How likely are the crew to have practiced such a scenario in the sim?

There do seems to have been a lot of QRH items to resolve.