TAM A320 crash at Congonhas, Brazil

[bsieker]

RTO with 2 RT inop?

atakacs,

the ground spoilers deploy when all conditions are met. Anytime. So, yes, retarding the offending thrust lever belatedly would have deployed them, and thus also activated the automatic brake.

Remember that they are also used in case of a rejected take-off, where the thrust levers are pulled back in the middle of the runway.

Which leads me to a question for the Airbus pilots:

Ground spoilers will deploy (and therefore trigger the autobrake), even if not armed, if the main landing gear is on the ground and at least one reverser is selected (and the other TL at or near idle or in reverse).

What do you do for a rejected take-off with both thrust reversers inop? Do you have to remember to arm the ground spoilers? Or select reverse anyway? Or just use manual braking?


ELAC, I enjoyed your academic and also the speculative discussion of stopping distances very much. I found the results very interesting, (and the distances surprisingly short), even if they don't answer the crucial human-factor questions in this accident.

[Rob21]

IMHO, one thing to consider is why the PNF seems so surprised: "OLha isso!!" (look AT this!!) and calls: "Desacelera!!" (decelerate!!). It looks to me that his (PNF) priority was calling the PF's attention to what makes the aircraft accelerate, or the T/L.
1. Note that the PNF "abandoned" standard calls after the "Spoiller nada!!" (spoiller nothing!!) call.
2. PNF says: "Olha isso!!" (look at THIS) and then says: "Desacelera!!" (decelerate!!).
3. PF says: "Não dá, não dá!!" (can't do, can't do!!)
It seems to me that the crew was trying to deal with some abnormal situation related to the T/L (stuck, bent, loose?) all the time between touchdown and the time manual brakes were applied (11 seconds).
IMHO, if the problem was as simple as a "forgoten" T/L in CLB position, I believe the call after "Spoiller nada!!" (spoiller nothing!!) would be something like "Manete!!" (Throttle!!).
And PF's reply wouldn't be "Oops!!", followed by a movement bringing the T/L back to iddle?
Just a thought...

[Danny]

Depending on some variables such as a/c speed and the amount of water on the runway, without spoilers, there is a very strong chance that the a/c was aquaplaning when brakes were applied. The video evidence shows that there must have been some standing water as the spray pattern from the operating reverser is very noticeable.

If it was aquaplaning, there is very little chance that the brakes would have had any significant effect until the wings were travelling slow enough to stop producing any lift. At 94kts, I would have thought there would still be significant lift and if only the spoilers had been deployed, they may have had a fighting chance of stopping within the boundary of the airfield or at least gone over the edge with less energy.

Can the ground spoilers on the A320 be deployed manually if they didn't do so automatically due to the thrust lever positions?

Edited to add: I realise that had they retarded the #2 thrust lever, system logic would have deployed the ground spoilers and activated the autobrakes. But... as a mere Boeing pilot, I am only used to having a great big handle that can be manually pulled up at any time which will, depending on air/ground logic, deploy the spoilers.

I think that this is where some Airbus and Boeing philosophies differ in that if I were to use the speedbrake (spoilers) when I shouldn't such as when in landing flap configuration whilst still airborne, I will get a warning. From my understanding of the A320, you only get the spoilers when the system wants you to have them. It can't be overridden. Please correct me if I'm wrong.

[ELAC]

I have re-read your excellent post but still fail to get an answer to a question I posted a few hundred messages ago: what would have happened (regarding the spoilers and auto braking) if the crew had retarded the 2nd T/L after touchdown. Would that be sufficient to disambiguate the logic or do you have only "one shot" for the automation to do it's magic ?

atakacs,

Sorry, the answer is exactly as bsieker has stated. At any time if the second thrust lever had been retarded enough to be considered "near idle" (15 degrees or less TLA below 10 feet) the logic for ground spoiler extension would have been made. Autobrakes would then have engaged assuming that manual braking had not already disarmed them.

ELAC

[EMIT]

Rejected T/O

BSIEKER,

In case of any rejected takeoff, you do not need to remember to arm groundspoilers, because arming them is one of the essential actions that has to be performed before the takeoff is started.
Arming spoilers is one of the items on the ECAM takeoff memo ("checklist on the screen" that actually monitors a number of switch actions/positions).

[ELAC]

What do you do for a rejected take-off with both thrust reversers inop? Do you have to remember to arm the ground spoilers? Or select reverse anyway? Or just use manual braking?

bsieker,

The armed condition of the spoilers is shown on the ECAM Takeoff Memo and checking it is a part of every take-off. The line "SPLRS.......Arm" is shown in blue until the PNF arms them at which time the memo becomes green.

Also a part of the memo is the line "T/O Config.....Test". This simulates a takeoff and checks the configuration of required systems, including spoiler position, but not whether the ground spoilers are armed. If all items are correctly configured the line changes to "T/O Config....Normal" in green. If any system is not in the correct configuration the CRC sounds and a Master Warning and ECAM Warning are triggered. However, those warnings won't occur if the ground spoilers are only not armed as opposed to not retracted. Having never tried this I don't know whether the line "T/O Config....Test" will change to "Normal" if the test is carried out with the ground spoilers not armed. The FCOM is silent on the matter, so it might be possible.

Once the before takeoff flow has been completed the Before Takeoff Checklist is read which includes the item "ECAM Memo.....No Blue", so any ommision to arm the ground spoilers should again be noticed at this time.

However, supposing it wasn't and the crew commenced a take-off after ignoring both to carry out the before take-off flow and then the check list (or intentionally decided to take-off without the ground spoilers armed), then for the ground spoilers to extend the speed would have to above 72 knots and at least one reverser would have to be in at least idle reverse. In this case a decision to not select reverse on either engine would result in the ground spoilers not extending to the best of my reading of the FCOM.

In practice I think in an actual reject, regardless of both T/R's being inoperative, most pilots would reflexively pull both thrust levers to reverse idle before they had the time to think about them both being inoperative, so likely they would still get ground spoiler extension even despite all of the missed items above. However, as you point out that is not following the procedure as written in the MEL page TripleBravo posted. So, for the TAM A320 at least (there is no similar procedure in the A330 MEL I have) it seems it is possible to do as you suggest, and one should be aware of it if operating an aircraft with both T/R's deactivated. Perhaps a revision to the MEL procedure might be appropriate?

ELAC

[ELAC]

Depending on some variables such as a/c speed and the amount of water on the runway, without spoilers, there is a very strong chance that the a/c was aquaplaning when brakes were applied. The video evidence shows that there must have been some standing water as the spray pattern from the operating reverser is very noticeable.

If it was aquaplaning, there is very little chance that the brakes would have had any significant effect until the wings were travelling slow enough to stop producing any lift. At 94kts, I would have thought there would still be significant lift and if only the spoilers had been deployed, they may have had a fighting chance of stopping within the boundary of the airfield or at least gone over the edge with less energy.

Can the ground spoilers on the A320 be deployed manually if they didn't do so automatically due to the thrust lever positions?

Danny,

It may be possible that aquaplaning contributed to the aircraft's speed off the runway, but, based on the AIT posted here http://www.pprune.org/forums/showthread.php?p=3457705#post3457705 and some preliminary FDR info posted here http://www.pprune.org/forums/showthread.php?p=3463553#post3463553, there has been nothing published that strongly suggests aquaplaning did occur.

From the above sources touchdown was in the touchdown zone at 140 knots and the aircraft travelled for approx 11 seconds at about that speed before maximum braking was applied. This suggests that deceleration started about 1000-1200m down the runway leaving at most 880m left to the end.

The AIT says that the aircraft over-ran the runway (in fact we know it left the side somewhere before the end) at approximately 100 knots. So, the speed decreased by about 40 knots within a maximum distance of perhaps 880m. As the kinetic energy of the aircraft varies in proportion to the square of its velocity the ratio of the aircraft's energy off the end of the runway to its energy at the start of maximum braking looks to be about 100²/140² = 51%.

That would suggest about a 49% decrease in aircraft energy achieved over at most 880m by nothing but the brakes and the one reverser while fighting against the other engine at 1.20 EPR and a wet runway. My guess is that this would be a pretty good result all things considered. In my guesstimate above the minimum ground stopping distance would have been 1520m (1820-300). If the stopping force available from the brakes remained fairly constant (and I don't know that to be true - if anything I'd expect it to rise due to greater weight on the wheels) by guesstimation it would have required at least 750m to reduce from 140 kts to 100 kts. So, if the aircraft actually achieved something within a 100m or so of that value then it's unlikely that significant aquaplaning occurred.

With respect to manual extension of the ground spoilers on the A320 the answer is no. There was some discussion of this a ways back in the thread.

ELAC

[ELAC]

I friend of mine told TAM'S A320 had msn 791
I friend of mine also told that Transasia's A320 had msn 789

Then I would ask another friend of mine if the following about TL fault

"Affected engine must be shut down, because when thrust levers are retarded, autothrust would remain in SPEED mode. During flare the thrust of the affected engine would increase to maintain the autothrust target speed."

was valid and if affirmative, any warning was then predicted.
Also, if affirmative, I would ask if such a behavior could be changed with later updates.

flyingnewbie,

If a TL fault had occured somehow causing the autothrust to remain in SPEED mode, then as the aircraft decelerated when the brakes were applied the thrust on the #2 engine would have increased in an attempt to maintain the target speed. If that had occurred I suspect that either the Airbus AIT or the FDR info that has made it's way into the public realm would have provided strong evidence of this.

ELAC

[Dream Land]

If it was aquaplaning, there is very little chance that the brakes would have had any significant effect until the wings were travelling slow enough to stop producing any lift. By Danny

Yes, and I would hope that all pilots operating high performance jets would understand this, apparently not.

[duwde]

I don't know if someone has already posted this, but...

Here's the FDR Graphs just released...
http://download.uol.com.br/fernandor...r_graficos.pdf

[BOAC]

Originally Posted by Dream Land

Yes, and I would hope that all pilots operating high performance jets would understand this, apparently not.

- indeed, and what I have been banging on about now since the beginning of this thread, that it is only the reversers, a small amount of airframe drag, minimal 'rolling resistance' and time/distance that will slow you down then until you break the aquaplaning.

[bsieker]

EMIT, ELAC,

thanks a lot for clarifying the "ground spoilers on RTO" issue, I wasn't aware that ground spoilers were armed for take-off.

Originally Posted by ELAC

In this case a decision to not select reverse on either engine would result in the ground spoilers not extending to the best of my reading of the FCOM.

Yes, that was my understanding, too.


Thanks,
Bernd

[teropa]

That graph is really disturbing reading.

Seems that just as ENG1 idle rev was selected, ENG2 accelerated from approach ~1.09 EPR to ~1.28 EPR and then stabilized to around ~1.19 EPR.
ENG2 TLA indicated CLB for the entire period.

Interesting to note that full manual braking action (disregarding the most likely accidential and small applications of brakes during the initial landing roll) was initiated at 18:48:34, just 7 seconds after nosewheel touchdown was registered.

I am starting to think that this was not a pilot error. Why would ENG2 accelerate, even though it was left to CLB. Shouldn't the Thrust Lock have just locked it to ~1.09EPR. Now instead, it accelerated to nearly ~1.3 EPR which is approximately what you would see as CLB power in those circumstances. No wonder they couldn't stop.

Obviously huge amounts of right rudder were also applied to overcome the thrust asymmetry. Maybe the ENG2 didn't respond to TL positioning? Obviously their previous landing was made also with the other TR inop, and they had no problems.

What is also disturbing (or the FDR is not telling the truth due to some underlying problem in the TLAs?) to see is that in their 2nd previous landing they only set the TL2 to IDLE (while reversing TL1 as should). But in the previous landing (ie just before the accident) they reversed both TLs as one would with both TRs operating?

Was there a reason they changed their way of handling the TL2 during the day? First they would pull it to IDLE, then they would pull it all the way to REV, and then finally they left it at CLB (that's what the FDR claims at least).

Tero

[PBL]

I don't know if someone has already posted this, but... Here's the FDR Graphs just released...

Good. So the first thing to point out is about the "delay" concerning manual braking, that was in the Airbus Telex.

Spoilers partially extend with WoW on one main, and reverse one one engine (provided the other engine is at idle, or also in reverse). Full extension is with WoW on both mains (FCOM 1.27.10).

Please note that WoW, brake pedal deflection, TLA and so on are depicted in the transcript as being sampled at one-second intervals.

Wow on L was registered at 18:48:25, WoW on both (with WoW on R) at 18:48:25 (FDR p2). "No spoilers" was called at 18:48:29+, five seconds after theoretical begin of spoiler deployment. Manual braking from nothing to max (80 degrees deflection) took place over the two seconds from 18:48:34-36.

There are less than 5 seconds from "no spoilers" call to activation of manual braking. Given 1 second (minimum!) reaction time, that is at most a 3-4 second "delay". Similarly, one could argue that it would be unreasonable to expect a "no spoilers" call to happen earlier than 18:48:27. That would allow 2 seconds after when they should be fully activated for the two actions: judge they are not, and call it.

So one could argue that the "no spoilers" call took 2 seconds longer than optimal, and the manual braking 3-4 seconds longer than optimal. That is 5-6 seconds longer than optimal, and paints a different picture from saying "manual braking was delayed 11 seconds".

That is my first point. Here is my second.

Let us compare optimal autobrake-with-no-braking against manual braking.

I have argued that optimal in this situation ("optimal" here means something like: perfect behavior by perfectly programmed robots) would have been a "no spoilers" call at 18:47:27, and, as reaction to this, manual braking to full starting at 18:47:28.

Had one been landing with manual braking, then manual brakes would have been activated as a reaction to WoW on both mains, which was at 18:48:25. Allowing a reaction time of 1 second, this means manual braking would have been activated over the 2-second time frame 18:48:26-28.

But we are not perfectly-programmed perfect robots. A human-factors issue arises. One does not expect "no spoilers", so the PM waits an extra second to allow him to believe his eyes that the spoilers haven't deployed. So a call at 18:48:28 (now the call at 18:48:29 is not looking so slow). And the PF reacting to this over an unexpected-event reaction time, which is understood by many engineering psychologists to be between 1 and 2 seconds, and we have application of manual braking at 18:48:30-32.

That is four seconds difference between the situation in which one is landing with manual braking, and the situation in which one is landing with autobrake and one must apply manual braking. 4 seconds at 70 m/s is 280 m.

Hydroplaning or no, thrust levers to idle or no, perfect reactions or no, this 280 m represents what they lost simply through the choice of autobrake over manual braking, which they made in the air when everything was OK.

I conclude that 300 m is the price you nominally pay in an abnormal braking situation for having chosen autobrake over manual braking. Food for thought.

PBL

[NigelOnDraft]

Posting those graphs really was goiong to bring out 1000s of Amateur accident investigators (of which I would be one if I decided to read too much into them... )

Seems that just as ENG1 idle rev was selected, ENG2 accelerated from approach ~1.09 EPR to ~1.28 EPR and then stabilized to around ~1.19 EPR.

Eng 2 appears to increase power as Eng 1 reduces power from the TL movement... which is entirely logical while ATHR is engaged to maintain an IAS etc. It is clear from the Graphs when ATHR was active etc.

Why would ENG2 accelerate, even though it was left to CLB. Shouldn't the Thrust Lock have just locked it to ~1.09EPR.

Why should it maintain 1.09 when ATHR was engaged and Eng1 EPR reducing and IAS reducing (slightly)... seems to be to be working as advertised NB the ATHR system sees one engine as essentially "failing" so applying power to the other one to maintain speed is surely what you would hope for?
NoD

[teropa]

Good points, NoD, and I agree. That's what should happen in a normal case.

Not pretending to be an investigator. Just an interested person, who has eyeballs, Mark I.

Why does the ATHR disengage only at 3 seconds after TR1 selected in MAX REV (2 seconds after TR1 fully deployed)? Is this normal? So it doesn't disengage at the selection of IDLE after all?

Tero

[PBL]

Teropa notes

Originally Posted by Teropa

What is also disturbing ....... to see is that in their 2nd previous landing they only set the TL2 to IDLE (while reversing TL1 as should). But in the previous landing (ie just before the accident) they reversed both TLs as one would with both TRs operating?

I don't know whether we know yet that the same crew made the 2nd-previous landing at Congonhas. Can someone say? We do know that the same crew made the penultimate landing in Porto Alegre, in which, as you note, both TLs were reduced to idle as per SOPs.

This shows, despite all the back-and-forth here about ergonomic thrust-lever design, that this crew knew how to handle thrust on landing according to the 1-TR-INOP SOPs. Indeed, I can't understand why any pilot would ever have doubted this. But it is nice to have it in black and white (well, black and red, actually).

PBL

[borghha]

Good question teropa, I was wondering too why the AT came off that late.

And this struck me too: during the 2nd previous landing (other crew) both TL were set to idle but REV was only selected on engine 1. How come SOP was not implemented there either?

[PBL]

borghha wrote

Originally Posted by borghha

during the 2nd previous landing ....... both TL were set to idle but REV was only selected on engine 1. How come SOP was not implemented there either?

Do we know that that was not TAM 1-TR-INOP SOP at the time?

PBL

[Max Tow]

From the superimposed CVR/FDR, the reverse number one only comment rather looks like an instruction. But as PBL remarks, the same speaker had apparently been handling pilot at POA and used both. The previous landing at CGH used just one. A most odd sequence, unless a deliberate departure from SOP at CGH due to concerns about a PHX-like handling effect on the known slippery CGH r/way.