SFO- Asiana 777 crash

[evaamo 549]

Hi Tom, as someone who has heard horror stories from many of my friends who currently (or formerly) fly for airlines in Asia, I can only say you're 100% spot on! The rigid and almost reverential culture towards senior members in the cockpit makes almost impossible to implement modern CRM. And from what I understand, it's even more so in Korea.

 

I agree with you with regards of it being an accident waiting to happen... I'm almost surprised we don't see this kind of mishaps more often.

[vr-pilot 11]

The A320 has little traps, you must disengage both FD's when hand flying and not following a computed path, especially in OP CLB or DES (Airbus version of FLCH) if you don't switch of the FD you can get yourself into a lot of trouble and trigger Alpha floor.

It is really "confusing", that Airbus FBW hand flying with "FD on" in conjunction with the A/T logic can put you quite easily in a stall/alpha floor (IIRC even overspeed?) situation. The principle "if you do not follow the FD I will behave like a mad cow A/T logic" is one of the system features that should be overcome someday...

 

Back to Asiana 214:

Watching the NTSB video I was very surprised seeing the plane at 1600 ft ALT sinking with -1000 fpm (3000ft MCP GA ALT) and at the same time the PF presses FLCH (while in A/P+A/T mode initiating a go around!). What did the PF think this would help him getting the plane lower and slower? (Perhaps displacement activity?) He disconnects the A/P, holds back the thrust levers and thereby puts the A/T (for sure unknowingly) in a sleep mode.

 

A/T in HOLD mode causing the "Low Airspeed Protection" to fall in a "Snow White slumber" is IMO in fact surprising.

 

The PF nevertheless managed to "VMC-establish" the plane EXACTLY at 500ft AGL with 137kt IAS and full flaps, but unfortunately with idle thrust... (11:27:15)

 

Until 11:27:38 (for 23 seconds!) the PF continued on idle thrust, pitchung up continuously,  until the aural airpseed alert comes up, further 4 seconds it takes until thrust lever advancement initiates... in all 6 seconds to full power then being only 60ft above the ground (engine spool up time?) The "go around call" at 20ft AGL with -700 FT ROD is "in sync" with the impact... :(

 

The PF had 23 seconds time to let the engines spool up. He (and the other three crew members) must have been relying on the A/T while concentrating on the visual approach...

 

Better pilot training/awareness, A/T automation "Snow White slumber" removal and ATC procedure easing (or "ATC procedure denial by flight crews enforcement ")  are my thoughts about what could be learned.

 

Greetings,

Claus

 

BTW: I just remember a landing of mine when I was put behind a B767 with my little C172 less than 2 minutes behind. It was like "land now or prepare to be number three or four on a different runway". So I accepted the controller's "advice" and slammed my C172 from 1000ft at 2 miles final abeam threshold in front of the following B737 traffic. I think also the controller noticed my wings rocking like wild when I was fighting the wake turbulence of the B767 on this beautiful, sunny, hot and windless day...

So "ATC denial" is something one has to learn. My lesson was learned and I still don't know today how I manged my C172 not to make a full roll at 30ft above the runway. I was milliseconds before the decission to support this full roll intentionally as fighting with engine roll momentum and full aileron deflection was at the limit. :wacko:

[Don t moderate me this way 882]

(...)

The PF nevertheless managed to "VMC-establish" the plane EXACTLY at 500ft AGL with 137kt IAS and full flaps, but unfortunately with idle thrust... (11:27:15)

(...)

 

 

Idle thrust implies a not stabilized approach ... and so did the sink rate of about 1.150 - 1.200 fpm at 500 ft.

 

The PF must have known the 500 ft 'gate' was coming, he must have known he was not yet stabilized before passing that gate, he must have had time to mentally prepare for a go-around at the 500 ft gate (the latest) - and so must have had the instructor pilot.

 

I concur with Tom's cultural argument as well; strictly speaking each and every crew member was responsible for a safe landing, IMO psychological/cultural aspects basically prevented each crew member from speaking up - until it was too late.

[vololiberista 205]

It is really "confusing", that Airbus FBW hand flying with "FD on" in conjunction with the A/T logic can put you quite easily in a stall/alpha floor (IIRC even overspeed?) situation. The principle "if you do not follow the FD I will behave like a mad cow A/T logic" is one of the system features that should be overcome someday...

 

Back to Asiana 214:

Watching the NTSB video I was very surprised seeing the plane at 1600 ft ALT sinking with -1000 fpm (3000ft MCP GA ALT) and at the same time the PF presses FLCH (while in A/P+A/T mode initiating a go around!). What did the PF think this would help him getting the plane lower and slower? (Perhaps displacement activity?) He disconnects the A/P, holds back the thrust levers and thereby puts the A/T (for sure unknowingly) in a sleep mode.

 

A/T in HOLD mode causing the "Low Airspeed Protection" to fall in a "Snow White slumber" is IMO in fact surprising.

 

The PF nevertheless managed to "VMC-establish" the plane EXACTLY at 500ft AGL with 137kt IAS and full flaps, but unfortunately with idle thrust... (11:27:15)

 

Until 11:27:38 (for 23 seconds!) the PF continued on idle thrust, pitchung up continuously,  until the aural airpseed alert comes up, further 4 seconds it takes until thrust lever advancement initiates... in all 6 seconds to full power then being only 60ft above the ground (engine spool up time?) The "go around call" at 20ft AGL with -700 FT ROD is "in sync" with the impact... :(

 

The PF had 23 seconds time to let the engines spool up. He (and the other three crew members) must have been relying on the A/T while concentrating on the visual approach...

 

Better pilot training/awareness, A/T automation "Snow White slumber" removal and ATC procedure easing (or "ATC procedure denial by flight crews enforcement ")  are my thoughts about what could be learned.

 

Greetings,

Claus

 

BTW: I just remember a landing of mine when I was put behind a B767 with my little C172 less than 2 minutes behind. It was like "land now or prepare to be number three or four on a different runway". So I accepted the controller's "advice" and slammed my C172 from 1000ft at 2 miles final abeam threshold in front of the following B737 traffic. I think also the controller noticed my wings rocking like wild when I was fighting the wake turbulence of the B767 on this beautiful, sunny, hot and windless day...

So "ATC denial" is something one has to learn. My lesson was learned and I still don't know today how I manged my C172 not to make a full roll at 30ft above the runway. I was milliseconds before the decission to support this full roll intentionally as fighting with engine roll momentum and full aileron deflection was at the limit. :wacko:

 

 

As I said earlier-

 

This is why jet engines should never be allowed to spool down to idle during an approach. They take too long to recover (compared to propellers that immedaitely increase the airflow over the wings) in which time the aircraft has sunk further with ensuing consequences!

 

The aircraft in the article below was producing at least 82% n2 if not more in order to power away safely as it did!

 

http://www.vc10.net/Memories/testing_earlydays.html#GARVM%20at%20White%20Waltham

[3-2-1-Now 126]

Why are people discussing glide slopes and use of AT?

 

1) GS was INOPERATIVE at the ground station - it was a localizer approach, with transition to visual.

 

2) You can use AP until MDA - but not below.

 

I wasn't very impressed with the use of vertical speed autopilot mode so low to the ground with the altitude pre-selected to an altitude above the aircraft - good way to fly into terrain in its own right! If they needed to select higher they should have been hand-flying it (with the FD OFF).

 

After seeing the NTSB video and the modes the crew selected on the flight guidance, it was gross incompetence that caused the crash.

 

Overall, it seems another case of no-one flying the aircraft, with everyone becoming obsessed with a single problem (too high on path).

 

 

 

Nothing wrong with using the automatics to help reduce workload, besides 1000 ft is not that low. It was a visual approach, they could have flown it VNAV or just stuck with V/S or FPA down to 200ft.

 

Using the A/P down to minimums is fine as long as you are ahead of the aircraft managing the energy and using the best modes for the job.

 

This turned from a high energy approach (KSFO seems to be famous for slam dunks)  to being slow, un-spooled and no way near stable approach criteria.  

 

Lots of automation involved in your solution to the problem. "Energy" has little relevence here - they were already doomed before the airspeed even dropped below Vref.

 

Why don't you consider this: the automated approach went to hell in a hand-basket (but could still be SAFELY recovered at 6 nm), so SWITCH IT ALL OFF AND MANUALLY FLY IT?

 

The weather was good, so there was no reason not to.

 

You want to COMPLICATE an already bad situation by trying to program the FMS just 7 nm from touchdown for an approach that was never briefed. A good rule of automation: if the automation isn't doing what you want, get rid of it and hand fly. In the case of Asiana, they were visual with the field! They crashed a perfectly good aircraft in VMC.

 

Let me post this video again:

 

http://www.youtube.com/watch?v=R0rYX-Jn6o8

 

Best regards,

Robin.

[Don t moderate me this way 882]

Best use of (the appropriate level of) automation doesn't necessarily imply reprogramming the FMC.

 

I agree with the hand-flying thingy but Asiana's policy seems to have been different - the NTSB would like to see this policy changing towards more 'hands-on' proficiency.

[FLEX1978 373]

Let me post this video again:

 

http://www.youtube.com/watch?v=R0rYX-Jn6o8

 

Best regards,

Robin.

 

I think that video has been posted to death.... energy management has everything to do it!  They went from high and fast to low and slow. On an NPA approach you configure early.

 

The reason I suggested other methods safely utilizing automation is because the PF stated he was fatigued and uncomfortable flying a visual approach into KSFO.  A simple solution is to go around once unstable and plug in the RNAV (GPS) PRM RWY 28L.   When did I say they should start plugging an approach at 7nm? Where on earth did you get that from?

 

You say you don't like the fact that they are in V/S with the target altitude pre-selected above present alt?  That is standard procedure for an NPA  and happens hundreds of times daily.  Set missed approach alt and remain in V/S. The screw up was reverting to FLCH

 

No offence my friend but try to keep what you think is right from flight sim separate from facts.

 

Fortunately culture is changing, 10 years ago out here in the middle east a missed approach was frowned upon with certain carriers, now all I hear is talk about CRM and DO NOT be afraid to Go-around.  Truth - guy's used to be scared of getting called into the office for becoming unstable and performing a missed approach.

 

I hope certain Asian carriers are also hammering this home to their flight crews.

[Christopher Low 5,359]

 

Overall, it seems another case of no-one flying the aircraft, with everyone becoming obsessed with a single problem (too high on path).

 

That reminds me of Flight 401, when that TriStar crashed in the Everglades.

[3-2-1-Now 126]

No offence my friend

 

None taken, but I know of at least two airlines where you will get busted for using VS with the altitude set such it will never capture. It is extremely bad practise (and very dangerous)!

 

 

try to keep what you think is right from flight sim separate from facts.

 

My comments are not "FS-isms". As Asiana proves (in a very long line of crashes), pilots in the real-world do things that are highly questionable through to out-right dangerous. Your suggestion of using VNAV when they can't even handle a simple localizer-only approach is the wrong idea. They were in VMC, and visual with the field. The altitude deviation was not unrecoverable. They took a small error in altitude and turned it into a crash, because they were screwing with automation instead of flying the aircraft.

 

Why didn't the PF have his hands on the controls below 3000 ft? Did you know that is a requirement? You throw "energy management" and "stabilized approach" around as if they're the only things that matter. Futile, when the failure was in basic airmanship.

 

Best regards,

Robin.

[Werner747 107]

When I looked at the reconstruction, by the time the pilots had switched to FLCH mode, I had already all the hair on the back of my neck standing up!   When by the 4 to 5 nm point already, a good long hard look should have been given and the go around considered.   

 

I think what may have happened here is that the pilots got so rushed by the incorrect approach path that they got more and more desperate to get it back on track, and in so doing started to do some less intelligent things which is why the FLCH error happened to.   Whenever an approach reaches that stage, a go around is the order of the day.   This seems to have been a mild to moderate case of "get-there-itis" in a way.

 

I concur that the pilots should have done a better job, but like others have said, I will not judge or place blame here.   But it remains a clear fact to me that they were simply not applying the principles of CRM and simply not paying attention to flying the aircraft in this critical phase of flight.   That is the one thing that hits me square in the face - who on earth was flying the aircraft during all this???   

 

Looking forward to the final report though...

 

Kind regards

[FLEX1978 373]

None taken, but I know of at least two airlines where you will get busted for using VS with the altitude set such it will never capture. It is extremely bad practise (and very dangerous)!

 

 

 

My comments are not "FS-isms". As Asiana proves (in a very long line of crashes), pilots in the real-world do things that are highly questionable through to out-right dangerous. Your suggestion of using VNAV when they can't even handle a simple localizer-only approach is the wrong idea. They were in VMC, and visual with the field. The altitude deviation was not unrecoverable. They took a small error in altitude and turned it into a crash, because they were screwing with automation instead of flying the aircraft.

 

Why didn't the PF have his hands on the controls below 3000 ft? Did you know that is a requirement? You throw "energy management" and "stabilized approach" around as if they're the only things that matter. Futile, when the failure was in basic airmanship.

 

Best regards,

Robin.

 

Hey Robin,

 

First thing, what is the name of the  777 operators you are talking about? slightly worrying that any any airline would ban flight crew from using Boeing's and every other 777' operators method of flying a NPA.  Of course hand flying is encouraged when 'Workload permits'  slightly hard to believe what you are saying so I would like to ask around about this.

 

upload pic

 

 

 

 

Your suggestion of using VNAV when they can't even handle a simple localizer-only approach is the wrong idea

 

Read again, my suggestion is one of many options that was open to flight crew. Primarily Go-around and give yourself more space and time.  To say it is wrong to use the automatics to assist if available is laughable, I won't even carry on with that part of the discussion

 

 

 

Why didn't the PF have his hands on the controls below 3000 ft? Did you know that is a requirement?

 

You are joking right??

 

Quote -----------You throw "energy management" and "stabilized approach" around as if they're the only things that matter. Futile, when the failure was in basic airmanship-------------

 

Basic Airmanship in a 200+ tonne plus machine is energy management!  obeying company SOP and adhering to stable approach criteria is basic airmanship!   I am not throwing anything around, do you think they still would have switched to FLCH if the aircraft was not high and fast? No, they would have remained in V/S until the MDA. 

 

They let the aircraft get ahead of them 'Bad energy management & airmanship' They then proceeded to select a mode they didn't fully understand and failed to monitor to situation.

 

The PF took manual control of the aircraft and still managed to crash it.  The fact that you think a go-around, reassessing situation and using all available tools to assist is a bad idea clearly shows you have never been involved in any form of airline training or operation. At least I hope that is the case!  :)

 

 

 

 

 

 

.

[Christopher Low 5,359]

It's that cultural problem that needs to be sorted out more than anything else. If anyone on the flightdeck thinks that they can't speak out (or take immediate action) because of "respect" for the senior officer, then you will always have the potential for disaster. That simply can't be allowed to continue when the lives of hundreds of passengers are in your hands.

[KevinAu 441]

While culture has been and probably will always be a crm threat with Asian air carriers, I don't see it as having been a factor in this accident. This looks simply like three guys who got so wrapped up with getting the plane down from a too high position that they just focused all their attention on the runway and vasi and plain dumb forgot to scan the instruments and assumed that the thrust would take care of itself as they were used to.

 

Any pilot could have done what they did. One does not have to be of Asian descent to do what they did. Even at my previous airline, there had been several incidents where guys just got distracted, focused on something or just didn't pay attention and forgot to push up the TLs as the plane leveled off and got into the shaker or pusher. And this was a US airline flying a plane that did not have ATs and these guys were caucasian.

 

Mismanaging the guidance panel modes and setting yourself up for a big fail is also something not unique to Asian peoples. Pilots of any ethnicity are perfectly capable of failing to downgrade the automation and just take over and hand fly the stick and throttles. Especially pilots who are used to the Airbus.

 

I haven't seen anything here that could be uniquely identified as an Asian factor in this accident. It did not appear that any of the three noticed the airspeed and AT issues until it was too late. So there wasn't any issue of somebody noticing a problem but being afraid to speak up. This crash was merely a case of fixation and inattention.

[Tabs 2,790]

It does get recognition, not sure why you would think otherwise? How exactly can ATC be blamed? Nothing about the approach path was dangerous or beyond the capability of the aircraft. The crew mismanaged the aircraft's energy state and ignored warnings. 

 

Everyone one else that day managed to get down safely with the G/S inop.

 

Other 777s from a literal smorgasbord of world airlines landed prior to the accident that morning with the exact same clearances and restrictions and had no issues. I mean, they had a 14 mile final and ended up 34 knots below Vref despite the 180/5 restriction - cmon! It is not difficult to slow a 777 to Vref+5 from 180 in 5 miles while in a high drag state - happens literally all the time at airports all over the world. If the crew were unable to perform that request then it is their duty to 1. inform ATC they're unable, and 2. execute a go-around. 

[KevinAu 441]

Other 777s from a literal smorgasbord of world airlines landed prior to the accident that morning with the exact same clearances and restrictions and had no issues. I mean, they had a 14 mile final and ended up 34 knots below Vref despite the 180/5 restriction - cmon! It is not difficult to slow a 777 to Vref+5 from 180 in 5 miles while in a high drag state - happens literally all the time at airports all over the world. If the crew were unable to perform that request then it is their duty to 1. inform ATC they're unable, and 2. execute a go-around.

Just because the outcome of the previous 99 approaches from the high setup were successful does not mean that there is no problem. All it takes is that 100th to end that lucky streak.

 

One of the initial links of this accident chain was the above glideslope setup from ATC. The next link in the chain was the mismanagement of the guidance modes in response to the above glideslope situation which exacerbated it. All it takes to prevent an accident is to break one link in the chain. ATC was a link in this accident chain and many similar ones. Unfortunately their role in contributing to an accident is often unrecognized. If ATC was more cognizant of the importance of stabilized approaches, then there may be improvements to safety.