|
I don't say the crew is entirely not to blame, but they it seems as the automation failed first and foremost. They managed to discover just such a glitch, but because they were not prepared for system failure as they should have been, the discovery struck them when they were poorly positioned to recover. It seems to me that the airline should have deployed proper test pilots, or at least one pilot who could be trusted to follow the manufacturer's SOPs. |
As far as I understand this is the main cause, pilot errors came only after that The main cause was the crew attempting a stall manoever at low level and before that attempting a test flight without the proper precautions and briefings. The sensor problem caught them with their trousers down, anything can be expected to fail on a test flight and you must assume that it will go wrong. That is what tests are about - to identify any problems. Why do you think professional test pilots are a special breed. |
What kind of Test
There are not many things that you do to an Aircraft during maintenance that require an air test. In this particular case was there anything that was being tested during the accident event that was mandated in the maintenance manuals? Or was this a test 'created' to act as somekind of handover from one operator to another? When you buy a second hand car you will often take the car for a test ride. But what are you testing? Is there a parallel between this and handing back an Aircraft. There are no legal requirements (are there?) for transferring an a/c nor maintenance requirements (usually) so what tests will make the new operator 'happy'. Was this whole event avoidable if the transfer took place in the hangar with a good old kick of the tyres?
|
Or was this a test 'created' to act as somekind of handover from one operator to another? The aircraft was about to be returned to ANZ after a lease. ANZ specified the same handover procedure as Airbus use when delivering a new aircraft. Instead of asking Airbus to do the job properly the operator obtained a(partial?) unofficial copy of the test schedule and tried to DIY. Read the official report for the sorry tale of misunderstanding, old information, inadequate briefing and general foolishness. |
Yes - but were they legally required to perform that particular test?
|
Yes - but were they legally required to perform that particular test? There is probably no specific airworthiness requirement to conduct that specific test at the end of a lease (but ICBW). If the lease contract specifies a handover procedure then that part of the contract is legally enforceable in the same way as any other contract. If you borrow my aircraft you do so on my terms and I am NOT about to accept it back with just a walkaround and a kick of the tyres. |
Bizdev
"legally required...." It may be splitting hairs, but nothing of this accident speaks of any crime, per se. All parties were attempting to perform duties agreed to in a contract. No contract would require any action that is illegal, or it would not be a contract. The mass of legal issues may be extensive, but almost certainly criminality was not an issue. |
Main cause
The main cause was the crew attempting a stall manoever at low level The example FDs in the report are very enlightening, whilst they initially had a warning of the need to use manual trim, the warning vanished due to the Airbus changing Flight Laws. To my mind the causes were: 1) Washing it in an unapproved method - missing covers and with a fire hose. 2) Failure to follow the SOP for the manoeuvre - crew didn't calculate speeds that the protection should activate at and aborted when it didn't activate. 3) Failure to manage the trim in the stall recovery - crew didn't check it during the stall recovery, and whilst the FD had warned earlier that manual trim was required it vanished from the FD at the time of greatest need. Yes there was other naughtiness going on about flight testing without officially informing ATC and doing the test at the wrong level, but I don't think those actively contributed to the accident. |
The crew was not supposed to believe that the automation would operate correctly - the purpose of the flight was to detect any maintenance-related glitches, including those that affected automation. Furthermore, if what you are saying is true, the pilots could very well expect the elevators, flaps, rudder or any other part of the plane to fail - who in their right mind would fly such a plane!? or at least one pilot who could be trusted to follow the manufacturer's SOPs. Think again. The main cause was the crew attempting a stall manoever at low level and before that attempting a test flight without the proper precautions and briefings. A very important factor was also that there was no sufficiant warning for the crew about this failure. I therefore conclude that crew errors only contributed to the fact that the aircraft was not performing "as advertised", and that most importantly the automatic switching of flight laws prohibited the auto-trim system to move from the nose up position - and this caused the second, fatal stall. |
Their cavalier attitude towards ATC, ie. speed control; lack of informational communication with ATC .... and lack of prudent speed control, dependency (maybe) on automatic nose down trim during attempted recovery, which subconsciously perhaps the pilot at the controls was used to aircraft doing (I am unfamiliar with A320's, really all Airbi, and have no desire for that to change). Have no idea if they briefed expected flight control logic/action and what to expect prior to initiating the event.
|
Mistake after mistake after mistake.
They were bent over the barrel with their underpants around their ankles long before the sensors sneaked up behind them. The report describes a comedy of errors which should never have even started. Leave test flying to test pilots, they are trained and paid to expect and understand the unexpected. Starting a stall manoever at low altitude was only the last piece of idiocy in this tale of foolishness. It does not matter how foolproof you make the machine, the world will breed a better idiot to outwit you. |
Firstly, AOA sensors were not serviced, so even if the crew was looking for a glitch caused by maintenance they wouldn't expect it with AOA sensors. the fact is the first point of failure were the AOA sensors due to inproper maintenance/handling/washing. Further, I assume that if an aircraft is washed or painted, there is always a possibility of sensors being left inoperative or misreading? Presumably a main aim of a 'test' flight is to check for such glitches? |
It seems to me that these two statements contradict each other. Why would they suspect any problems with AOA sensors if there was no maintenance done to them? I assume that if an aircraft is washed or painted, there is always a possibility of sensors being left inoperative or misreading? Starting a stall manoever at low altitude |
They didn't start a stall manouver, they wanted to demonstrate speed/stall protection. Each item on the test schedule must be tested to see if it works CORRECTLY OR NOT. Nothing can be assumed. This is a TEST not a demonstration or a routine ticking of boxes. A competant test pilot assumes that each and every test could fail and has prepared for all eventualities including the worst case. Starting the stall test at low altitude was a stupid error made by a crew who were not competant test pilots and should never have attemped to fly a test flight. Scrounging an unofficial copy of the manufacturer's test schedule does not make you competant to conduct a test flight. |
You really cannot go fly "presuming" all kinds of glitches. |
If the point of the "demonstration" was to ensure that a crash as a result of a failure would occur before passengers were embarked, I guess the flight was successful. Unless that was the goal, though, the crew should not have started with the assumption that the system being demonstrated would work.
|
Hi,
They were bent over the barrel with their underpants around their ankles long before the sensors sneaked up behind them. The report describes a comedy of errors which should never have even started. Leave test flying to test pilots, they are trained and paid to expect and understand the unexpected. Starting a stall manoever at low altitude was only the last piece of idiocy in this tale of foolishness. It does not matter how foolproof you make the machine, the world will breed a better idiot to outwit you. Experience of the CBD: - 12 709 flying hours, including 7,038 on type. 128 hours in the last three months, all on type. 14 hours within thirty days, all on type. Any flight time in the past 24 hours.? - Service time: End of the last service before the accident flight: 2 November 2008 21 h 24. • Start of service on the day of the accident: 4 h 30. • Resting time: 13:30 min. Experience of co-pilot: - 11 660 flying hours, including 5,529 on type. / 192 hours in the last three months, all on type. / 18 hours within thirty days, all on type. / No hours flying time in the past 24 hours. - Service time: End of the last service before the accident flight: November 4, 2008 at 20 h 52. / Start the service on the day of the accident: 4 h 30. / Rest time: 120 h. From a french page: While the low speed tests were carried out in haste, it is also true that the crew was not alerted that two probes were blocked effect. This is a serious anomaly. Indeed, if the pilots had been warned that two probes were blocked effect, they never sought to test the proper operation of protection systems in which these probes are the main element. The certification document SC 25 requires in its Subsection 1309 (c) that pilots are informed of any failure so they can take appropriate action. CS 25.1309 (c) Information concerning unsafe conditions Operating System Must Be Provided To The Theme to enable crew to take corrective action appropriée. A warning indication Must Be Provided if immediate corrective action IS required. Systems and Control, Including indications and Annunciation Must Be Designed to minimize crew errors, Which Could create Additional hazards The pilots did not have a visual alarm and / or sound informing them of the unavailability of two probes impact which is contrary to the standards required for certification. The BEA does not say in his report. Page 16 of its report, the BEA dares say: "Between 15 and 15 h 04 h 06, the probes of incidence 1 and 2 are blocked and remain frozen until the end of the flight to values affect local quasi- identical and consistent with values of incidence of cruising without the crew perceive it. " It should of course read "without the crew being informed." It is intellectually dishonest. The blocking of 2 probes incident resulted in the miscalculation of the characteristic speeds (and Vαprot Vαmax), rejection of ADR 3, the direct passage into law and made inoperative automatic compensation. This succession of automation in a very complex system has not been seen by the crew. The pilots did not understand what was happening. (Remember Habsheim, Bangalore, Ste Odile, Congonhas ..., the pilots did not understand what was happening.) If the BEA agrees that "the real situation of the global aircraft was not known to the crew," he does not question the technology as the Airbus CENIP Brazil has in its report on the accident an A 320 of the company TAM July 17, 2007: "The Automation Of The aircraft, however complex, capable of Providing Was Not The Pilot With sufficiently clear and Accurate stimuli, To The Point of favoring Their Understanding of What Was Happening In The moments just the after the landing in Congonhas " The BEA also forgot to mention that feedback has still failed because two similar events took place in a French company before the crash of Perpignan. If all players in the accident on 27 November 2008 had benefited from these lessons learned, good awareness and good safety, it would have saved seven people dead and a plane. |
Forgive me for jumping in, but there's been so much BS on this thread.
First, in doing test flights whether they are experimental test flights, maintenance check flights or research flights the mantra is "Plan the flight and fly the plan." I've been engaged in flight test for thirty-plus years and may have been tempted to deviate from the flight card (or told to), I've always resisted, which many explain why I'm still around. I'm not any better than the others, just more methodical (and more wary of what can go wrong). I can not imagine doing this test essentially in the traffic pattern. If you missed the point at a safe altitude, you need to go back to the safe altitude and conduct the test. And don't do things on the fly (no pun intended.) If you're doing a test such as this, where the alpha-protection logic should engage at XXX knots, and it doesn't by XXX-2 knots, I'd take the airplane back and figure out what happened on the ground. I've seen some negative comments about voting out one alpha probe, but then allowing it to trigger stall warning. Isn't this the conservative thing to do? I don't have any problem with this at all. As far as decluttering the mode annunciations during the recovery, I feel partly guilty. In some of my studies, I recommended doing just this to clean up the PFD when a unusual attitude recovery was to be made. After the A-330 accident during flight test in Toulouse, I realized the error and no longer recommend decluttering essential information. This means, however, we need to move this off the PFD and have the annunciations adjacent to but not on the PFD. Sorry about the rant. Goldfish |
Goldfish,
Your post is justified. Just one thing : To out vote one alpha probe or one ADR is not an issue, but to not tell IS. |
Hi,
This is interesting .... Crash Perpignan yet the failure of feedback! (Page 84 of the BEA report). "An event on the A320 has been reported by an airline after the publication of the interim report." In fact, there were 2. The BEA does not give the dates of these events (or the name of company), but does not specify that they occurred after the accident on November 27, 2008. We have reason to think that these precursors two serious incidents occurred before the crash of a French company in Perpignan. If all players in the accident on 27 November 2008 had benefited from these lessons learned, good awareness and good safety, it would have saved seven people dead and a plane. Excerpts from the BEA report: "The circumstances of this event were determined from maintenance records, the minutes of the aircraft equipment and the testimony of the copilot. The test probe 1 indicated a blocked probe to 5 degrees. Blockage of the probe 3 was found in the manual verification of freedom of movement. The record review indicates that icing residues and muddy water has probably led to the blocking probes on the flight. During this flight, the alpha floor function did not activate. The PF found that the speed was less than ten knots in the V max expected alpha. The crew felt the aircraft sink and the captain decided to discontinue the investigation. The PF made a maneuver similar to a recovery stall. There was no stall warning. On another A320 from the same operator, abnormal values of incidence were also found during a check flight. The three probes have been examined in the workshop, which revealed the presence of fluid in the bodies of two of them. " These two incidents are precursors of the crash events in Perpignan. Les dossiers noirs du transport aérien |
Goldfish
Thanks for your expert comments. How does an airline pilot become competant to conduct tests such as these ? I know that there are a few excellent schools such as ETPS Empire Test Pilots' School but they cannot produce enough qualified test pilots for all of the routine airline test work. Is it just a case of senior captains being given the task and surviving or is there some formal best practice that should have been followed. These poor guys were thrown into the deep end of the pond and clearly did not have the experience or knowledge to cope when it went wrong. Surely there should be rules somewhere to prevent good pilots being suckered into this kind of mess. Did you just drift into the job and learn by experience or were your formally trained before going into your testing career ? |
How does an airline pilot become competant to conduct tests such as these ? Surely there should be rules somewhere to prevent good pilots being suckered into this kind of mess. Did you just drift into the job and learn by experience or were your formally trained before going into your testing career ? NoD |
HN39,
An engineer will correct me if I'm wrong but this is what I think : As soon as the aircraft is back on the ground, the PFR (Post Flight Report) will tell that one ADR has been rejected and the report should mention the reason why that ADR has been rejected : In this case one AoA sensor was disagreeing with the 2 others. (A normal operation flight would have also triggered the rejection of one ADR) Now, it depends how far the engineer wants to investigate : He can test the system, see no fault and sign the aircraft ready for flight, or it must be possible (?) for him to extract the data of AoA sensors 1 and 2 (3 not being recorded) and analyze those data in more details ... |
That's how the engineers put it all those years: It's the pilots' fault if they don't recover an airplane malfunctioning. Easy!
|
CONFiture;
I deleted my post after noticing that the matter had been discussed by NoD one page back. Thanks for replying. heavy.airbourne; Just another bit of info from the CAA site that speaks for itself: Check Flight Schedule 246 for A319/A320/A321 - High Incidence Protection - Recommended FL130; Limits 10000ft AGL to FL140. - Clearance should be obtained to operate in an altitude block of about 5000ft. regards, HN39 |
FMA Calls?
The CVR transcript does not contain any FMA calls - presumably this was SOP in GXL.
When they put the a/c into Alternate Law there was no formal reading of ECAM message. When USE MANUAL PITCH TRIM appeared on the PFD's, nobody read it out. And nobody read ECAM, which would have displayed a DIRECT LAW caution. Do you see a connection between non-use of FMA and ECAM discipline as recommended by Airbus, and their failure to recover ? |
The read to me indicated a dire lack of professionalism, and I presume an ego driven can-do attitude on the part of the Captain. After all, he was a conveyor of type ratings on the aircraft to presumably lesser skilled (on this aircraft type) aviators than he. His apparent qualifications of flight simulator instruction and thus far, far superior system knowledge than your average lowly line pilot would certainly imbue him with astounding flying skills and decision making ability.
|
How much of a factor would fatigue play in this? I was struck by the fact that the German pilot and co-pilot left Frankfurt at 4am. Even if the flight had taken place at 12.30 as intended, that would still have been over eight hours since making an early start. By the time of the flight, they must surely have been more than a little tired. I'm surprised they weren't put up overnight in Perpignan before the flight. Might a bit of get-home-itis have set in?
|
CAA Flight Check Handbook, Section 3 page 29 onwards covers most if not all that should have been considered before and during the flight. Particularly para 4 Flight Safety; and Tech 2, para 7 Stalling and Slow Speed Checks.
A sad epitaph after the fact. |
Bearfoil #1380
"Legally required" was a poor choice of words on my part - what I am getting at is that I doubt whether many, if indeed any, of the tests performed on this flight were required/mandated by the maintenance manual (following maintenance) or any National Airworthiness Authority (NAA) legislation/rules when transferring the aircraft to another operator.
|
One might be easily impressed by the thousands of hours of both pilots, and easily believe that they should be amongst the most quailified to perfom those "test flights" (well, I would prefer the word "acceptance flights") Such an assertion would forget too easily that a significant percentage of that flight time (!) has been spent on autopilot, or reading magazines in the cockpit, or even sleeping, not mentioning going to the toilet or chatting with the flight attendants in the galley... and there is also the time spent on the ground.
So all those impressive flight hours dont't really mean a lot, as far as acceptance flights are concerned. Now as rightly said by Goldfish85 and Ancient Geek, there are specialised crew for those duties - even if acceptance flights are more than ususally called "test flights" - and even the recognised test pilot schools like ETPS or EPNER have dedicated courses for acceptance pilots. They usually require 6 months, against 12 months for the full course. Also the big manufacturers (Airbus at least) have in-course programs, to allow selected pilots with recognised experience, to fulfill some of those flight duties after being issued with an appropriate qualification. In any case, there will be a test card, with points, and for each of them relevant safety parameters, altitude being of paramount importance. And for stall approaches, there is a big idea : incremental, incremental.... |
NZ opinion....
From the NZ Dominion post, the NZ ALPA opinion on this.
Letter: The pilots weren't to blame - union | Stuff.co.nz |
From the NZALPA letter in the above post ... This failure was beyond the pilots' control and knowledge; additionally, it's not clear whether the accident wouldn't still have occurred had the flight manoeuvre been carried out at 14,000 feet instead of 3000ft. mm43 |
But the specification that the the alpha protection test should be carried out at FL140 and not less than 10000 feet AGL should have been a big clue that failure was possible and recovery would take a lot of altitude.
Ignore the instructions at your peril. It is entirely possible to argue that the crew were suckered into conducting a test that they were not qualified to conduct but at the end of the day the buck stops at the left seat. The pilot error in this case was accepting the task. Hubris - maybe. Management pressure - maybe. The buck still stops at the left seat. |
Originally Posted by The Ancient Geek
The pilot error in this case was accepting the task.
Hubris - maybe. Management pressure - maybe. The buck still stops at the left seat. regards, HN39 |
Originally Posted by HazelNuts39
(Post 5967702)
The transfer of aircraft from one operator to another happens all the time. In the absence of a proper regulatory framework for these flights, I wonder how many of the pilots doing the acceptance flights are 'adequately prepared' for that task. Perhaps the regulators are the real culprits, not the individuals that died in this accident.
But even if those regulations or similar had covered this flight, would it have made a difference ? I am not sure - the flight was conducted based on a document obtained from the manufacturer unofficially / illicitly, and specifically not approved by them for this purpose. The crew then proceeded to completely ignore that document as regards the alpha-protection test. Would they not have just ignored regulations too in order to squeeze the test in ? |
I think it is well accepted by all that the test was largely unprepared and also performed at an absolute inappropriate altitude.
Now, can we look further and analyze how the system did react to the erroneous AoA information ? The BEA report is very poor in that perspective, it is evasive not to say secretive. It seems the load factor played a crucial role in the different phases after the initial stall warning … Where is the load factor graph ? The ELAC (Elevator Aileron Computer) needs AoA information in order to ensure the control of the elevator, the horizontal stabilizer and the ailerons, but, as we know it, that AoA information was erroneous, but considered as valid by the system ... What has been the consequence on the ELAC response to the pilot inputs ? After the initial stall warning, the Capt side stick has been mainly forward for the first 10 seconds but the elevators barely reached the neutral position due to (according to the BEA) the load factor. Where is that load factor graph ??? |
The real question is what was this flight for? Transfer from one airline to another after mayor maintenance, in other words to make sure everything was working.
They must be ready for something NOT working or a failure. They were not prepared, qualified or ready for it; that is the bottom line. Yes they had a failure, but that is exactly what they were looking for. They were not able to identify it. During this test (and yes, it is a test); the Flight Test Engineer calculates a minimum speed that properly trained and qualified Pilots would NEVER go below because the aircraft will be outside the protection envelope and it will stall. They put themselves in that situation, should not do that flight without proper training. Learning experience for the rest of us. Harsh, but true. They paid the ultimate price for their mistake, God bless them and their families. G |
"They paid the ultimate price for their mistake, God bless them and their families."
Yes, the pilots paid the ultimate price for a perilous situation the Captain placed himself and the aircraft into. The poor trusting souls in the cabin, their lives snatched away from them. |
Hi,
Ok .. the test was performed at a low altitude and this is a error of the pilot(s) But .. you can't discard this from the scenario: First, a critical failure occurred and this failure was not enunciated to the pilots. Second, the aircraft automation safety system failed because of that critical failure. This is a black point and must be fixed |
| All times are GMT. The time now is 19:20. |
Copyright © 2026 MH Sub I, LLC dba Internet Brands. All rights reserved. Use of this site indicates your consent to the Terms of Use.