TAM A320 crash at Congonhas, Brazil

[ratchetspanner]

Octavius

Each MEL item is given a category as to the deferral time allowed before engineering need to address and fix the defect. Cat A is usually before next flight but dependant on what is unserviceable, Cat B is 3 day limit, Cat C is 10 day limit and Cat D is 120 day limit (normally cabin defects or non airworthiness items). Hours and cycles are usually used for repeat inspections on damaged panels, dents, rib 5 etc,etc. Hope this answers your question.

[Max Tow]

Soaring the Skies
You misread my post - I gave Warsaw as example of non-TR related overrun.


By the way, can anyone adv if it IS the case that Retard warning can be cancelled when only one TL retaded?

[SoaringTheSkies]

hmm... MaxTow, Warsaw was neither related to TR nor to TL.

But at any rate, four out of 12 hull losses related to TL mishandling would be quite significant. (not sure if I have those numbers right, though)

pj

[Max Tow]

Soaring the Skies
You're absolutely right. It was an overrun but not TR or TL related. However that actually serves to further emphasise the point in that as far as I can see the only TL mishandling related accidents have been when a TR is locked out and these accidents have constitured a significantly high proportion of total A320 accidents.
Yet even given the HF issue of why this defect situation has been on occasion mismanaged in past, I'm still unclear why further action should be needed if A) the airlines adopted the SB and B) operational procedures followed the revised MEL for TR lockout?

P.S. Any opinions from AB drivers on whether in light of the Phoenix accident report, one would be happy to have reversed both (with a poss consequent directional stability implication) under the conditions prevailing at CGH that night?

[armchairpilot94116]

to maxtow

REgarding the cessation of "retard" from the TW final report:


14 5
Appendix 7 Response from Airbus
Company for the Stop of
“Retard”

Q: 2/ Why does the message "Retard" stop 2 s after the nose gear
touched down? Is it as per design?

A: As soon as one TLA is set to REVERSE (whatever the other TLA position), the
internal FWC signal "TLA inhibition" becomes true. If "TLA inhibition" is true,
RETARD is inhibited

(more on the subject from the same report)


To Airbus Company
1. Reviewing the design of stop mode of Retard warning sounds or
accommodating other warning methods to ensure that the warning will
continue before the thrust levers are pulled back to Idle notch after a
touchdown has affirmed.（ASC-ASR-06-03 –006）
4.2 Safety Actions Taken or Being Planned35


According to the Airbus Company
1. In response to: Reviewing the design of stop mode of Retard warning
sounds or accommodating other warning methods to ensure that the
warning will continue before the thrust levers are pulled back to Idle notch
after a touchdown has affirmed.

Airbus Company Response:
Airbus has developed a specific warning when one throttle is set to reverse
while the other is above idle. This warning generates an ECAM warning "ENG
x THR LEVER ABV IDLE", a continuous repetitive chime (CRC), and lights the
red master warning light. This new warning is implemented in the FWC
standard "H2F3".
A Service bulletin will be issued very soon on this subject

[TyroPicard]

I have looked at an internet translation of the VEJA article that marciovp refers to, (thanks marcio) and if I understand it correctly the thrust of the briefing (Brigadier General Kersul, during the reading of the dialogues of the pilots of the A320 of the TAM in the CPI of the Aerial Apagão) is as follows...

There is no evidence of technical malfunction...
An air transportation system cannot ignore the possibility that pilots will make some mistakes...
Aircraft should have warning systems to alert the pilot to his mistakes..
The TAM A320 did not have an efficient warning system for this mistake...
Nor did the runway used offer some safety margin..

Progress.
TP

[Max Tow]

Armchairpilot

Re TPE. Interesting: why on earth design a warning that in this situation ceases when the problem is not resolved...misleading to the point that I'd have thought a fix would be essential. I assume there was some logic in the original design - any ideas?

[AlexL]

Originally Posted by ELAC

Quote:
Am I wrong that there is just no action for the PF to perform wrt the throttle levers in a conventionally designed aircraft?
pj
You seem intent on believing what you want to believe regardless of the facts.
On *most* conventionally designed aircraft you have to pull the thrust levers to idle when you flare, there is no such thing as autothrust in a very large percentage of airplanes. That's the way it's worked on the B727 and every other non-FBW aircraft I've ever flown. Even on most conventional aircraft with autothrust, such as the B757, this is usually disconnected at the time the pilot disconnects the autopilot leaving the thrust levers under manual control. If you don't pull them back yourself when you land they produce whatever thrust the lever demands. This was never pointed out to me in training as a design flaw.
The difference you seem to be getting at is between the Airbus and certain recent Boeings which are landed manually with autothrust active. The Boeing TLs move whereas the Airbus ones do not. There are arguements for either system when looked at holistically, but the Airbus system is not one iota more complicated or different in philosophy in this particular respect than a B757, B727 or a DC-3 for that matter. You want idle thrust, pull the levers to idle. Simple.

your are still missing hugely important diffence.
In the Boeing et al, the thrust matches the lever position - if you do not pull the levers back, then the thrust doesn't come back. Doesn't matter wether you are manul landing or autolanding - the levers will move the same way
In the Airbus the thrust changes without pilot intervention or the levers moving.
For example in this accident.
In the Airbus - during the apporach - the thrus is at a normal approach setting, then during the flare the thrust comes back to idle, then once landed the thrust spooled back up again - what did the pilot do to cause the thrust setting to change twice? Nothing.
In a Boeing / MD etc etc, the ONLY WAY that that sequence of events could have happend is if the pilot pushed the thrust back up after landing.
I.e for the boeing to have thrust spooling back up on the rwy after landing requires a positive pilot input. Wheras the airbus only requires an error of omission, i.e the pilot forgets to do something.
I still maintain that if a simple error of omission has such catastrophic consequences then it is the fault of the man-machine interface.
Just as many people are saying that forgetting to pull back the levers on landing is a completely incomprehensible thing for a pilot to do, then so are aircraft with non-moving thrust levers.
I HAVE to pull my levers back to land - the aircraft won't land without it (well it will, but a long, long, long, long way down the runway - 99.9% of pilots would have gone around by that point on a short runway), then as soon as my wheels hit the ground, my spoiler / autobrake logic is already satisfied.
On the airbus the thrust will idle, and the aircraft land without the thrust levers being pulled back. The aircraft then needs a further pilot action to satisfy the spoiler / autobrake logic. However alien it may seem to some, that action can be, and has been forgotten.
THAT my friend is the crucial and now fatal difference between the two systems

[armchairpilot94116]

to maxtow

I agree with you 100pct. The ASC of TAiwan recommended to Airbus that "RETARD" should not cease until both TL are at idle. But Airbus response was to activate a chime and an ECAM message and to light the Red master warning light. And it wasnt even mandatory or in fact implemented across the entire fleet apparently. Clearly the lessons from the Taipei accident were not learned. Perhaps the Taipei accident was not serious enough to warrant sufficient attention due to no loss of life? Well we certainly have loss of life now dont we!

I think the "RETARD" command should NOT end unless both TL are at idle and only then. This has been mentioned before by others in this thread as well who also saw the logic.

[ELAC]

ELAC" one must consider the remote possibility that throttles don't work properly...either in a cable controlled, or a fly by wire.
if pulling throttles to idle is a natural and normal thing, and it doesn't work, why waste time and try again?
my little checklist is for the case when moving throttles doesn't reduce thrust.
and again, I asked if it would have saved lives in this particular case? for some reason moving the throttles didn't work, or wasn't properly done .
have you ever had a throttle that didn't work? or a fuel cutoff lever that didn't cutoff the fuel? it has happened...and will probably happen again

bomarc,

Your initial question was:

For this unique accident, would the following memory items have saved lives?

My answer was accordingly for the case as we understand it here. There is no indication that either of the thrust levers didn't work and in this case the immediate cause of the problem appears to be that a thrust lever wasn't moved, not a thrust lever that was moved but without effect.

Other circumstances might dictate other actions, but for this unique accident the most effective memory item would have been: "Thrust levers....Idle".

ELAC

[TyroPicard]

AlexL
You don't quite have the A320 A/THR system right.

In the Airbus - during the apporach - the thrus is at a normal approach setting, then during the flare the thrust comes back to idle, then once landed the thrust spooled back up again

On the airbus the thrust will idle, and the aircraft land without the thrust levers being pulled back.

On an Autoland RETARD mode engages at 40 ft. RA and reduces thrust to Idle at the appropriate height. It would remain at Idle even if PF did nothing with the thrust levers. "Retard" callout happens at 10 ft. RA as a command to PF to move the T/Ls.
On a manual landing the A/THR maintains Vapp and if PF does not move the T/Ls the A/THR will increase thrust to maintain Vapp as you trundle down the runway. It does not reduce to Idle. PF retards the T/L to Idle during the flare - the system calls "Retard" at 20 ft. RA as a reminder if PF has not already moved them.

The aircraft then needs a further pilot action to satisfy the spoiler / autobrake logic.

Not true - suggest you read PBL's post No. 585 on page 30.

Cheers TP

[ELAC]

ELAC,
Your post worries me when you say that what we as pilots want is a plane that allows one eng in rev whilst another remains in fwd thrust.

Rananim,

I think your worries might be misplaced.

Admittedly, it would be a very unusual circumstance that would ever see one engine developing forward thrust while the other is engaged in reverse, but in deciding whether or not to interlock the reversers I suspect the designer has to make relative risk assessments of very low probability situations.

Using what we believe to have occured here as a basis for discussion you have to consider the following possibilities:

If there was a reverser interlock, would the failure or restriction of Rev. 1's engagement have triggered recognition with the pilot that TL2 was not retarded? The answer is speculative at best, but as anyone who's spent time in sims knows it can be very hard to shake a mental model once it's been set, especially in a very short time frame under high stress. The pilot response might simply be more confusion "Why is Rev #1 not working ... it's #2 that's locked out and I haven't selected reverse on that one!". The instintive reaction might very well be to pull harder on the TL that you expected to produce reverse, not to pull back the TL from which you were expecting nothing at all.

Also, the case for an interlock in this situation has to be built not just on recognition of the source error, but then on the correct following actions which would vary between different situations. On a short runway the very first response to a lack of Rev function due to an interlock would most likely be firm manual application of the brakes. This follows the general rule of making what you intend to happen, happen by manual intervention if the automated system does not produce the desired result. A subsequent decision to go-around due to recognition that the reason for the interlock preventing reverse thrust actuation is that the other TL is still above idle could be fatal since, if any significant stopping attempt has begun with the brakes or other means, there is no longer any guarantee that the aircraft can accomplish the rejected landing safely.

From an aircraft system designer's perspective I think this situation presents a Hobson's choice: there are a multitude of ways to be wrong, but no clearly definable way to be right. Given this my guess is that the design concept follows providing the best aircraft response for the most likely set of pilot actions in the circumstance. If a pilot retards one thrust lever to idle and then engages reverse without moving the other lever, it is more probable that the pilot intends to land and stop. If on selection of reverse he does not recognize the mis-handling of the other thrust lever his decision to stop using other means (as T/Rs are not required parts of the stopping process) is more likely to remain rather than be reversed. Conversely, if he does recognize the mis-positioned thrust lever his actions could be to either advance both and attempt to get airborne or to retard the mis-placed one and continue to stop. That decision is a 50/50 one that depends on factors beyond the designers control or the system's understanding.

Ultimately then the balance of probabilities is that a pilot having landed and having retarded only one thrust lever to idle and selected reverse on that engine is more likely to continue the landing than attempt to get airborne again. If this is the more probable though not certain case then from a design perspective the correct approach would be to give the pilot what he's commanded, the reverse on the one engine in the knowledge that while the split thrust lever condition is inherently ambiguous and not optimum for either a stop or a go the probabilities lie with the stop and in this case all means available and commanded by the pilot are going to be required.

Going back to Airbus' stopping distance calculations in the TransAsia report it's worth noting that the stopping distance for the condition of one TL in max reverse, one at approach thrust, no spoiler and manual medium braking is well within almost all normal runway lengths (assuming a normal touchdown point) in a dry condition and most in a wet condition if the brakes are applied in a timely fashion. I suspect that as a designer that would be the final tie breaker in terms of what the correct logic for the situation should be as rejected landing performance versus obstacles is completely incalculable in the circumstances.

Looked at another way, imagine if the pilot in this instance had initated immediate manual braking when the lack of deceleration was noted but had still gone off the end at a lower speed with engine #1 producing forward idle thrust instead of the maximum reverse thrust he commanded. Would there not be a howl of outrage focussed on why the autothrust system did not give him the reverse thrust that he commanded? Would we not now be noting that they would have made it if only maximum reverse thrust on at least the one engine was available?

ELAC

[BOAC]

We have some experienced AB heads here - do any of you have any inkling of what is causing this apparent 'mind-set'? Is it that the A/T system, as TP says, WILL reduce the thrust in an Autoland and BECAUSE A/T apparently remains engaged on a manual landing there is some 'vestigial' brain function that says I do not need to? In both cases there is no visible difference on the throttle quadrant. Would the answer be to teach A/T disconnect in the flare on a manual landing thereby reinforcing the 'pilot' inputs? Something needs to be done, and I don't think issuing a reminder to pilots to close the levers will crack the inbuilt block.

Picking up on TP's answer, somewhere back in the thread there was a mention of 1.2EPR on the No2 (Nod's post #1113?). IF this is pukka, is that what you would expect from the A/T as it tries to maintain Vapp with No2 thrust? It seems quite low to me. Is it possible that a partial throttle back from CLB had taken place and would the EPR then be lower?

[Max Tow]

BOAC
Bear in mind the TL accidents plus the mishandling incidents reported by Ppruners have only been on a/c with one TR inop. (unless I'm missing something). So is the "mind set" simply a human instinct to not use the inop TR combined with a "retard" warning on unmod a/c that is worse than useless (literally) because it apparently gives you a thumbs up when you have not resolved the problem, if the Taipei explanation is correct?

[TyroPicard]

do any of you have any inkling of what is causing this apparent 'mind-set'?

Not me! The only common factor is one Reverser Inop, and (I presume) the old MEL advice to use the good engine only in reverse. Here's one thought... Compare this with a OEI approach, where only the live T/L is in CL detent (shut down engine T/L is at Idle) and PF only holds one T/L, only moves one T/L to Idle during the flare - could that be a potential mindset that would cause this problem?

Would the answer be to teach A/T disconnect in the flare on a manual landing

A/THR disconnects by moving T/Ls to Idle - when the TLA (remember the donut?) reaches existing thrust, further rearward movement reduces thrust. I assume you mean use the Disconnect button as well? Not a good idea as someone is bound to push the button before moving the levers and the result is Climb thrust just when you don't want it!

somewhere back in the thread there was a mention of 1.2EPR on the No2 (Nod's post #1113?). IF this is pukka, is that what you would expect from the A/T as it tries to maintain Vapp with No2 thrust? It seems quite low to me. Is it possible that a partial throttle back from CLB had taken place and would the EPR then be lower?

Reading James7 plagiarism I conclude that when A/THR disconnects on this particular landing the thrust is locked by the FADEC on the No.2 engine, the one left in CL detent. Long time since I flew an EPR engine (V2500 shows N1) but IIRC 1.2EPR is somewhere between a reasonable approach value and Climb thrust and would therefore indicate an accelerating engine at the moment thrust is locked.
TP
PS I don't think Autoland experience will produce that mindset - manual landings are far more common .. and the proportion of Autolands in Brazil must be pretty low!

[marciovp]

Newer software...

Se o A-320 da TAM tivesse a nova versão do software, o alerta “Retroceda!” prosseguiria até que ambas as alavancas tivessem sido retornadas para idle. Pela transcrição da gravação da cabine, às 18h48min21s6, o Fadec emitiu o som duas vezes, antes de surgir o barulho do movimento do manete. O avião estava a cerca de 6m de altitude, em pouso. Dois segundos depois, a tripulação na cabine escutou mais uma vez a ordem “Retroceda!”.

If the A-320 from TAM had the new software version, the alert "Retard" would continue until both T/L were on idle. From the transcription of the voice recorder, at 18h48m21s6, the FADEC sounded twice before the noise of the mouviment of the T/L. The plane was at 6 m altitude, landing. Two seconds later, the pilots heard uma more order "Retard".

I found this interesting:
http://catless.ncl.ac.uk/risks/8.49.html#subj3

[sarah737]

Landing in Congonhas on a wet runway you shouldn't hear the retard at all. The retard is a reminder in case the pilot is too late in retarding. The chain of events started right there at 20ft...

[bomarc]

ELAC

from the times wire service (LA times)

"Maia said aircraft equipment failure had not been ruled out, and that it was possible the pilots had shifted the throttle lever correctly but that the plane's computer had failed to respond."

For a moment, let us give the benefit of the doubt to the men and not the machine.

we must have options for this scenario.

I believe my little checklist would have prevented some if not all deaths in this case and that all pilots must have in their little bag of tricks all sorts of options.

[PAXboy]

ELAC, what a sensible and level headed analysis. You have not presumed any fault to machine or person (unlike others) but set out some reasonable grounds for design decisions that will have been considered and reviewed by countless numbers of pilots over the years.

[TyroPicard]

hi bomarc
The A320 has two Angle sensors on each Thrust Lever. If the two sensors disagree about the TLA, on the ground, but not at TOGA or FLX/MCT (i.e. take-off thrust setting), the FADEC selects Idle thrust irrespective of thrust lever position. An ECAM warning occurs and as part of the drill the pilot backs this up by selecting idle.
If neither sensor produces a valid signal, on the ground, the FADEC selects Idle thrust. An ECAM etc......

There has been no mention so far of a related ECAM warning, and I feel sure it would have been mentioned if it had occurred. Both (known) TLA failure modes produce thrust Idle on the ground, and we have heard that 1.2 EPR was recorded on No.2 engine. Obviously some deep technical investigation required.
TP