Airbus crash/training flight
Pegase Driver
Join Date: May 1997
Location: Europe
Age: 74
Posts: 3,686
Likes: 0
Received 0 Likes
on
0 Posts
I am always careful with the self proclainmed " Private incident investigators " theiy are paid by someone ( generally the Victim's famillies , not sure who pays this guy , but he ain't working for free surely ), with an agenda. He might have a point, but his speculations are as good as the BEA.
I am not an Airbus pilot nor a Airbus sofware engineer , even less an aircraft painter at EAS, but trying to stall any aeroplane while on base leg to see what happen does not sounds to me to be really a technical issue.
I am not an Airbus pilot nor a Airbus sofware engineer , even less an aircraft painter at EAS, but trying to stall any aeroplane while on base leg to see what happen does not sounds to me to be really a technical issue.
Join Date: Jul 2008
Location: UK
Age: 69
Posts: 475
Likes: 0
Received 0 Likes
on
0 Posts
but trying to stall any aeroplane while on base leg to see what happen does not sounds to me to be really a technical issue.
So there are two possible paths to take with this investigation:
1) it was crew error full stop. Had they been flying higher this perhaps would not have happened, end of investigation and wait for the next similar accident or
2) expand the investigation and potentially uncover failings or mistakes or oversight within the maintenance environment. Were there issues with 2 out of 3 AOA sensors? If yes what happened?
Were correctly trained staff employed?
Has the company a proper quality system? etc. etc. etc.
Now depending on the answers to question 2 this scenario may well end up incorporating technical issues as well.
There are other departments involved in the safety equation and nobody can state with 100% certainty that this accident would not still have occurred at 20,000 or 30,000 feet. We don't know.
I agree, the big picture is that they were testing the aircraft towards the edges of the flight envelope to check that all systems perform as they should. Obviously one of the possible outcomes is that the aircraft DOESN'T perform as it should! If allowance hasn't been made for this eventuality, you can find yourself, as this crew did, in a nasty spot. 
Join Date: Jun 2009
Location: Australia
Posts: 49
Likes: 0
Received 0 Likes
on
0 Posts
Airbus
All out of my league but. Heck arn't these AB's supposed to stop this happening? Instead of all these wonderful protections, the plane actually runs wild? I've been well into stall buffets from crop-spray height to 100 feet or so, and my planes didn't go bonkers. Makes me angry these french bureaucrats.
Pointer to good summary post pls ?
Read through most of this thread, and the interim report, but I can't get a straight picture between the 'aoa' sensors being called 'essential' and opinions that the a/c 'protected' itself into the sea.
Can someone point me to a good summary post pls ? Particularly interested in how data from the stuck AoA sensors may have misled systems or denied protection and whether the protections were in fact operating during the final sequence.
Can I also ask someone with patience how often max trim is needed in flight ? An amber warning that manual trim is needed seems a bit quiet, is there a case for warning that max trim is being applied ?
Can someone point me to a good summary post pls ? Particularly interested in how data from the stuck AoA sensors may have misled systems or denied protection and whether the protections were in fact operating during the final sequence.
Can I also ask someone with patience how often max trim is needed in flight ? An amber warning that manual trim is needed seems a bit quiet, is there a case for warning that max trim is being applied ?
Trash du Blanc
Join Date: Mar 2001
Location: KBHM
Posts: 1,185
Likes: 0
Received 0 Likes
on
0 Posts
The real danger in imminent stalls is that you are way below min control speed should a wing engine falter, or compressor stall, or catch a bird as you are recovering. Then you get to pull the good one back, drop the nose and speed up....
Join Date: Jan 2005
Location: W of 30W
Posts: 1,916
Likes: 0
Received 0 Likes
on
0 Posts
Originally Posted by Mr Optimistic
Particularly interested in how data from the stuck AoA sensors may have misled systems or denied protection
Where are they on that ?
Probably too busy to extract some specific words from the CVR that will conveniently help to fully blame the crew.
Typical BEA stuff.
Join Date: Jul 2010
Location: New Zealand
Posts: 9
Likes: 0
Received 0 Likes
on
0 Posts
So it was all the German pilots fault.
A single cause airliner accident is rare. Most are the result of the multiple causes.
The following is mainly taken from the BEA report into the loss of D-AXLA.
The traces for the AoA sensors show them working untill approx FL320. The traces then show that they did not moved again. (Page 33).
The latest reports refers to an aircraft wash which suggests that they were contaminated with water and as a result were frozen into the position that they were at that time.
As seen with other accidents, ice evidence can melt after a crash making definitive statements about icing difficult.
The first request for the slow speed check (including an alfa prot check) was made by the Air NZ Captain (who was overseeing the test flight) at approx FL120 which accorded with the advisory paperwork. (page 19) shortly thereafter, the A/C begins a continuous decent taking it out of the advised altitude for the low speed test.
The handling Captain suggested that the low speed check be done later or skipped. This at 5000'. (page 20).
At approximately 4000' the handling pilot asked the Air NZ Captain if he should conduct the slow speed check and this was affirmed. The handling pilot (who was also following ATC instructions and discussing nav with his F/O) eventually initiated the low speed check at about 3000" (page 20).
Due to what? configuration? control law? inoperative sensors? Alfa protection did not initiate. There is no reference in the report as to whether or not the crew had any expectation of this outcome. The aircraft stalled, eventually diving into the sea, which would indicate that they did not.
x x x x x x
From what has now been published the post painting checks for clearance of external sensors were not carried out adequately, eventually leading one way or another to the failure of the AoA sensors to function correctly. This failure played a part in events that has yet to be explained.
Alfa protection did not occur.
Inadequate training? This crew should have been prepared for Alfa floor protection failure and the check captain appears to have been unaware of the altitude requirement.
Inadequate paperwork? The paperwork should clearly identify the risk of Alfa floor failure. The Customer Manual assumption is that the Alfa protection will work.
If the test had been carried out at say FL120 the question remains would the crew have been able to recover the aircraft. What is shown in the traces published on page 34 is the disconnect between the pitch inputs of the handling pilot and the position of the elevators which in the final instance were diametrically opposed to each other. The french have intimated that the programing of the flight control computers is not an issue. Was this the result of the Aoa failure?
A single cause airliner accident is rare. and this isn't one of them.
A single cause airliner accident is rare. Most are the result of the multiple causes.
The following is mainly taken from the BEA report into the loss of D-AXLA.
The traces for the AoA sensors show them working untill approx FL320. The traces then show that they did not moved again. (Page 33).
The latest reports refers to an aircraft wash which suggests that they were contaminated with water and as a result were frozen into the position that they were at that time.
As seen with other accidents, ice evidence can melt after a crash making definitive statements about icing difficult.
The first request for the slow speed check (including an alfa prot check) was made by the Air NZ Captain (who was overseeing the test flight) at approx FL120 which accorded with the advisory paperwork. (page 19) shortly thereafter, the A/C begins a continuous decent taking it out of the advised altitude for the low speed test.
The handling Captain suggested that the low speed check be done later or skipped. This at 5000'. (page 20).
At approximately 4000' the handling pilot asked the Air NZ Captain if he should conduct the slow speed check and this was affirmed. The handling pilot (who was also following ATC instructions and discussing nav with his F/O) eventually initiated the low speed check at about 3000" (page 20).
Due to what? configuration? control law? inoperative sensors? Alfa protection did not initiate. There is no reference in the report as to whether or not the crew had any expectation of this outcome. The aircraft stalled, eventually diving into the sea, which would indicate that they did not.
x x x x x x
From what has now been published the post painting checks for clearance of external sensors were not carried out adequately, eventually leading one way or another to the failure of the AoA sensors to function correctly. This failure played a part in events that has yet to be explained.
Alfa protection did not occur.
Inadequate training? This crew should have been prepared for Alfa floor protection failure and the check captain appears to have been unaware of the altitude requirement.
Inadequate paperwork? The paperwork should clearly identify the risk of Alfa floor failure. The Customer Manual assumption is that the Alfa protection will work.
If the test had been carried out at say FL120 the question remains would the crew have been able to recover the aircraft. What is shown in the traces published on page 34 is the disconnect between the pitch inputs of the handling pilot and the position of the elevators which in the final instance were diametrically opposed to each other. The french have intimated that the programing of the flight control computers is not an issue. Was this the result of the Aoa failure?
A single cause airliner accident is rare. and this isn't one of them.
Scotty Beamup
Was not checking that Alfa Protection occurs at the proper time the purpose of the test? It seems to me that they should have realized that it did not occur and get out of there pronto. 
Alfa protection did not occur.
Inadequate training? This crew should have been prepared for Alfa floor protection
Inadequate training? This crew should have been prepared for Alfa floor protection
Per Ardua ad Astraeus
Join Date: Mar 2000
Location: UK
Posts: 18,579
Likes: 0
Received 0 Likes
on
0 Posts
I only have the preliminary report, but I have to say that whatever contributory factors arose from AoA sensor failure, having recovered the a/c to 2320ft, 138kts and 7 degrees nose up pitch, it would appear to have been a failure to control pitch plus repeated use of aileron to pick up a fully stalled wing which caused their demise.
The HP removed the full nose-down pitch command, applied a hard nose-up demand and used full low-wing aileron at an unrecorded speed but probably below 100kts and possibly below 46kts.
The HP removed the full nose-down pitch command, applied a hard nose-up demand and used full low-wing aileron at an unrecorded speed but probably below 100kts and possibly below 46kts.
Join Date: Jan 2005
Location: W of 30W
Posts: 1,916
Likes: 0
Received 0 Likes
on
0 Posts
Originally Posted by Scotty Beamup
What is shown in the traces published on page 34 is the disconnect between the pitch inputs of the handling pilot and the position of the elevators which in the final instance were diametrically opposed to each other. The french have intimated that the programing of the flight control computers is not an issue. Was this the result of the Aoa failure?
It was more to make sure the Airbus was not overstressed before it hit the water ...
Where is G graph by the way ... ?
Join Date: Jul 2010
Location: New Zealand
Posts: 9
Likes: 0
Received 0 Likes
on
0 Posts
What I thought was, what was the fundamental difference between this crew and the French test pilots who first flew an airbus A320 to Alfa floor.
And the simple answer to that is that the French test pilots approached the initiation of alfa protection with the greatest of caution. Being test pilots they had considered all possible outcomes including the possibility of non activation of that protection.
This crew did not because they are line pilots and their experience, in the sim on approach and inflight demonstrations all had positive outcomes.
The only thing left to alert them to the possibility of failure was the CA Manual but it reads "During deceleration observe - The alfa floor activation".
And so the difference between these two crews was "the expectation of possible outcomes", and what this crew needed was someone or something to alter their "expectation of possible outcomes".
But that did not occur.
And so there was no protection for them, in their training, in the paperwork, and finally in the aircraft systems themselves.
And the simple answer to that is that the French test pilots approached the initiation of alfa protection with the greatest of caution. Being test pilots they had considered all possible outcomes including the possibility of non activation of that protection.
This crew did not because they are line pilots and their experience, in the sim on approach and inflight demonstrations all had positive outcomes.
The only thing left to alert them to the possibility of failure was the CA Manual but it reads "During deceleration observe - The alfa floor activation".
And so the difference between these two crews was "the expectation of possible outcomes", and what this crew needed was someone or something to alter their "expectation of possible outcomes".
But that did not occur.
And so there was no protection for them, in their training, in the paperwork, and finally in the aircraft systems themselves.
Join Date: Jun 2009
Location: Australia
Posts: 49
Likes: 0
Received 0 Likes
on
0 Posts
Except
The pilot was asked to test his alfa floor, then asked if he should "do it now". I have no idea what the indications would be, but surely if your "doing it", you'd be looking for it, even if you wern't a test pilot?
I think BEA's conclusion that this was pure pilot error is indeed oversimplifying but:
Executing manoeuvers to test critical envelope protections at altitudes where failure of these protections leads inevitably to a crash comes pretty close to the defintion of reckless.
Would more altitude have made sure they would not crash in the given scenario:
No
Did the altitude they peformed it in make the crash inevitable:
Likely Yes.
Moreover, the rushed test caused a rather quick approach of Alpha_Floor. This left insufficient time to notice and properly react to Alpha_floor not activating.
During proper test flights approach to Alpha_Floor is done slowly.
Due to the fact that there is quite some margin left to Alpha_Max this normally leaves sufficent time to properly react to such a situation.
This is the reason there are normallly detailed procedures for these kind of tests which reflect the technical dependencies and margins behind which cannot always be derived from pure gutt feeling.
I think in the first place it was indeed important to highlight the piloting aspect of this accident.
This should however not exclude to have a look at the techincal issues in a more detailed analysis
Executing manoeuvers to test critical envelope protections at altitudes where failure of these protections leads inevitably to a crash comes pretty close to the defintion of reckless.
Would more altitude have made sure they would not crash in the given scenario:
No
Did the altitude they peformed it in make the crash inevitable:
Likely Yes.
Moreover, the rushed test caused a rather quick approach of Alpha_Floor. This left insufficient time to notice and properly react to Alpha_floor not activating.
During proper test flights approach to Alpha_Floor is done slowly.
Due to the fact that there is quite some margin left to Alpha_Max this normally leaves sufficent time to properly react to such a situation.
This is the reason there are normallly detailed procedures for these kind of tests which reflect the technical dependencies and margins behind which cannot always be derived from pure gutt feeling.
I think in the first place it was indeed important to highlight the piloting aspect of this accident.
This should however not exclude to have a look at the techincal issues in a more detailed analysis
Per Ardua ad Astraeus
Join Date: Mar 2000
Location: UK
Posts: 18,579
Likes: 0
Received 0 Likes
on
0 Posts
Originally Posted by henra
Would more altitude have made sure they would not crash in the given scenario:
No
Did the altitude they peformed it in make the crash inevitable:
Likely Yes.
No
Did the altitude they peformed it in make the crash inevitable:
Likely Yes.
Yes, the technical factors are important, but:-
a) did the software or some mechanical effect prevent normal stall recovery action?
b) would correct stall recovery have prevented the accident?
I wonder what the 'passengers' were saying during this?
The accident was not 'inevitable' due to the low altitude.
A developed one: Rather not.
Therefore my comment regarding the speed decrease with which they approached Alpha_Floor.
Once they noticed that something went wrong they got into even deeper trouble by the (Mis-) Trim and sudden Thrust induced Pitch up.
Without enough altitude the chances of recovery in that situation were surely slim.
I wouldn't rule out that a very proficient test pilot would have been able to recover from that situation but probably not normal airline pilots. It's simply not what they are trained/supposed to do.
They have been in deepest test pilot country and at an altitude where even those would have never performed these maneuvers.
Result: predictable
Join Date: Jul 2002
Location: UK
Posts: 3,093
Likes: 0
Received 0 Likes
on
0 Posts
BOAC:
a) did the software or some mechanical effect prevent normal stall recovery action?
a) did the software or some mechanical effect prevent normal stall recovery action?
Firstly, given the herculean level of pedantry that AI went to when designing these systems and the amount of ground and air testing they did, and sometimes continue to do on the prototype testbed, and also that the systems were designed to do everything in their power to avoid a stall situation, I'd be shocked if such a fault didn't show itself in testing very early on.
Secondly, the MidiBus has been hauling passengers around for 22 years now. Even if it hadn't shown up in testing, it would likely have shown up in line flying by now - certainly at a previous handover test like the one in this case.
The only automatics-related caveats I can think of relate to trying to force the nose down when TOGA is applied, and I'm pretty sure that the AI personnel present would have caught that kind of thing if it had happened. Also, I'm presuming that the FDR doesn't reflect this situation, as it would have been discussed to death on here by now if it had been.
Per Ardua ad Astraeus
Join Date: Mar 2000
Location: UK
Posts: 18,579
Likes: 0
Received 0 Likes
on
0 Posts
DW - yes, I am with you - the reason I mentioned it is because some seem to be looking at some sort of technical problem causing the accident when, as I think you can see from my posts, I put the major cause at incorrect stall recovery.
Obviously the technical issues must be looked at, of course.
I do not agree with your assertion, however, "that would likely have shown up in line flying by now - certainly at a previous handover test like the one in this case." - bear in mind the AoA sensor was thought to be faulty on this flight - how many other 'test flights' would fit that bill?
Obviously the technical issues must be looked at, of course.
I do not agree with your assertion, however, "that would likely have shown up in line flying by now - certainly at a previous handover test like the one in this case." - bear in mind the AoA sensor was thought to be faulty on this flight - how many other 'test flights' would fit that bill?
Join Date: Jul 2002
Location: UK
Posts: 3,093
Likes: 0
Received 0 Likes
on
0 Posts
I'd hope that it would be caught fairly early on - given that the AoA sensor is (or can be) a pretty important piece of kit in the software decision-making process, I'd have thought that they'd have "failed" instruments on the test rig - one by one and then in every possible combination, to see how the control logic adapted. Again, if the system didn't pick up that the AoA sensor was U/S and adapt it's behaviour to compensate then that could be a potential deal-wrecker... But I'd have thought fairly unlikely nonetheless.