Bait?
 

Definitely not intended to bait anyone, and I appreciate the reply!

suninmye - you stated - "If available thrust reduces, the 777 autopilot will fly the ILS glideslope until the speed reduces to just above the stall. It will then ignore the glideslope, lower the nose and descend just above stall speed. If the 038 had hit the ground in this configuration everyone would not have walked away. That statement is based on having seen it a few times in the simulator. Full up elevator, (unheard of in a 777 landing) had been applied at the appropriate time just before impact to cushion it so that many of the occupants thought on touchdown that it was just a hard landing."
-----------------------------------------------------------------------

Do you have an official statement that supports your position? The FDR data released by the AAIB doens't show enough data to support your statements, which is why I ask.

There is evidence of an increase in pitch attitude, and a slight decrease in speed, just prior to impact but slight changes in pitch and speed were present just prior to the last FDR data.

Also, if the a/c is being governed by low speed protection, what effect would moving the yoke, to increase AOA, have?

@ WojtekSz
The ILS does not lead the aircraft to the begining of the runway but to the touchdown zone abeam which the glide path antennas are located. On the ILS you usually cross the threshold at 50 ft AGL and touch down some 300 m further.

Keeping the AP on probably means that the crew was expecting thrust to be recovered without deviating from the ILS. Disconnecting the AP and flying below GP to maintain some energy to reach the threshold at a decent Vz is most probably feasible (as in the sim) but would require immediate identification of the loss of thrust.

The OCH on an ILS is decreasing with the distance from the threshold, but gives some room to follow a "flatter" path. I was basically saying that since the acft had enough total energy to reach its actual touchdown point (in the grass) it could probably have reduced its impact Vz by saving energy in not following the GP, which seems to have been demonstrated in the simulator.

S.F.L.Y.
 

Hindsight is great isn't it!

I have been told by a couple of sim instructors ‘in the sandpit’ that with a few months of 'hindsight' and accumulated knowledge about this accident, a fair number of crews were 'surprised' with this scenario in the sim; and some of those that claimed to know the method to achieve the actual outcome or 'better', suddenly 'froze' in the sim. Others followed the method I had used on the day, but with hindsight were able to change the flaps earlier, and they still only achieved the same distance or worse! The majority of crews faced with this problem, crashed well short of the perimeter fence; so the people who claim to have done any better; have more forethought than me or less headwind!

It would be interesting to know if those that have “made the r/w” in the sim have used timings based from when we officially had rollback at 720ft or when we, on the day, actually realised we had a problem at 480ft. At 480ft in the sim, are pilots actually then taking the time to decide if this is a ‘fuel’ problem, an EEC problem, a windshear scenario, a double engine failure with EICAS failure…. the list is endless… or..jumping straight to the conclusion that it must be ‘rollback’, which of course was an unknown event on Jan 17th 2008!

Unfortunately for any crew, if this ever happens again (for whatever reason, but it shouldn't now anyway) will be that it is now an emergency procedure with a memory C/L, so would have to be carried out to adhere to SOPs. Will they… close the thrust levers for 10 secs to warm the FOHE or will they change the flaps or both. Two of those will be wrong according to any new SOPs…a big grey area. Pilots will naturally take some seconds to correctly identify the failure as ‘rollback’ and even then might not believe it; so to carry out this drill at low altitude will be scary. Not many pilots I know would like to close thrust levers for at least 10 secs below 400ft on an approach. The spool-up time would be too late. Please also bear in mind another real life rollback event where the crew knew of the event on the BA38; but in a relaxed state of mind in the cruise, they still took some time (in minutes) to realise this was a rollback before they reacted.

I accept that the r/w has been achieved in some sim sessions, but I am not in possession of the sim inputs for my scenario. I appreciate that some posts that make reference to making the r/w are backed up with “with the benefit of hindsight”, however, I don’t know what the settings were set to, or if the drill was only ever practiced with a “complete knowledge that there is a rollback at 480ft and what can you do to get the aircraft down” – as opposed to treating it like an unknown emergency situation ... but would you actually have wanted to take the risk of landing on the tarmac, with massive potential for a fire?

If it was always treated as an emergency situation, then, for those of you who made the r/w, I am sorry for my explanation above, but I am sure you can understand why I felt the need to explain to those non flyers the information needed when coming to a conclusion that the engines are suffering ‘rollback’ as opposed to any other emergency.

Boeing have tested this to the limits in their sim; as you might expect, so you will be able to read their results in due course when the AAIB final report is issued this year.

I, and my family, await its’ release eagerly.

Had the a/c made it to the concrete, 10,000 pounds of fuel was waiting to ignite in the ensuing spark fest. I don't see how the a/c could have made the strip had one practiced the profile for weeks. Given they made it over the fence, missed the hard stuff and no one croaked, what's to complain? Good grief. The Fuel, the fuel, boss.

bear

Question remains, would it be recorded the moment the spars moved from their open position or would a record be made only if the spars reached their closed position? In other words, if an uncommanded partial closing of the spars had occurred which rectified itself before they reached the fully closed position, would that have been recorded? If only recorded when reaching closed position: spars take up to 15 seconds to fully close so if an anomaly lasted anything less than 15 seconds and the spars moved back to their open position, they could have restricted fuel flow considerably without having a record of it and, as reported, the spars would be found in the open postion.

The question isn't to know if it was or not technically possible to make the runway. Before being limited by any technical factor the aircraft was limited by its crew's situation awareness. Appropriate actions cannot be taken before the crew realize and identify the problem The pilots were obviously very confused as they let the AP flying the aircraft at a dangerously low airspeed without taking manual control, most probably because they didn't have time to realize they should give-up on the ILS trajectory.

As a pure coincidence, the flap retraction probably caused the AP to disconnect seconds earlier than with full flaps, which gave the crew enough height to recover from the subsequent 8 degrees pitch reduction and high Vz (100 ft lost in 2.5 sec --> 2.400 fpm).

@Mmmayday38
whatever you did at the time has worked out fine. With the hindsight i cannot find better way of doing this. I believe many should be grateful for your instinct reaction to the problem. Because i think that it was instinct flying more than anything else - even reading through the desription takes longer time than you had to make decisions.
The sim test as described by M.Mouse show clearly that to reach positive results they had to start corrective immediately at 700ft. It is nice to know what would happen ;)
Would the new breed of low hours pilots know how to react in such demanding situation? Would they have enough flying instincts developed?

@suitcaseman + S.F.L.Y
thanks for correcting on the ILS :)

@ S.F.L.Y
i believe you are right that appropriate situation awareness has significant influence on security of the flying. I do not know if the pilots may/should have turned the AP earlier but i assume the rules were not broken as this would be clearly stated in the post accident report.

do you really believe that retraction of flaps was a coincidence? Commended 17 sec before impact while the AP disconnected some 7 sec before impact?

Are you trying to prove it was human error?

@ WojtekSz

I don't know what is so unclear in my statement but I never said the flap retraction was a coincidence (how could it be?!). What I meant is that by coincidence (as it wasn't the crew's intention) this flap retraction accelerated the occurrence of the pitch down movement at AP disconnect. Should the aircraft had few more knots to loose before reaching its lowest speed this AP disconnect and pitch down movement would have happened later, closer to the ground, with no more room to reduce the high Vz. Few seconds before the impact the Vz was way too high as the aircraft initiated a lazy stall.

You are talking about the low hour pilots and flying instinct. Where is the flying instinct when you are VMC, approaching stall speed with no available thrust and that nobody stops the AP to increase the attitude? The most basic flying instinct would command to take over manually just like in a trim overrun. Luckily the aircraft did it by itself and disconnected the AP as it reached unsafe limits. Is it really the aircraft's job to disconnect the AP when speed gets so low?

The autopilot disconnect was caused by a control input. From the report found here:

M.Mouse, you're right about this control input which is hardly reflected in the data report (no pitch decrease before AP disconnect). Anyway this happened 6.5 sec after speed dropped below 110 kts. At that point there was nothing else to do than lowering the nose to save speed (by increasing Vz). It's not only a matter of trading Vz to save speed to fly above Vs, extra speed is subsequently required to absorb some of the Vz prior to impact. Maintaining a speed just over 100 Kts couldn't reduce the Vz below 1400 fpm.

@S.F.L.Y
what i read from your comments is the overall idea that if the PF would have done something (lowering the nose) earlier the plane could have landed as usual. Is
that right?
What i do see from the accident report is that PF has reacted (increased thrust) about 5 sec after the speed got below 135kt, and this would be about anyone would take to wait for the AT to correct the speed. At that time i would not expect ANY pilot to start lowering the nose - would you believe otherwise?

We shouldn't have the same report since in mine the X axis is graduated in 2 sec increments. My reading shows that after the speed got below 130 kts you have 13 units before AP disconnect, meaning 26 seconds... How many Kts of deviation are tolerated on approach during a JAA IR test flight? The same graph shows 7 seconds below 110 Kts before AP disconnection. With no offense to the crew this seems to be a lot of time, especially when Vref was lost for over 20 secs.

The report uses the term "movement," which would seem to cover that:

we have the same report ;)
the difference is that you seem to believe that the first crew action was commanding nose down while i seem to follow the report thinking that 'the autothrottle and the flight crew commanded full thrust' and this has happened much earlier - just after the speed got below 135kt or latest
And to me it is fully understandable as adding more thrust is what most pilots do to add speed while landing :). And since B777 is a jet hence it took a while to acknowledge no speed increase. And all this is the time AP was constantly increasing pitch.
I would not consider take this time as no action.

suitcaseman: i am fully with you on this :)

At least two guys are having fun. I don't care about criticizing the crew, I'm just interested in understanding how they reacted or not and why. I just observe a certain lack of action while speed decayed and I want to learn from it. If it happened to them it might well happen to me as it would probably happen to most of us. We very often act far from the most logical strategy and analyzing this is certainly a good step in improving flight safety, at least better than randomly distributing safety medals. Now if this also amuse people then it's even better.

Thanks god my personal experience of emergencies is limited to a single good ending event. Not because I'm a sky god but because I had good support from the ground.

@S.F.L.Y
could you pls elaborate on what would be the most most logical strategy in this case? This might help the me (possibly also others) to understand your point.

Again, this is not the point. Some found out that different tests in the sims could lead to different results. Fact is that in real life crew's reactions are what they are, and not always appropriate since the human brain needs to update its situational awareness and subsequently its strategy. This is why we are trained to perform some reflex actions to some alarms. A normal reaction to speed deviation from Vref is thrust correction, which was attempted by the crew in this particular case. Of course it takes some time for the crew to realize the action is ineffective. At this point the speed keeps going down and the crew naturally focus on sorting out the thrust issue as its difficult to admit you are now flying a glider. This particular point is the most critical as this is when the have to accept the problem and look for a different strategy. It is very interesting to me to see that actions were taken on the flaps (drag reduction strategy) prior to taking manual pitch control. It's probably a much easier mental process to give up on the initial flap setting than on the AP ILS approach. Taking over manual control of a critically low speed and powerless aircraft is a natural decision for any pilot. Why did it happen so late? My point is not to criticize the pilots, but to understand why they didn't take the actions we could expect in such situations.

In other words I don't want to know if I would do better than them, I want to understand why I might be doing the same thing. Hope this time my point is clear.

Apart from one post that mentions flap to 20 and using ground effect and flaps to 30 with loads of information to pre plan in a sim.

The above does/may not include what was going wrong with which systems and how much information was avail for the crew.

The aircraft may of had more problems than just all donks rollback.

The crew were near the end of some very long flights and all that goes with that.

It still appears to me they put that bird down in a good spot with the right speeds and angles to make a nice story.

Max respect to all the crew, you made a great team on that day, you will never buy a drink in my company, happy new year and all the best for 2010.

Mis-dagin
 

My comments were based on my experience of recreating it in a simulator several times and 2nd hand reports from the guys who were at the front.

The flight data recorder did not show the final up elevator movements as it stopped recording before impact.

In the recreation we attempted to deal with it as though we were not expecting it. We failed to make the runway on all attempts and ended up pretty much in the same position. There is only about 30 seconds from first recognising a problem to hitting the ground and the first 15 seconds or so is spent trying to evaluate the problem and there is a sense of disbelief that neither engine is responding. The EGTs are normal and uniform and there is no Eicas indication of engine failure.

The low speed protection you refer to is only valid with the autopilot engaged. With the autopilot disengaged there is an autothrottle wake up which would have been of no use as it simply advances the thrust levers which had already been advanced but the engines had not responded.

A flap 25 approach requires less thrust than a flap 30 approach and faced with a loss of thrust with flap 30 extended there is an instinct to raise the flap by one stage. In a normal goaround on the 777 even from 50 feet you still immediately retract the flap from 30-20 or from 25-20 even before you have obtained a positive climb because the sink is negligible and there is a reduction of drag.

In the recreation there is a high rate of descent before touchdown but it is fairly instinctive to initiate the flare early and commensurate with the ground rush rather than the radalt calls. To raise the nose and arrest the descent rate requires a much greater application of up elevator than normal. Approaching the stall the elevator feel makes the elevator artificially heavier but if you pull a bit harder you can still obtain full up elevator and incredibly effect a touchdown IAS of just under 100 knots. We ended up with full up elevator on all touchdowns and I understand this is what happened for real.

Just bear in mind that only one passenger was injured. I believe that was because they got the flare exactly right. Without that the result may have looked like the Turkish 737 at AMS. Well done chaps.

Two factors are involved in such kind of flares: the timing and the available energy. The less you have of the second, the less margin you have for the first. As you said the disbelief most probably contributed in delaying the moment where they ad to admit this would be a deadstick landing. Until the AP disconnection at a very low speed, they probably were probably keeping hope of making a normal ILS. When they finally got the proper situational awareness the speed was already way to low and nose had to be pushed down. Fortunately the plane had just enough height to recover the few knots to neutralize some of the Vz.

I think AMS was a different story, the aircraft made a much more ample flare (it had more energy to do that) leading in a much higher impact attitude at much lower speed. In that case casualties were not caused by the aircraft rate of descent but by its high nose up attitude which led to a violent front part momentum on impact.

S.F.L.Y
if you would finally accept that flaps 30 are just like huge airbrakes then getting away with such brakes is clearly the most important item when trying to stay in the air, especially that at some 400ft is a shade too early for a flare.

You may be right that for some time the crew had waited to see the results from retracting the flaps. But i would not bet too much on the crew expecting to stay on the ILS path and make it to the runway securely as usual.
(1) AP has not turned itself - it was turned out by the control movement from the flight crew. (2) Nose was pushed down and at the same time they have used ailerons+possibly rudder to turn right (in order to miss the concrete runway surface to avoid fire)
They had made a perfect flare - you can see it from the main wheels crushed into the wings - but what is more important is that they had managed to keep the plane flying (arrested speed decay and made typical stall avoidance maneuver to gain the speed). With the vertical speed constantly increasing the higher the plane speed the better was the chance to glide to the ground instead of falling on it. And with higher speed the control movements had better chance of success.

@ maynardGkeynes,

Thanks, movement is indeed what the report uses for system description. This seems to cover the issue when movement is interpreted as "any" movement. Perhaps the fuel cut-off switches (switch position) and spars are also separately recorded parameters.

The report also mentions - quote: "Any uncommanded movement would have been recorded on the FDR and warnings would have been enunciated on the flight deck." - unquote. Regarding the warnings, they would indeed be enunciated, but only if there is a disagree between the spar control relay and spar valve position (stuck valve). Apparently no warnings if the valves, slaved to the control relays (not the cut-off switches), follow relay command if i'm correctly interpreting system operation . . . but i could be wrong.

Anyway, the report is quite definitive regarding movement being recorded so i will leave it to the experts involved with the investigation.


Regards,

Green-dot

This is exactly what I'm saying. Now you seems to be convinced that retracting flaps from 30 to 25 is a much higher priority than to stop the AP pulling on the nose...
On a constant path (ILS) a flap retraction requires an AOA increase, which is opposite to what is required to increase the speed (AOA reduction). Retracting flaps in order to reduce drag at constant AOA would require a steeper path and in order to increase the speed the AOA would need to be reduced. The speed didn't decayed because of the flaps but because the AP was constantly increasing the AOA. That's the first thing to stop in order to preserve energy. Flap retraction from 30 to 25 is only improving gliding efficiency once AOA and speed are under control.

I would not bet a lot on maintaining the ILS path as the flying crew priority ;)

Anyhow, following sequence was used:
1. push the throttles fully up, wait for results
2. repeat push the throttles up, wait for results
3. reduce flaps, wait for results
4. push the nose down, steer to land
5. get the plane on the ground, evacuate all passengers
what would you propose to improve in this sequence? how would it influence the outcome?

Obviously after the second unsuccessful attempt of thrust increase something else had to be done to save speed. Retracting flaps on the ILS is not increasing/saving speed, on the contrary, maintaining the ILS (which btw leads far beyond the threshold) is actually the only reason that caused the speed to decay.

When you fly a glider, speed is not constantly decreasing because of the lack of engine or because of a wrong flap setting. Speed only decreases if you increase the AOA, ie when trying to follow an unachievable flight path.

When the crew detects speed deviation it first acts on the thrust. If this is unsuccessful, there is only one way to save/increase speed: acting on the AOA/flight path. As such, without available thrust, AP disconnection and AOA control is the number one priority, acting on the flaps while the AP keeps increasing the AOA is totally irrelevant. While I understand such a wrong sequence can happen with short reactions delays, I'm really surprised that you can't get it after a couple of days of discussions behind a computer.

175 feet
 

A very interesting discussion . If it's correct that the autopilot wasn't disconnected until 175 feet there would surely have been hardly any time to attempt anything much by way of manual flying.

Are we going off-track here?
 

I'll not throw mud or ridicule any particular poster in this forum.... but there are claims being made which would have a 1st year student of aerodynamics choking on his beer.

Please remember guys that this unique event came at the end of a long sector..and it was an event which was neither expected nor trained-for.

Double-engine roll-back late on approach with inadequate/confusing instrument readings. We've had 2 years now to consider this event - supported by special and update bulletins from the AAIB.

Those guys had less time to diagnose and react than takes most of us to punch a number into our cellphones. I could be in a minority here but my belief remains that the happy outcome more than justifies the fact that, despite their limited very options, they discharged their obligations in a thoroughly professional manner. :D Happy New Year to all bm.

As many previous posts have summarised:

Until the rollback, the a/c was in a configuration of flaps, gear, thrust and AoA that maintained the g/s at the selected speed.

Thrust is reduced, a/p attempts to maintain g/s at expense of speed by increasing AoA while the crew is sorting out what is going on.

The drag from the flaps will accelerate the speed decay resulting from loss of thrust; so, reducing flaps will reduce the rate of speed decay at a cost of having to lower the nose to regain speed.

You then end up substantially lower, but in this case the resultant approach path fortunately happened to remain above the obstacle clearance slope until hostile structures and terrain were cleared.

In gliders a high drag approach is a steep approach. With certain flapped gliders, maximum flaps give you a very steep approach. A good way to get yourself in trouble with such a glider is to select more flaps early on final than will allow you to reach your desired touchdown point because decreasing flaps may require you to lose more height on final than you might have available:uhoh:

In contrast to jets, a glider with max landing flaps will approach substantially nose down until flaring just before hitting the ground -- same applies to a C-172 forced approach with 40 flaps. In both cases the airspeed decays very quickly when the nose comes up.

A jet with max landing flaps approaches nose up courtesy of thrust.

It would be interesting to hear of any simulator landings with max flaps and idle thrust from height.

Suitecaseman, I'll not enter you sarcastic game. We are talking about low speed in a specific configuration. In their situation it was no more a matter of slowing down the speed decay but to immediately stop and reverse it. Reducing flaps on the ILS is not improving a low speed situation, on the contrary.

Do you seriously think the guys at Boeing designed the flap 30 only to induce more drag and burn more fuel on approach? Look at the report and see what happened when the flaps were retracted: the plane went slightly below GP causing even more nose up AP input. The low speed and reduced flap setting immediately triggered the stick-shaker.

If 105 kts is too low with flaps 30, how couldn't it still be too low with less flaps? With a flap 30 config at such speed there is no other way than acting on the AOA to reduce the margin with Vs. Retracting flaps while increasing nose up inputs will only get you closer to Vs and this is why the stick-shaker activated when the flaps reached 25. With 2 pilots in the cockpit you can't tell me that playing with the flaps was the only option they had. If you don't take manual control when you aircraft is about to stall with no thrust then please tell me, when do you think it's time to fly?

In a glider you don't use drag to control speed but slope. Speed is controlled by primary flight controls.

S.F.L.Y
i have asked if you could propose some better actions to take - and what is your porposed action:
sorry but seems that you should rather concentrate on gliding than discussing flying jets or aerodynamics of the wings with flaps.

Looks that You are the only one that does not get it: so for the last time: flaps enabled LONGER stay in the air (aviate!) - if you look at the diagram from Report you would see that after nose-down command the speed has stayed relatively constant but the flight path (repeat - fligh path NOT the ILS path) decayed - the plane started to fly faster toward the ground. If you would draw the Radio Altimeter line appropriate for this speed (last 6 sec of flight) immediately from the point when the flaps were commanded you would see that the plane would have arrived on the ground some 4secs earlier - with much higher vertical speed and 200 meter earlier. Results would be devastating.

Not that the majority is always right but please make an effort of reading once again last 2 days worth of posts. Then discuss your point with anybody you know who flies big jets. And respond if you would still have anything new to add...

suitcaseman - "I suggest you ask your flying school / instructor for a refund. Speed is determined by a combination of pitch attitude (AoA), thrust and aircraft configuration."

So if speed is changed by a/c configuration, and going from Flaps 30 to 25 was a good thing, why weren't more flaps retracted?

misd-agin
 

Flame me if you will.

It's a trade-off. Too much flap retraction at that height and speed, you drop like a stone. Hope you're not a pilot - no offence intended.

My apologies, misd-agin, I see you are a pilot.

WojtekSz, for such a case of thrust deficit, deliberately accepting to let the glideslope drift over would definitely extend the gliding distance.
Not here to judge the BA38 crew as I would be more than happy to have done as well as they did, but purely aerodynamically speaking, SLFY is not wrong.

Mmmayday38, as we have the privilege to have you on board here, would you like to comment on those words, or is it too early ?

suitcaseman - you're losing me. They were at Vref -17/18 kts. Yes, reducing flaps decreases drag. It also decreases lift. If they really wanted to decrease drag Flaps 20 (mid range slats) is a large drag reduction.

You mention engines, etc, etc. In this example they were a glider. Power was low and fixed.

I don't know, and I doubt you know since you havn't published the numbers, if reducing flaps to 25 from 30 was a net improvement in performance in this case. I tend to doubt it and think that maintaining airspeed was more important.

But we've had months to consider this and still don't know the answer. The crew had seconds.

Guys, we all agree that at such a low speed the priority is to stop it from decaying and eventually recover some. Without thrust the only thing to do is to reduce drag. Retracting flaps is definitely reducing drag, but this is only effective when you don't create more drag somewhere else (ie by increasing the AoA...).

In terms of drag reduction retracting flaps might be a good thing, but totally irrelevant when you simultaneously let the AP add more drag by increasing the AoA trying to follow a wrong flight path.

To answer WojtekSz's question, the first thing to do when you realize you have no more available thrust is to recover and maintain the best gliding speed, which you won't achieve through flap retraction. In other words the PF should take manual control, control speed and eventually call the PNF for flap reduction.

WojtekSz, can you tell me why according to you the AP shouldn't have been disconnected before selecting flaps 25? Why maintaining the ILS all this time? To me the AP just spoiled all the benefits of the flap reduction which induced even more drag as the AP had to correct a slight deviation below GP induced by the flap reduction.

What was Sully's first action when he lost both engines? Take manual control or aircraft reconfiguration?

I don't know what Captain Sullenberger did but in the East Midlands B737 accident some years ago I seem to recall that the captain immediately taking manual control dramatically increased his workload and certainly did not help his attempts to deal with the problem with tragic consequences.

In my own company it is a broad rule that using the AP and handing control to the co-pilot is the preferred initial actions which then allows the (usually) more experienced captain to concentrate on managing the problem with as few distractions as possible.

Non-pilots will possibly not appreciate how much mental capacity manual flying involves especially with a non-normal condition.