TAM A320 crash at Congonhas, Brazil

[DingerX]

I guess there's not much point in indicating:

But ... what you have just confirmed is that: if a TL is left advanced on a 777, you will not get TR or GS? That sounds to me (as an outsider) exactly the same as the Airbus under discussion?

You know, this thread is getting somewhat long, and the more informed have probably left about 600 posts ago, but it doesn't hurt to reread what has already been posted.

A. The AB logic is only for spoilers, not for reversers. So the argument, if correct, is that the following input:

One TL not at idle, TR commanded on other, GS Armed

Will result:
320: Forward thrust on one engine, Reverse on other, no GS, no autobrakes

777: Forward thrust on one engine, no TR, no GS, no autobrakes.

Initial reports had the aircraft attempting a Go Around. The consensus was that fundamental airmanship (equal to "no landing can occur without TLs set to idle") dictates that, once the reversers are out, a Go-Around is no longer an option.

Talk all you want about design and what should have been done, but if these facts are true, than the 320 has a hole in the logic: there is a condition where the aircraft will allow one command that results in the irrevocable decision to stop while rejecting other (revocable) commands to stop.
Sure, the problem is mitigated by fundamental training (retard all TLs), and specific training (do not reverse thrust before GSs come up). And, sure, I have no doubt that "holes in logic" like this exist for all makes of aircraft.

B. My understanding is that that is not an F-16 but a CT-155 Hawk trainer out of Moose Jaw.

[SoaringTheSkies]

Lemurian,

I've done this numerous times in this thread, but I'll be glad to do it again:

So what do I think are contributing factors to this accident:

a) reduced sensory feedback channels in cockpit

neither the throttles nor the spoilers lever actually move to feed back current status. All status on thrust and spoilers has to be obtained from the displays, leaving us with just optical input which requires quite a few brain cycles to process. Look at how much brain tissue is related to the different sensory channels, and you'll get an idea how quickly information from those channels is available to us (the more brain, the more processing is needed, the longer it takes).
This is no problem at any time in normal ops and probably not even in most emergency situations, but as has been said earlier by the gentleman who had the bad fortune to have to eject and the good fortune to survive, your brain goes into weird modes. (btw, that's part of my grief with some of the postings here, of course everybody, pilot or just enthusiast, can look at the books in bright daylight and will immediately know what needed to be done to save those souls, and I'm 100% sure if the pilots of that flight were among us, they would know just as well. It's the way our brain processes information that sometimes gets us into trouble.)

b) a logical "dead end street"

The ground spoilers/autobrakes logic is very stringent in what it requires to be true before it will itself allow the retardation devices to do their job.
I don't have the logic diagrams in front of me, but from memory, it's ((wheel spin up) AND (one or both squat switch depressed) AND ((both thrust levers at or near idle) OR ((one thrust lever at or near idle) AND (one thrust lever in reverse)) OR (both thrust levers in reverse))
Oh, and of course AND GS armed but let's take that for a given.
Heck, most laymen would require a minute or two to understand those conditions in plain English!

Back in Warsaw, it was concluded that one of the reasons for the accident was the fact that one of the squat switches was merely millimeters away from it's switching point, but it was still signalling "FALSE" that is, not on the ground. The whole equation was therefore FALSE. Nothing the pilots could have done.
Now this time: After the original omission, the TR left in CLB, was a fact, do you think they'd slap on their foreheads and go "oh buddy, look we've forgotten the thrust lever up there, let's take it to IDLE!"? No!
They have not merely forgotten about it, it was pushed off their list! Now, on the runway, no Spoilers, what did they probably have on their mind? What's the chances the TL they had forgotten about in a stressful situation just seconds ago was on top of their minds now? Speeding down a short, slippery runway with no means of braking? I bet the stress they were in now made what they experienced just a few seconds ago look like an afternoon walk.
I'm sure they had their mind set on exactly two things: brakes and spoilers. Who knows in what order.
Now, this is where I think they've been let down by the system: you can hit the brakes, and as PBL explained, they did so with tolerable timing (intolerable by the runway, but tolerable by human standards). They moved the spoilers lever. Do you want to try to guess why? Because it's the only damn thing directly relating to the spoilers! Tunnel vision! No room to think about the logic equation from up there! No time to rethink your approach and discover your omission.
Yet the logic (and not the computer!) had no way out for them but to answer the 200 souls question: what precious lever have you missed, gentlemen? The clock is ticking! I want your answer!
Wasn't there a chance that the plane might have been hydroplaning? That the main wheels have not spun up? We know why the spoilers didn't come up, they did not! Who knows what they thought the reasons might have been.

Now, I don't want to hear how this or other things could have happened on a Boeing, an MD, an Embraer or a Canadair. They are different Airplanes. They probably have their own design flaws. Heck, the fact that the Lauda Air 767 crashed years ago clearly had a good Boeing contribution. This however is about this Airbus. This is about a system that only takes one answer. And about brains that get into modes where they just can't come up with that one answer.

So it's back to my three requirements to any automation system:

- prevent error
- forgive error
- fail gracefully

pj

ps: sorry that this might be a little less "dry" in tone than what might be considered adequate for this forum.

[BOAC]

- prevent error
- forgive error
- fail gracefully

Thank you for that post, PJ. Like Woodvale I have been in the 'brain is jelly' mode before pulling the handle. I think you have summed it up well.

I also think your 3 guidelines should take their place alongside the laws of robotics.

[ChristiaanJ]

SoaringTheSkies,
Thanks.
I tried to say about the same thing ages ago, but said it far less clearly.

[DingerX]

not objecting to much in that polemic, but maybe some insight into the "brain is jelly" mode:

The beauty of the brain is its simplicity (and apologies to the professional neurologists in the audience for my poor representation): it's basically a bundle of neurons wired to sensors. What makes it work is a box in the back that releases chemicals, reinforcing the neural connections whenever something "important" happens: that reinforces certain neural pathways. If neurons fire without that box reinforcing it, those connections get weaker (this is what happens in REM sleep: neurons fire randomly, weakening the connections and "freshening" the mind). How to judge something important is part of the system as well, so that when and how much reinforcing chemical is released becomes very nuanced.

During a "high-pucker factor" event, certainly, on one side, adrenaline is going to make maintaining total situational awareness difficult. On the other side, that little box is gonna be dumping at Trauma levels: every second is going to be "remembered" in stark (if selective) detail.

All this is to say, part of the "Weird mode" the brain is in, is after-the-fact. The "time dilation" effect is how we remember these events, but is not how we experienced it. We remember it as being twice as slow, but that doesn't mean we think twice as fast.

But, again, there are professional studies on these things, and the good folks who build these things keep current in them. No system is designed in ignorance of these factors; but that doesn't mean they never fail.

[SoaringTheSkies]

Not quite true, Dinger,

actually, when in real stress, we focus on saving our lives.
We do that by switching off, or rather reducing, some of the higher level brain functions which take a lot of time and creativity to perform. One of the things that will fade very quickly is what is called "declarative memory". You know you use it all the time. It's got quite fuzzy borders on both ends. Can you think of the name of your second girl friend's brother (if she had one)? I guess you'll be able to after mulling over that for a few minutes. Can you tell me what three or four requirements have to be met before the ground spoilers deploy on an A320? If you had been unable before this thread, you sure can by now. But if you're not flying the plane, you'll forget it. That is, you'll forget how to remember it. Remembering things from the declarative memory is a rather creative act. You have to think of things that connect to the thing you want to come up with.
Maybe the best explanation is what you do when you've mislaid your keychain somewhere in the house. It's not in one of it's three usual places, so what do you do? If you're any way like me, you'll try to remember what you've done, where you've been since you unlocked the door. That will help. Declarative memory.
And you also know that you get less and less good at this when you get under stress.
However, if you know how to ride a bicycle, I'm pretty sure than even a rather high level of stress won't make you fall off your bike. That was different when you were a child. Has it ever happened to you back than that someone hollered your name and you looked around and - fell on your nose?
Riding a bike is one of those things that slowly trasitions into procedural memory.
In aviation, memory items are design to reside exactly here, in procedural memory. Repetition makes memories go there over time.
Procedural memory is very fast to access, it's usually very simple stimulus-action type of memory.

In stress, we're very good to access procedural memory. It's where everything is stored that we needed to get away from a predator at the dawn of humanity, probably even earlier than that.
We are very bad at accessing our declarative memory, though.

pj

[borghha]

Excellent posts, SoaringTheSkies. And thanks for summing it up!

It is quite true that in emergency situations our brain is programmed to get away from the danger as soon as possible: flee! The adrenaline kick stimulates all fysical abilities and makes us act fast. But this implies also that higher brain functions (the declarative memory) are less readily accesible at that moment. On the contrary, the brain makes sure to learn from the experience: it activates a memory booster (the amygdala) that strenghtens the saving mechanism, in order to make us react even more rapidly the next time round. This is one of the reason why people suffer from the post traumatic syndrome.

Where it concerns the case at hand (one TL left in CLB), I was thinking of an example involving keys too: each morning I take 4 items with me when going to work: wallet, keys, cell phone and service card. No problem, I 'll always remember them. But if for some reason this routine gets disturbed by eg. a fifth thing I have to take with me only that day, it happens quite often that I forget one of the four routine ones...
Still find it irresponsible of AB and regulators not making the logic update mandatory after two earlier incidents of the same kind, luckily with less or no people killed. It could have prevented this accident. Commercial arguments seem to be very powerful.

[ChristiaanJ]

SoaringTheSkies,
Thanks again for your post.
In a way, this is HOTAS, "hands on throttle and stick", telling you by physical feedback what you're telling the aircraft to do.
The rest of your awareness should be outside and "ahead" of the aircraft.
If you have to duck back down to read the "video game" and interpret it, it's already too late.
Too simplistic, sure, but I hope you understand what I mean.
CJ

[Lemurian]

StS,
Under the guise of serene objectivity, your post smacks of prejudice and bias.
The main flaw of your argument is that it can be thrown back at you :

a) reduced sensory feedback channels in cockpit

I'd say that :

• The totally unusual throttle stagger was a peripheral vision cue
• The feel of one small button (the top of one throttle) is vastly fdifferent from that of a handle (the two T/Ls together)
• The direct visual cue of a very dynamic engine performance display was also a very important information.

HOWZAT for sensory feed backs ?

b) a logical "dead end street"

You are either dishonest or totally unable to read a flow chart -and in this case you're disqualified to talk about what you don't know - ; as for the auto spoiler operation my understanding is a lot simpler - and in fact, any objective observer would only see that the logic has added safety built-in.
Let's review it :

• Spoiler handle ARMED + T/Ls at or near IDLE + 1 GROUND condition --> DEPLOYMENT
• 1 T/L in REVERSE + 1 T/L at or near IDLE + 1 GROUND condition --> DEPLOYMENT
• 1 or 2 T/L in REVERSE (+) other at or near IDLE + 1 MLG strut compressed --> PARTIAL DEPLOYMENT.

The *GROUND* condition is either both struts compressed + Rad ALT < 6 ft OR wheels spun up > 72 Kts.
Note that the *Spoiler ARMED* condition only appears once.

I'm sure they had their mind set on exactly two things: brakes and spoilers.

Boldest words I've seen lately !

Now, I don't want to hear how this or other things could have happened on a Boeing, an MD, an Embraer or a Canadair. They are different Airplanes.

Why ? Because you might find that , regardless of the type, they could have found themselves in the same predicament, and therefore destroy you anti- Airbus stance ?

[Lemurian]

Rananim

NO SIR.You wont get TR#1.Both TL's must be idle to get TR.They wont get GS either(gear not tilted,TL's idle).Crucially,this leaves the GA option.And theres your big difference.The plane wont dig a big hole for you and your 200 passengers.

I don't know where you found your info.
Excerpts from the 777 system description manual :
A. Thrust reversers.
... Reverse thrust is only available on ground. The R/T levers can be raised on;ly when the forward T/L is at the IDLE position.
The EEC inhibits reverser isolation valve actuation and reverser deployment unless the aircraft is on GRD with the engine running...
Note the singular mode.
B.Speedbrakes
...In the ARMED position, the speedbrake lever is driven aft to the UP position when :
-Landing gear fully on the ground (not tilted)
-T/Ls at IDLE.
On the ground, when either reverse T/L is moved to the reverse idle detent, the speedbrakes automatically extend. The speedbrake lever does not need to be in the ARMED position...
Did I read correctly ?
Or is there a misunderstanding on the definition of a tilt ?

[marciovp]

Ifs

To try to wrap things out. First and foremost, in hindsight it is easy to know what could have been done better. Let me go into the ifs that perhaps would have made a difference in this sad and terrible disaster:

• If TAM did not fly with the reversors locked out...
• If TAM did not put more fuel than necessary in that plane...
• If TAM did not fly with so many passengers and employees...
• If the co-pilot also had a long experience with A-320s...
• If it was not raining...
• If Congonhas had a longer runway and escape zone
• If Congonhas was not slippery as reported by other planes
• If the pilot decided to land at Guarulhos instead...
• If the pilot did not forget to idle the right TL
• If TAM had installed the new warning system for this situation
• If the pilots would go around when the GS did not arm
• If the pilots discovered the position of the right TL before landing

I guess, the disaster would not have happened if 1, 6, 8, 9, 10, 12, just one of them, was present.

2,3,4,5,7,11 could have made a difference in terms of less pressure and distraction on the pilots.

Oops... I forgot this possibility of not being able to move the TL. Yes, bomarc, "desacelere" in Brazil specifically means "reduce thrust". And "nao consigo" means what you said in your poster: difficulty moving something or doing something.

Some very good folks in Brazil, accustomed to the slang/colloquilisms of Portuguese, have contacted me and explained that the words used on the CVR/Transcript meant that there was difficulty in moving something (thought to be the throttles, or the devices used to control acceleration), just like words in English that might have meant difficulty in opening a Jar...something was stuck.

While arguments in favor of the plane vs. the pilots have been quite well written, there is still a chance that moving the throttles was not possible.

[bomarc]

First off let me say I don't care for boeing or airbus.

Now that I've got your attention.

An airbus pilot that I know explained to me how the thrust reversers are selected on the airbus.

This explanation made it abundantly clear that pulling or trying to pull both thrust levers to idle would be a natural, instinctive if you will, method of landing the plane. While there isn't a reverse lever/throttle, something akin to that is required to enter the reverse regime.

Some very good folks in Brazil, accustomed to the slang/colloquilisms of Portuguese, have contacted me and explained that the words used on the CVR/Transcript meant that there was difficulty in moving something (thought to be the throttles, or the devices used to control acceleration), just like words in English that might have meant difficulty in opening a Jar...something was stuck.

While arguments in favor of the plane vs. the pilots have been quite well written, there is still a chance that moving the throttles was not possible.

Add to this, there were two pilots who could equally get to and see the thrust levers/throttles. The pilot in the right seat, someone hired as a captain, would not give up his life if he could have pulled the throttle to idle.

Ask yourself pilots in the right seat, would you let someone kill you, just to keep your job?

We may remember that it was the copilot (right seat) who finally got thrust reversers to deploy on the Southwest Airlines 737 at Chicago Midway...some 18 seconds after touchdown.

We've talked about going around after the landing...yes an option, but we have the time to think about it.

During the early years of both automobiles and airplanes, stopping both was sometimes accompanied by the operator/pilot/driver saying, "WHOA"...just like a horseback rider might have done.

Some of the earliest French planes required the throttle to be pulled back to increase power. How did this make sense? Well, pull back on the yoke to go up, add power by pulling back on the throttle to go up.



So, look at all the schematics of the airbus. Read everything you can about it. Wonder about the human factors. Play with it in the simulator...but ask yourself what you would have done in the right seat!

[Rob21]

The "relationship" between pilot and throttle was weakened by auto throttle.
Want to reduce speed? Adjust the dial, no need to move "levers". It's time to do the flare? Don't worry, engine thrust will come to idle. Forget the TLs...
Reverse now? OK, let's move the "thing" (oh, yeah throttle) back.
TLs are no longer primary control of speed, as we learned at flight school.
Nowadays, the speed dial on the panel is primary control, am I wrong?
When an Approach Controller requests you to reduce or increase speed, where do your hands go? Speed dial or TL?
Moving throttle is the last "intimate" remembrance pilots have that those levers control speed. Even though pilots don't mess with it very much any more, to "see" it move gives us the illusion (remembrance) that TLs are controlling our speed. Auto throttle disconnected, flare, nice, look at those TLs coming to idle... They are controlling my speed, good!!
IMHO, maybe it's time for aircraft engineers to rethink cockpit design. Pilots transitioned well (I think) from yokes to joysticks for aircraft atitude control. But TLs are still the same "shape" they were 50 years ago, and they don't do what they used to do any more. Aircraft engineers, don't be afraid to change those old fashioned throttles. Put some nice acrylic levers, that change colors. Green, thrust forward. Red, thrust reverse. With a glimpse, we will know if they are OK. I believe pilots will learn very quickly how to deal with the new design, like they did with joysticks.
Again, sorry if I got carried away...
Rob

[Dani]

If someone wants to insinuate that the TL no 1 was stuck in climb mode, that still doesn't take away the blame from the pilots.

What would you do in a split second if you see that it's stuck and engines are not spooling down? Right, cut the engine by the fuel lever, called Master Switch on the Airbus.

Dani

[armchairpilot94116]

What we know is a true-ism: history tends to repeat itself

SB not being made mandatory contributed to a tragic accident (airlines not faultless either for not making those recommended changes). Pilots fighting against their machines after mishandling of a vital control and not understanding what their predicament was. Congonhas wasn't the first and unfortunately probably won't be the last. But more attention needs to be paid to make a good product better.


read PBL's article here regarding the 1994 CAL A300-600 crash in Nagoya
(scroll down to 1994)

http://www.rvs.uni-bielefeld.de/publ...ei/taipei.html

Does anything sound familiar?

[SoaringTheSkies]

Now, I don't want to hear how this or other things could have happened on a Boeing, an MD, an Embraer or a Canadair. They are different Airplanes.

Why ? Because you might find that , regardless of the type, they could have found themselves in the same predicament, and therefore destroy you anti- Airbus stance ?

Ok, Lemurian, after this, I'm entirely sure who's firmly in one turf, and of the two of us, it's not me.
You want me to be an anti-Airbus person? Well, if it gives you peace of mind... Feel free to put me in any turf you see me fit, but be aware that I'm not going to change for you. It's all in your perception.



pj

ps: it's a pity to find someone so deeply buried in the trenches.

[borghha]

Quote

Some very good folks in Brazil, accustomed to the slang/colloquilisms of Portuguese, have contacted me and explained that the words used on the CVR/Transcript meant that there was difficulty in moving something (thought to be the throttles, or the devices used to control acceleration), just like words in English that might have meant difficulty in opening a Jar...something was stuck

.

Sorry, but to imply that 'não da' means 'it's stuck' is jumping to conclusions! não da indicates that it was impossible to reach the desired result, ie. 'desacelerar', to reduce speed.

[RoyHudd]

I suspect the poor PF's brain was stuck, not the TL. He had probably left one TL generating thrust, and had removed his hand from it, believing that it was in the idle detent. He just selected full reverse on the other (non-locked out)engine. (The data recorder will reveal all, this is just my simple hypothesis.)

And PNF did not think to visually check the TL positions with all that was happening. Easy to imagine, unpleasant consequences for all.

[SoaringTheSkies]

If TL #2 was stucked they would have shutdown ENG 2 immediatly.

just out of curiosity. The logic for the ground spoilers refers to the thrust lever positions. It does say nothing about the engine running or not.
Would a switched off engine with the thrust lever out of idle still be inhibiting spoilers or not? I would suspect it would be taken out of the equation. Would be interesting to see that in the FCOM.

pj

[hetfield]

@STS

Yeah, think so too. Anyhow the pushing force would be eliminated.