Asiana flight crash at San Francisco

[Machaca]

kaokao

On July 11 NTSB Chairwoman Deborah Hersman stated:

"There is no anomalous behavior of the autopilot, of the flight director, and of the auto-throttles, based on the FDR (flight data recorder) data reviewed to date."

With over one thousand 777's in operation, the investigators would have immediately informed the operators if a fault were posing a significant risk.

And as several have explained, it is the pilots job to override the automation if needed and fly the plane.

[7478ti]

Qualification includes recency or currency

Concur, ....but I specify "QUALIFICATION", as distinguished from a "License" per se (e.g., some European state's notion of a T/R's purpose) because QUALIFICATION is meant to include Training, Checking (evaluation), and Recency (currency). In addition Qualification addresses the full range of Knowledge, Skills, and Abilities for operation of a specific vehicle and mission for that specific flight (unlike many "Licenses", or a degree such as an M.D. degree, which may not necessarily even "Qualify" a person to perform a particular medical procedure).

I say "virtually" regarding the pilot intent issue, only because of the "outlier" cases of a few historic unfortunate deliberate acts that have in fact been attributed to seasoned commercial pilots (intending harm) over the decades (I can cite specific examples).

[suninmyeyes]

For those who have not flown a 777 and have expressed incredulity and lack of understanding of what could have happened at SFO I would like to explain a few things. I have seen it demonstrated in the simulator and it all happens very quickly.

Being high and fast at 1000 feet led to idle thrust. This led to a decelerating speed.

If you take a snapshot at 500 feet the height and airspeed looked about right. From having been high and fast the crew would have a sense that things were back on track. The biggest clue that things were about to go wrong was most likely the word "HOLD " at the top left of the primary flight display that indicates to those in the know that the autothrottle although engaged will not increase the thrust. Had the word at the top left been SPD (Speed) the accident would not have happened.

From 500 feet the plane starts to go slightly low on the visual profile. As all the private aircraft pilots on this thread have told us if you are low on a light aircraft you add power. That is not the case in a commercial airliner. If you are low you raise the nose and the autothrottle adjusts thrust to control the speed unless it is in hold mode.

If you have gear down flap 30 and idle power and are decelerating below Vref and you raise the nose then the rate of speed decay happens extremely quickly. At this stage if you are not aligned with the runway you inevitably spend more time looking forward and turns involve both ailerons and spoilers and invoke more drag. Apart from practice in the simulator this was probably the first approach either Captain had done where the autothrottle did not maintain the Vref speed. The landing checklist was not completed until 500 feet. The landing checklist in the 777 is electronic and normal airline practice is that when the checklist is complete the non-handling pilot announces it and the handling pilot would then have to look inside the flight deck below 500 feet at a critical time to confirm the checklist was complete. It's a bit like checking your text messages in a car while overtaking. Had this checklist been completed earlier it would have been one less distraction.

So to recap, although at 500 feet height and speed were ok the thrust levers were at idle, the alignment was not right and the nose was then raised with the speed decaying rapidly. It was probably only about 20 seconds after things looked OK at 500 feet that the the aircraft was below 200 feet and the situation was non recoverable due descent rate and lack of speed and low thrust setting. Jet lag does inevitably slow down the recognition process.

I am not in any way excusing what happened, I am just pointing out that it can happen extremely quickly with distractions which is why most 777 operators employ standard operating procedures that involve being stabilised at the right position and speed at 1000 feet with thrust levers back up and speed at vref controlled by the autothrottle and landing checklist complete. None of these happened in the above case which led to a lot of distraction and regrettably in that very busy period it looked like the crew became overloaded and missed the rapidly decreasing airspeed.

As regards changing procedures at SFO. I do not think it is necessary. Thousands of approaches were flown before this without accidents and thousands of approaches will be flown after this without accident. I'd also like to venture that in the unlikely event that the same crew were allowed to fly an airliner again they would never make the same mistake again. An accident is a painful learning experience.

[Lonewolf_50]

Tom: thanks for the reply, not sure why my post was deleted. Maybe it was too long.
suninmyeyes:

If you are low you raise the nose and the autothrottle adjusts thrust to control the speed unless it is in hold mode.

That seems to be the problem. Either the pilot adds power, or the A/T adds power (an automated feature, type dependent) but when low and slow there is a need for power to correct. Retired F4 named this pages ago: it is an energy management matter. Pilot (or an auto feature) adds energy via power (engines). Merely changing pitch does not add energy, it trades energy. If pilots are taught to fly the nose and the plane will take care of the power (because it seems that frequently, it is set up to do just that) then the root cause of an automation induced (assisted?) trap seems to be staring us in the face. (Granted, one can get low and slow in an aircraft without a bit of automation as well ... particularly if distracted ... )

In this respect, flying a visual approach in a heavy is no different than flying a King Air, with the major exception of response time. The pilot in a heavy has to be further ahead of the aircraft than the King Air pilot due to how long it takes to overcome inertia in landing configuration, at approach speeds, in the case that a correction is needed or a Go Around is required.

A similar comparison: driving a PT boat versus driving a Navy Destroyer. Response time / inertia: big difference.

If you have gear down flap 30 and idle power and are decelerating below Vref and you raise the nose then the rate of speed decay happens extremely quickly

From behind the power curve to further behind the power curve ... energy deficit.

Apart from practice in the simulator this was probably the first approach either Captain had done where the autothrottle did not maintain the Vref speed.

Aviation Week asks (July 15 2013 issue) Automation Paralysis?

The landing checklist was not completed until 500 feet.

Do you think this was due to accepting the 180/5 call from ATC?
Someone made a point on that early on in the thread.

[suninmyeyes]

Lonewolf

I do agree that 180 knots to 5 DME was a contributory factor to not being stabilised if they conformed to that. I personally would have either said " unable" or reduced speed at least two miles earlier as has been suggested earlier.

[PEI_3721]

suninmyeyes, thanks for an excellent and thoughtful explanation.

I would add / ask that the differences between previous A320 experience and the B777 may also have had influence, e.g. A320 autothrust when engaged will maintain speed without thrust lever movement, as in this accident.
What are the equivalent correct/incorrect A320 autothrust annunciations?
Does type training pay enough attention to these differences?

Also, the flight technique – raising the nose, should be based on the ‘normal’ picture, the attitude with respect to the real world for the aircraft type; had this pilot sufficient experience on this type to establish a normal picture?

Would the crew have any additional cue of speed error from the control force, assuming it deviated from the trim condition; or does the B777 have an auto trim follow-up system in manual flight?

I would disagree about not changing SFO procedures; previous success is not evidence of safe – acceptable risk operations. Accidents are a combination of factors – “the unforeseen and often unforeseeable concatenation (linking) of diverse events, each one necessary, but singularly insufficient.” (James Reason).

“The landing checklist was not completed until …” We often ‘kill time’, but lack of it can be a real killer; how many occasions did your post refer to time.

[phil gollin]

There is a majority (but not universal) feeling that the ATC advisory "180 kts to 5nm"/4-degree approach was a possible contributory factor.

I am sure that ATC did not impose this as a whim, in fact a couple of people have indicated that it is a common call for noise abatement reasons.

Well, where does that policy come from ?

Is it an airport operating requirement ? Is it just something that they would prefer ?

[Chronus]

asiana crash and the visual trap

Here is a link to a little piece written by Perry Thomas.

The Visual Trap

I just wonder in accepting a visual approach whether the crew breached company SOP, were prepared for it in their landing brief or had been admiring the maginificent scenery which has been commented upon by another contributor to this forum.

It is of course the great Summer season over here in Europe and many charter operators are busy hauling our sun worshippers to wall to wall sunshine destinations, where the likes of PAPI`s may be difficult to read. For those heavy drivers Mr. Thomas` article may be worthy of a quick read.

[Bill Bader]

As regards changing procedures at SFO. I do not think it is necessary. Thousands of approaches were flown before this without accidents and thousands of approaches will be flown after this without accident.

And as several have explained, it is the pilots job to override the automation if needed and fly the plane.

And an inexperienced crew should not be flying 300 people around

I've followed this entire thread, and I've read every post. These three points become painfully clear.

So many people have disagreed, usually allowing for pilot fatigue or a lack of hands-on skill. But, when I buy a ticket to fly, what I DON'T see in the fine print is the following:

"The ticket purchaser agrees that flying an airplane is demanding work, and crews may be fatigued or may lack the ability to manage the aircraft without anticipated computer assistance. Therefore our passengers are reminded that, from time to time, aircraft may stall and crash into the Atlantic Ocean or smash up whilst landing on a bright sunny day in San Fransisco. Should you be on such a flight, and should you thusly be killed or mangled, or should your spouse, parents, or children thusly be killed or mangled, we ask you to bear with us. Thank you for your understanding in this matter."

I think apologists should ask themselves whether such a caveat would seem quite okay and reasonable. Maybe in bright red, on their ticket.

Similarly, other professionals, from your friendly brake-repairman to your surgeon, might also like you to be so understanding. They, too, might be tired or new at the job.

[BuzzBox]

This is an interesting phenomenon in this thread. It was established very early on, within a day or so of the accident that there was nothing "slam dunk" about this approach.

A Squared: I was merely trying to point out that approaches into SFO are not always as 'easy' as some of the jocks on this forum seem to allege.

Having now had the time to trawl back through over 1,000 posts, it seems the flight in question was vectored in down the coast to join left base for 28L. That brings its own difficulties due to the high terrain to the south of the airport. The aircraft might have been 'aligned on final by 14 NM from the runway threshold at 4300 ft', but what speed was it doing? The Flight Aware data seems to suggest it still had a ground speed of around 230 knots. That's not an impossible situation to be in, but nevertheless it presents a challenge in achieving a stabilised approach if not handled correctly. Again, a reasonably fresh crew at the top of their game might not have a problem, but it's a different story altogether for a tired long-haul crew.

Why they didn't recognise the situation earlier and do something about it (ie go around) will no doubt be a focus of the investigation.

I think apologists should ask themselves whether such a caveat would seem quite okay and reasonable.

Bill: Nobody, including me, is 'apologising'. Most of us are merely trying to understand how this happened in a bid to find ways to prevent it happening again.

[Dolfin]

Originally Posted by PEI_3721

I would add / ask that the differences between previous A320 experience and the B777 may also have had influence, e.g. A320 autothrust when engaged will maintain speed without thrust lever movement, as in this accident.
What are the equivalent correct/incorrect A320 autothrust annunciations?
Does type training pay enough attention to these differences?

The really big difference between Boeing and Airbus is that the Boeing gives a tactile feedback by moving the controls and thrust levers, while the Airbus does not give any tactile feedback.

Thus, flying an Airbus, it is required to check always visually, what the bird does => change a parameter, then see on the instruments what the bird does, even (especially) on a visual approach.

That in turn means, as a former Airbus driver, at least this pilot should have been used to checked more often the instruments and so he should have detected the airspeed / N1 / EPR disagreement very soon, just by checking his instruments. Basic jet flying.

Regarding the training issue, see my post #1794.

Just my 2 cents.

[jack11111]

16 yo victum/vechicle investigation?

Something is bothering me terribly: is there going to be an internal investigation of the run over girl...or is it this going to be sweeped under the "FOAM"?

[Dolfin]

SOP

Originally Posted by suninmyeyes

Being high and fast at 1000 feet led to idle thrust. This led to a decelerating speed.

Well, and what should have happened at 1,000ft if the aircraft was too high and too fast?

Most operaters I know call this an unstabilized approach and require the crew to GO AROUND!

Originally Posted by suninmyeyes

The landing checklist was not completed until 500 feet.
...
So to recap, although at 500 feet height and speed were ok the thrust levers were at idle, the alignment was not right and the nose was then raised with the speed decaying rapidly.

The aircraft was not stabilized at 500 ft either, another chance for a GO AROUND missed, not even the checklist was read at 500 ft nor adequate power set!

That has nothing at all to do with not understanding automatics or changing to a new type or having two guys in a new role or on a new seat or lack of a bloody glideslope.
That is nothing less than (multiple) disregarding SOP by not performing a GO AROUND at well predefined points for an unstabilized approach.

And, btw, the Airbus has an electronic landing checklist as well, so the trainee was familar with that.

[galaxy flyer]

Jack11111,

I'm sure there will be, the fire service debriefed any event like. The NTSB will probably include some of the findings in its report. But, by then, it will be page 3 news.

[fire wall]

Enough bullsh1t from general public and 100 hr private pilots
I have flown into KSFO numerous times with 21 years flying Boeing a/c, stated only to advise that this is not ill informed comment. Whilst the traveling public and pvt pilots are entitled to their opinion, this wild speculation does nothing to get to the crux of the accident.

1. The a/c was NOT given a slam dunk approach. Aligned on final at 4000 odd feet at 14 miles is not a slam dunk.
2. 230 kts at 14 miles is NOT a challenge.
3. Stabilised approach criteria was NON existant.
4. The a/c was on profile at 400 ft..... but with a ROD at close to double the stabilised equivalent, way outside the ballpark.
5. The 3 muppets up front said NOTHING. (I won't honour them with the term pilots)
6. A/T hold is a normal A/T sub mode. It only becomes a "trap" when you don't know what you are &^%#& doing.
7. It is preferable that the guys up the pointy end know what they are *&$^ doing !
8. All ULH operations have a fatigue element. It is a professional's responsibility to mitigate that fatigue element by inflight rest, professional competence and knowing your game. The last two items were NON existent.

These morons masquerading as pilots deserve jail time, the charge professional incompetence.

[misd-agin]

14 DME, 4300', 230kts, or even 250kts, is now being presented as some sort of tricky arrival?

It isn't. Period. Routinue arrival trying to save time is 250kts to 10 DME @3000'.

And every long haul crew is tired. Most are lucky if they get 2 hrs of sleep during a long haul overnight flight. Three hours of sleep, unless on a four man crew, is rare.

Four man crews are the most rested. Asiana was a four man crew.

[RoyHudd]

Try 2-crew long haul ops, and you'll know what tiredness is. UK charter companies are specialists in this exquisite form of cost-saving torture. Cheap tickets, exhausted pilots, no rest areas for 3-crew when they happen.

[Punchespilot]

Firewall ,Good post ! True indeed.


On another note , regarding speed brakes , i am wondering what is the likely hood the approach was flown with the speed brakes up ?

[despegue]

Manual flight with autothrottle/ thrust is not a good idea, and I do not practice it, neither does my company. either all ON or all OFF. combining makes things jot only more complex, it confuses both the aircraft and the crew.
It seems more and more likely that this is one of the contributing factors.

Over and over, I see Captains and FO's blindly relying on the autothrottles, never putting their hands on the thrust levers to feel their reaction ( Boeing here) or at least monitoring the thrust trend and value ( on AB).

[A Squared]

Originally Posted by BuzzBox

The Flight Aware data seems to suggest it still had a ground speed of around 230 knots. That's not an impossible situation to be in, but nevertheless it presents a challenge

Yeah, that's about what I see on Flight Aware also. Now I don't fly the 777, but a little googling leads me to believe that 230 knots is slow enough to extend the first two increments of flaps and extend the landing gear.

So, you're saying that being lined up on final 14 mile out, on a 3 degree glide, and slow enough to a) throw out the gear, and b) start throwing out flaps, is a "challenge"?

Maybe I'm missing something, but I'm not seeing the challenge. If you're not high, and you're slow enough to start configuring, and you've got 14 NM to go, it seems like easy money to me.