Emirates B-777 incident at DUBAI.

[facthunter]

That's why they used reduced thrust all the time, for take off and climb. ALL twin jets are overpowered on two. Nev

 

 

[dutchroll]

Yeah the A300 was a bit of a different beast. However the subsequent Airbus models (320,330,340,380) are all remarkably similar in the way they work. That's an intentional philosophy from Airbus.

 

But they cant press a TOGA switch/button, if such a thing in not on/in an airbus.

On the Airbus these days we are often told to think of the thrust levers/throttles themselves as "switches". There are a number of detents in the throttle quadrant and depending which one you advance them to, the engines and aircraft will do various different things. The "TOGA detent" on the throttle quadrant in an Airbus has much the same effect as pressing the "TOGA switch" on a Boeing, but through a slightly different mechanism. It commands full thrust and activates the flight director pitch and roll modes into their respective "go-around" functions. If the autopilot happens to be engaged when you do that, the plane will do it all automatically. If you're manually flying at the time you'll still get the commands but you have to do some of that pilot stuff to follow them.

 

The Airbus which sort of pancaked into the trees was the age-old screwup: Pilot has idle power set, descending for show-off low pass. Pilot suddenly realises his flight path is not looking good and wants much more power. Engines can't respond quick enough. Pilot crashes plane.

 

 

[Roundsounds]

I agree: how did he get to 175 ft with the engines at idle? That doesn't add up to me either.

If the TOGA switches were pushed when inhibited (below 5'RA for x seconds) a further push would be required when out of the inhibit - ie airborne plus x seconds.

 

 

[flying dog]

The Airbus which sort of pancaked into the trees was the age-old screwup: Pilot has idle power set, descending for show-off low pass. Pilot suddenly realises his flight path is not looking good and wants much more power. Engines can't respond quick enough. Pilot crashes plane.

Wow, that is different to the official story I heard.

 

They had the "FMC" in landing mode and so when they commanded takeoff power, the computer just bawked at them, as it thought "You can't do that in THIS config"

 

That is a cut down version and: Yes, it has holes. I can't remember the 100% word for word version.

 

When landing, you have to expect GO AROUND requests, but something was done incorrectly and so the plane continued on "landing".

 

Shame there was no runway on which to land.

 

 

[Bennyboy320]

They had the "FMC" in landing mode and so when they commanded takeoff power, the computer just bawked at them, as it thought "You can't do that in THIS config"

The aircraft was 100% serviceable & doing what it was supposed to, the pilot can always go around at ANY altitude & ANY FMGC mode by manually pushing the thrust lever to the fire wall i.e. selecting TOGA, the same thing almost happened to Jetstar except they didn't crash at MEL 21JUL07 JQ156, both incidents caused by poor systems knowledge.

 

 

[dutchroll]

You stick the thrust levers into the TOGA detent (ie, firewall them) on an Airbus and, as Benny said, you'll get full thrust no matter what "mode" (technically they're called "phases" on the Airbus) the FMC is in. Full forward means only one thing to the engine computers (FADECs). "Gimme all you've got!"

 

Even if it failed to sequence the computers from the "approach" phase into the "go around" phase, you'll still end up with maximum thrust and if you're hand flying, you just pull back on the stick to the correct pitch attitude and away you go. It's no different in principle to if you were flying a Bonanza.

 

At worst, if it was still in the approach phase, the flight directors would give you dud info, thinking you still want to descend on your approach path to touchdown. But if you have full thrust you just manually do that pilot stuff and pitch it up to 15 degrees and get the other pilot to turn them off if need be.

 

Also if you don't get the thrust levers fully forward and instead leave them in the climb detent with autothrust engaged, that's going to ruin your day too, because you're just telling the autothrust system it's "business as usual" and it thinks you're still on approach and want to land, so all it wants to do is maintain your approach speed. The Airbus thrust levers normally stay in the climb detent from after takeoff until landing when you pull them to idle. When they're in that climb detent, the autothrust engages and what that does is driven by what modes you're using with the auto flight system, and which phase of flight you're in (or rather the computers think you're in!).

 

However on an Airbus, ask for TOGA thrust and TOGA thrust you shall receive, always. If you don't ask, you don't get.

 

It's only two "clicks" (detents), but there's a chasm of difference with how the plane and its systems behave, between having the thrust levers in the climb detent and having them in the TOGA detent. You have to know this stuff if you're an Airbus pilot.

 

 

[onetrack]

But don't the Airbusses have an additional problem with regards to dual inputs? I'm thinking Afriqiyah Airways Flight 771 here.

 

 

[dutchroll]

Well.....not a "problem" as such. Just pilots fail to have a good knowledge of the system and what it tells you, and discipline in handing over/taking over control like our instructors taught us when we all learned to fly.

 

So the plane has a sidestick controller on each side, obviously one for each pilot. It is totally "fly by wire" so there's no physical link between the two. If simultaneous inputs are made, the computer "averages" them. I guess it was designed that way because if it did anything else, a dual input could lead to pretty extreme movements.

 

If one pilot pulls full back, and the other pilot pushes full forward, nothing happens because the inputs are averaged and cancel out.

 

However the plane will loudly announce "dual input!" at you if both pilots make a control input, as well as illuminating a warning light in front of each pilot. There is a pushbutton on each sidestick which if depressed (whichever one is pressed first) will instantly take priority from the other sidestick and if held long enough, will lock it out all together. This is how you take control if the flying guy is doing something really bad and you don't have time to talk about it.

 

The bottom line is that discipline over who is actually flying the plane is required (which has been the case ever since dual control planes were invented) and positive confirmation or action is required to have one pilot take control if dual input is shouted at you by the computer, but we've seen accidents where this discipline has broken down. AF447 which stalled into the Atlantic had this happen, among many other things. You'll see "dual input" in the CVR transcripts while Bonin was holding his sidestick fully back keeping it in the stall though he didn't understand that he was stalled, while the other guy has tried to push forward. No-one ever really did anything about it. The senior First Officer should've screamed at him "I HAVE CONTROL!" and/or pressed his takeover button.

 

It's just one of these basic principles which tragically broke down. Only one guy flies the plane at a time, and everyone in the cockpit needs to be fully aware of who it is.

 

 

[dutchroll]

Correction:

 

In fact both inputs are algebraically added, not "averaged". It does this up to maximum deflection that would be available to a single sidestick. So the effect of one pushing forward and one pulling back is the same - one essentially cancels out the other, or partially cancels it out. Sorry....been a while since I've done the ground school and I had to refresh myself because I couldn't recall the exact details!

 

It's on page 1,371 of our 4,255 page Flight Crew Operating Manual for the Airbus A330, if you must know.... (that's true!)

 

The principle is as I said though. It tells you, and there needs to be clear communication of who is doing what.

 

 

 

[facthunter]

A comprehensive report was released two days ago on the Emirates 777 accident. I don't have authority to pass mine on but I have no reason to believe someone couldn't access it by searching thoroughly. Nev

 

 

[Birdseye]

This is available:

 

ASN Aircraft accident Boeing 777-31H A6-EMW Dubai Airport (DXB)

 

 

[dutchroll]

It has been released on the GCAA (Emirates Civil Aviation Authority) website.

 

https://www.gcaa.gov.ae/en/ePublication/admin/iradmin/Lists/Incidents%20Investigation%20Reports/Attachments/90/2016-2016%20-%20Preliminary%20Report,%20AAIS%20Case%20AIFN-0008-2016%20-%20A6-EMW.pdf

 

As expected....monumental stuffup.

 

 

[onetrack]

The PF had both MLG on the ground initially - why didn't he put the nose down, and complete the touchdown? Because he'd touched down too far down the runway?

 

 

[facthunter]

It's interesting that Boeing do not have a specific rejected landing procedure. The wind changes were significant headwind tailwind component and 48 degress Phew.! Over 42 used to keep the B727 on the ground. Nev

 

 

[dutchroll]

I imagine so. It's a little hard to tell precisely from this report.

 

The initial touchdown at 1100m from the threshold is fine. That's well within what you'd call acceptable. But he had a fair bit of tailwind until just before touchdown. To us, this would be a red flag to go around because if it swings around to headwind just before touchdown you are now in overshoot shear and can find yourself floating uncomfortably down the runway trying to touchdown.

 

The interim report is sparse on the fine detail here but I'm sure the full report with full flight data and CVR matching will clarify it.

 

But the bottom line is what I said back at the beginning. Once you decide to go around just do it. Full power, set the go-round attitude for your plane. Worry about what happened when you get on the ground. Works in a Drifter, Cessna, or B777. They missed one crucial step. They had no thrust on!

 

 

[facthunter]

Don't retract the wheels till you absolutely sure you won't need them. I don't like the idea of being superior in ideas and skills to those who were there, because they were there and we were not.. It's fortunate that with so much damage and a fire, one poor fireman was lost when the centre tank exploded, unfortunately. Nev

 

 

[dutchroll]

I imagine so. It's a little hard to tell precisely from this report.The initial touchdown at 1100m from the threshold is fine. !

Sorry.....mistook metres for feet.

The touchdown was waaaay long!

 

 

[dutchroll]

My take on it so far based on the presented facts:

 

- the approach speed was pretty good throughout.

 

- going through 700ft the wind changed from slight headwind to a tailwind of 16 knots. The reported ground wind was a tailwind of several knots. This would perk our interest because that indicates a probability of overshoot shear.

 

- that's not a "severe" shear, but it is significant and does require a bit of care to handle.

 

- "overshoot shear" (increasing headwind or reducing tailwind) is actually fraught with complications when it's close to the runway because it results in the aircraft picking up excessive energy, having the thrust at idle to prevent the speed increasing (not desirable to have idle thrust late in the approach), and can result in a long float. Bad overshoot shear can often result in a go-round close to the ground as it becomes obvious the plane just won't touch down.

 

- he started the flare at 35 ft and that seems about right to me but it took 10 seconds from there to touchdown. That's a significant float.

 

- so re-reading it I realised the touchdown was at 1100m (not feet!) which is way too long. A touchdown at 1000-1500 feet from the threshold would've been right on the money for that size aircraft. You don't want it shorter than 1000 ft and not longer than 2000ft. He touched down at 3200ft.

 

- this explains the aircraft barking the "long landing" message at him.

 

- he has then attempted a go-round, as he should.

 

- he pulled the nose up for the go-round and the aircraft climbed because it still had some energy. They went into the standard Boeing procedure of flaps 20, positive climb, gear up.

 

- he never advanced the thrust for the go-round and the F/O didn't pick him up on it either and by the time they worked it out, it was too late.

 

There are a lot of automation complacency issues to come out of this.

 

 

[Old Koreelah]

Crickey Dutch, how do you blokes do it? I'm busy enough juggling just a throttle and a stick to land my little Jodel!

 

 

[dutchroll]

Crickey Dutch, how do you blokes do it?

It's ultimately just a plane. Disconnect all the automatics and it still flies in accordance with all the basic principles.

The big issues coming out in recent years in Airbus as well as Boeing are automation mishandling (especially autothrust modes), automation surprise, and related topics.

 

Both Asiana and Emirates were perfectly 100% serviceable B777 aircraft which are now total hull losses.

 

 

[cooperplace]

My take on it so far based on the presented facts:- the approach speed was pretty good throughout.

 

- going through 700ft the wind changed from slight headwind to a tailwind of 16 knots. The reported ground wind was a tailwind of several knots. This would perk our interest because that indicates a probability of overshoot shear.

 

- that's not a "severe" shear, but it is significant and does require a bit of care to handle.

 

- "overshoot shear" (increasing headwind or reducing tailwind) is actually fraught with complications when it's close to the runway because it results in the aircraft picking up excessive energy, having the thrust at idle to prevent the speed increasing (not desirable to have idle thrust late in the approach), and can result in a long float. Bad overshoot shear can often result in a go-round close to the ground as it becomes obvious the plane just won't touch down.

 

- he started the flare at 35 ft and that seems about right to me but it took 10 seconds from there to touchdown. That's a significant float.

 

- so re-reading it I realised the touchdown was at 1100m (not feet!) which is way too long. A touchdown at 1000-1500 feet from the threshold would've been right on the money for that size aircraft. You don't want it shorter than 1000 ft and not longer than 2000ft. He touched down at 3200ft.

 

- this explains the aircraft barking the "long landing" message at him.

 

- he has then attempted a go-round, as he should.

 

- he pulled the nose up for the go-round and the aircraft climbed because it still had some energy. They went into the standard Boeing procedure of flaps 20, positive climb, gear up.

 

- he never advanced the thrust for the go-round and the F/O didn't pick him up on it either and by the time they worked it out, it was too late.

 

There are a lot of automation complacency issues to come out of this.

as a Jab pilot, I know zip about the big planes, but I can't imagine any pilot forgetting to advance the throttle on go-around, so is it a good assumption that he thought an automatic system was going to do this but didn't? There's speculation earlier in this thread and elsewhere that the wheels touching the ground might have made the systems "think" the plane was on the ground so that stopped the throttles advancing. How does this sound to you, Dutch?

 

 

[dutchroll]

as a Jab pilot, I know zip about the big planes, but I can't imagine any pilot forgetting to advance the throttle on go-around, so is it a good assumption that he thought an automatic system was going to do this but didn't? There's speculation earlier in this thread and elsewhere that the wheels touching the ground might have made the systems "think" the plane was on the ground so that stopped the throttles advancing. How does this sound to you, Dutch?

Yes I reckon that's about right.

He had the autothrust engaged on the approach. Nothing particularly wrong with that.

 

Boeing use "TOGA" (Takeoff/Go Around) switches on the thrust levers to initiate go-arounds. Pressing these engages the servo motors on the thrust levers to automatically drive them full forward (on the Airbus you actually have to physically push them full forward - there are no servo motors on Airbus thrust levers which are designed differently), however those switches are inhibited after touchdown (based on radar altitude). But go arounds are almost always initiated from some reasonable height above the runway, so through all your training and even 99% of the time in real life you'd be used to pressing the TOGA switches and getting full thrust, not nothing. In that 1% case of having already touched down, that's not what happens.

 

Therefore the B777 procedure in the books if you go-around after touching down is to make sure the thrust levers get pushed fully up. But had they ever practiced this? I wonder.......

 

It appears they were caught out in the heat of the moment by how the system behaves in that rare scenario. All they had to do was physically push them up.

 

 

[Old Koreelah]

One interesting distraction (I'm good at those): I noted in the crash report was that this plane had done over 13 thousand trips and had done almost 60,000 hours. And people say that these days nothing's built to last!

 

 

[cooperplace]

there are lots of 747s with >100,000 hours; you've got to hand it to Boeing, their stuff is built really well. Imagine a car doing 100,000 hours!

 

 

[Old Koreelah]

there are lots of 747s with >100,000 hours; you've got to hand it to Boeing, their stuff is built really well. Imagine a car doing 100,000 hours!

Roughly equivalent to a car doing ten million km.