OZ 777 Crash At SFO Part 7

kanban · Tue Jul 09, 2013 11:13 pm

Quoting rfields5421 (Reply 79):
Let me say one other thing about cameras.

It's kind of a knee jerk reaction... consider all the flights in and out of all the airports and how frequently would the data be needed.. seldom.. Then someone will want cameras in the cockpit, others watching the FA's.. it never ends.

then General Av will complain they are being monitored as some government personal freedom violation. Actually when they has been an incident, camera footage or not, they seem to get to the bottom if it..

flyingturtle · Tue Jul 09, 2013 11:17 pm

Quoting D L X (Reply 77):

So, stalling at 50' is just about the worst thing that can happen, I suppose?

Was lifting the nose a "last ditch effort" to avoid nose first into the sea wall, then?

I think stalling at 50' or less should be survivable by most of the passengers... but don't ask me about their (spinal) injuries. Worse would be a stall at 200 or 300 ft. Too little height to recover, and too high to pancake on the ground.

Here, the seawall was the problem IMHO, and to avoid slamming nose-first into the seawall was their instinctive priority in this situation. I'm sure we'd talk about a wholly different picture if they didn't "land" so "long".

Regards, David

Klaus · Tue Jul 09, 2013 11:19 pm

Quoting boacvc10 (Reply 85):
I'm impressed with the candor and deep knowhow of the Chairperson in this briefing.

Indeed. She is very guarded against misinterpretation and is extremely deliberate, cautious but still quite open about the actual state of the investigation.

She is doing a highly competent job in an extremely difficult position. Excellent.

Quoting nm2582 (Reply 88):
Would there be any benefit to altering the stick shaker logic, such that it uses some kind of "look ahead" algorithm?

If I understand that correctly, the low-energy warning in the Airbus models attempts something like that (separate from the hard protection which would engage close to where the stick shaker goes off in a Boeing).

matt · Tue Jul 09, 2013 11:19 pm

Quoting cjg225 (Reply 94):
Yes, I thought I heard her say that, but I wasn't sure because I was typing part of an update the moment.

That's simply incredible! Does anyone know where, on that specific 777-200, the jumps seats are located in the vicinity of doors 4L and 4R? How many rear- and forward facing?

Klaus · Tue Jul 09, 2013 11:25 pm

Quoting flyingturtle (Reply 101):
Here, the seawall was the problem IMHO, and to avoid slamming nose-first into the seawall was their instinctive priority in this situation. I'm sure we'd talk about a wholly different picture if they didn't "land" so "long".

I think beyond the low-speed causation the acceleration tracks on the FDR will be highly interesting for the actual crash:
• How much forward speed did the pitch-up and probable stall remove?
• How much altitude did the pitch-up still gain even so (if any)?
• How much forward energy did the MLG and tail impacts on the sea wall dissipate?
• How violent were the multiple contacts to the ground in the forward and aft cabin?
• What were the g-loads incurred by each of the events?

n92r03 · Tue Jul 09, 2013 11:32 pm

Quoting matt (Reply 103):
Does anyone know where, on that specific 777-200, the jumps seats are located in the vicinity of doors 4L and 4R? How many rear- and forward facing?

No, and OZ's website seatmap is incorrect for this 772. We can assume the jump seats are behind the last row of pax seats, which puts them much closer to the damaged pressure bulkhead and hole. Unreal that someone can be ejected and live.

zanl188 · Tue Jul 09, 2013 11:34 pm

New NTSB B-roll...

Looks like a good bit of the nose gear is still in the gear bay... and the door to the forward cargo pit was able to be opened!
I'd expect the door frame to be distorted after the pounding this airplane took....

http://www.youtube.com/watch?v=CqSSrU9L9ok

codc10 · Tue Jul 09, 2013 11:37 pm

Quoting Klaus (Reply 98):

That's not a fabrication, it was written by one of the flight crew on UAL885 on 06JUL and posted to a private, internal UAL message board in which identities are verified and posts are made under real names. Unfortunately, this was duplicated almost in its entirety and shared throughout the web via email, which explains how the Curaçao Chronicle may have obtained it. The version posted was not totally scrubbed, either, as an internal ID appears on the first line of the email.

I would take that analysis as accurate, personally.

boacvc10 · Tue Jul 09, 2013 11:53 pm

Quoting zanl188 (Reply 106):
Looks like a good bit of the nose gear is still in the gear bay... and the door to the forward cargo pit was able to be opened!

Has the other engine been located as yet?

cptkrell · Tue Jul 09, 2013 11:54 pm

Quoting rfields5421 (Reply 79):

Mixed emotions about cameras focusing on a/c ops at airports, cost being my only question. But if tens of thousands of cameras currently record people running through a red light, or ripping off an ATM, would not such cameras be of some importance for an aircraft crash investigation? I dunno. A mishap like this is of course a real but rare tragedy. Wonder what the associated costs would be comparitiively, say, as cameras for entering and exiting the restrooms in the same airport? I'm getting ready to opine, that although very infrequently needed, it could be a positive implimentation.

That said, thanxs Starlionblue (Rep 245) for updating (correcting my misinformation posted about the FA (Ms. Lee?) in the posted photo. They are ALL heroes; weren't even thinking about being heroes, but they are.

Can't let this go... although everybody else did...aaexeplot's Rep 278 (Previous Topic Part) about having 10 years of experience in professionally insuring media releases being accurate, grammatically and spelling correct falls dramaitically after reading three (at least three; I stopped counting) big time spelling errors, and the last statement was...well.

BTW, I watched Brian Williams a while back (don't get me wrong...I didn't have my channel "clicker" here in the compuiter room) who stated that the wreckage would probably remain at the crash site for the remainder of the week. After watching Brian Williams for more than a few minutes, I usually have an uncontrollable desire to take a shower because he is so smarmy, but if the report is true that will answer a couple of previous questions. best regards...jack

fpetrutiu · Tue Jul 09, 2013 11:55 pm

Quoting boacvc10 (Reply 108):

Has the other engine been located as yet?

It came to rest in between the runways about 500 ft from the 2 o'clock position from the nose of the aircraft.

prebennorholm · Tue Jul 09, 2013 11:59 pm

Quoting PW100 (Reply 75):
Initially, this crash initially seemed to me like the LHR 777, but right now, it is quite frightening how much this is looking like the TK 737 at AMS.

Agree. There are striking similarities to the TK 737 accident.

But there are also differences. The TK plane had one radio altimeter MEL'ed. The crew was unaware that the faulty altimeter reading was fed to the autopilot. The OZ crew was flying a perfectly good plane.

But there is one more similarity: Both flights were training flights with the instructor as PNF. In both cases the PNF didn't do their job - they didn't watch the speed. Way too late did they realize that they bled off their Vref down to (near) stalling speed.

It raises the question, at least for me: Is it too much to combine the two jobs, instructor and PNF? Too much workload when things go abnormal?

Were both instructors fully occupied monitoring the performance of their pupils, while forgetting their job as PNF?

Would it be better to have the instructor in the jump seat and a "full time" experinced PNF in the front seat? At least during take-off and landing?

Aaron747 · Wed Jul 10, 2013 12:02 am

Quoting nm2582 (Reply 88):
At the point the stick shaker activated in THIS flight, the flight was already doomed. They had no altitude, they had no airspeed, and the only source of energy was several seconds away from producing any usable thrust (and then many more seconds away from rescuing them from the situation they apparently flew themselves into).

The stick-shaker is partially intended as a last resort, idiot-proof stall recognition/warning system.

The issue in play here is the severe airspeed decay that put them in that position, and the question is why it was allowed to occur.

Beyond the fact that there are two different sources of airspeed information on the Primary Flight Display (the vertical bar shows trend, the box shows the live readout), there are other signs of low airspeed. Even though the 777 is FBW, based on my understanding there is adequate "feel" in the controls. Sluggish control response and labored inputs are usually a sure sign of airspeed decay approaching the stall envelope.

For those reasons and more, the stick-shaker functions as I described above - the last resort face-slap to wake up. There are certainly situations, such as takeoff scenarios, where stall may be quickly approached via aggressive AoA, and in those cases I guess the stick-shaker is a more immediate recognition tool. But based on what we know so far, this accident is really a question of who was minding the store, or not, and why.

[Edited 2013-07-09 17:04:40]

Viscount724 · Wed Jul 10, 2013 12:07 am

Quoting boacvc10 (Reply 108):
Has the other engine been located as yet?

Photos of the other engine were published very soon after the accident.

Klaus · Wed Jul 10, 2013 12:10 am

Quoting CODC10 (Reply 107):
That's not a fabrication, it was written by one of the flight crew on UAL885 on 06JUL and posted to a private, internal UAL message board in which identities are verified and posts are made under real names.

Okay, that sounds like plausible validation. Having it leaked from an internal board instead of it being publicized directly also sounds like a more plausible context.

Quoting prebennorholm (Reply 111):
The OZ crew was flying a perfectly good plane.

Could be, but we don't actually know that yet. The NTSB will have a lot of validation work before them prior to them being able to make that conclusion.

Without any technical issues being an excuse for lack of monitoring, of course, if that was in fact the case.

[Edited 2013-07-09 17:10:53]

Moose135 · Wed Jul 10, 2013 12:16 am

Quoting prebennorholm (Reply 111):
It raises the question, at least for me: Is it too much to combine the two jobs, instructor and PNF? Too much workload when things go abnormal?

Were both instructors fully occupied monitoring the performance of their pupils, while forgetting their job as PNF?

Well, you pretty much want the IP in the other seat with full access to the controls when you have someone who hasn't been fully checked out yet in the other seat. And part of the IP's job of "monitoring the performance of their pupils" consists of watching what the airplane is doing - airspeed, altitude, attitude, configuration, and the like.

Starlionblue · Wed Jul 10, 2013 12:22 am

Quoting nm2582 (Reply 88):
Would there be any benefit to altering the stick shaker logic, such that it uses some kind of "look ahead" algorithm? If you take current airspeed, attitude, and configuration (flap setting, power setting, gear up/down, etc.) you can predict many seconds into the future when the aircraft would be on the threshold of stall. You also know the current power setting and how long the engine takes to spool up from the present power setting; so you can determine how much warning is needed to accommodate the engine spool up time. Essentially, the logic would be: "if engine spool-up time plus some safety factor is greater than the time to stall in current config, then activate the stick shaker".

This idea is predicated upon my assumption that stick shaker logic is fine at altitude (if you encounter the shaker at 15,000 ft, ease up on the stick, gain some speed, and you're fine) but perhaps not quite enough warning on final.

In theory, sure. However I think the number of false alarms you get would be large. The positional precision required is very high if nothing else, and GPS is not quite as magical as most people think. There can be errors.

You also risk confusing pilots.

IMHO, pilots have plenty of tools to warn them already. If those tools aren't giving them a hint, they are not going to listen to one more warning. Look at what happened to AF447. Plenty of information available that would have helped the crew, but they consistently discarded it since it did not support their hypothesis.

Quoting ABpositive (Reply 91):
Why not place the camera in the cockpit and store data in the VDR box. This way there is audio and video evidence for investigators and cameras are small enough now to be placed in cockpits unobtrusively.

The DFDR and CVR already give lots of info. This would seem quite intrusive. Given that airlines already use data to nail pilots who don't perform, what could they do with this? At least that's what pilot reaction to this sort of plan has been to date.

Quoting Aesma (Reply 81):
I think it's not about the attitude but about the gear, you don't retract it while you're still descending, since it might be needed to bounce off the runway.

Typical procedure would be to initiate a go-around and ask for gear up and the first notch of flaps. The pilot monitoring would retract the gear, but importantly ensure first that rate of climb was positive in case the plane sinks so much it hits the runway.

Quoting cloudboy (Reply 2):
We know low speed was a critical factor in this crash. Media reports make it sound like the aircraft was so slow that it was nearing stall speed - is that the case? We know it is well below normal landing speed, but the aircraft was not in, nor was it even near, stall speed, correct? I

Vref, the reference speed to be used on final until crossing the threshold, is Vs0 (stall speed in landing configuration) times 1.3, at least in a low-wind situation with a clean runway. Given Vref as 137, stall speed was 105. They were at stall speed. Also, even just above stall speed the controls start to get very mushy and the plane feels sluggish due to low and disturbed airflow. If you pull up at these speeds, even if you don't get a clear stall, the plane will not change direction but just keep on going..

Quoting cloudboy (Reply 2):
If that is the case, from a pilots point of view, how much weight do they place on airspeed vs approach angle and rate of descent while landing?

That stuff is of paramount importance.

Quoting Braniff747SP (Reply 3):
Apparently the stick shaker went, so that would point the airplane to being right above stall speeds.

Indeed. The stick shaker will activate at least 5% above stall speed.

Quoting cloudboy (Reply 5):
Where were the relief pilots at the time? We're they actually in the cockpit at the time, or were they upstairs in the rest cabin.

If they weren't in the cockpit, they were probably in jump seats of pax seats. The crew rest might well not be certified for occupation during take-offs and landings.

Quoting moderators (Thread starter):
Amazing stories. Hopefully helps dispel the notion that Asian carrier FAs overly focused on service and not sufficiently trained on safety. Service and safety are not mutually exclusive. These FAs are heroes!

Quite right. I recall stories about SQ006 flight attendants running back into the burning plane to save pax, including Irene Ang, who died while doing so.

Quoting slider from thread 6
LH was saying that the steep glide slope was due to potential noise restrictions and it leads their locations for aborted landings.

The glide slope is a bog-standard 3 degrees.

Quoting frmcapcadet from thread 6
Also would the pilot at the last few seconds nosed up to avoid a head-on collision with the seawall? This would seem to be a desperate but correct move??A desperate and incorrect move. At such low speeds on the back of the drag curve all pulling up does is change the pitch angle without changing the flight path angle. The only way to extend the glide is to pitch forward since best glide speed (for longest glide) is faster.

Quoting frmcapcadet from thread 6
But this plane was sinking at a faster rate than SOP IIRC, so it what point should they have increased engine power, and how many seconds would it take to recover altitude?

There's a reason DH for a full ILS is 200 feet AGL. It takes a while for the engines to spool up as you say. However more importantly the go around procedure assumes the speed is around Vref. If the speed was 20 knots below you'd need way more height, I'm guessing a couple hundred feet.

Quoting zkojq from thread 6
they deserve every cent of their pay check. Probably a promotion too. And they did all this while wearing reasonably low skirts which I thought would be impossible to run/climb over stuff in (are these factors considered when the uniform is designed?).

Presumably. In any case skirts cannot be too restrictive in normal service either since F/As do a lot of squatting and bending over during meal service.

Quoting flyingturtle from thread 6
What are the advantages/disadvantages of having the autothrottle on during a visual approach?

Reducing workload? Can one simply aim for the touchdown point (while respecting a healthy sink rate, though), while the autothrottle just cares about the right airspeed?

Reduced workload indeed. Let the autothrottles worry about airspeed.

Quoting tp1040 (Reply 44):

The ILS was out of operation and the effect , if any, on the events is not yet obvious.

Not really germane to the crash, is it common for it to out at a major airport for such a long time.???

I know there are different systems in place to safely land a plane, but the ILS being out of service makes it one less tool a pilot has to safely land a plane. Seems like an unnecessary risk and poor management by the FAA and the airport. As flier, I want all the resources available, up to date and working.

It is not uncommon. It was only out on that one runway. Also, on a clear day the ILS should not be needed by a trained pilot. It doesn't really add to risk if it isn't there.

[Edited 2013-07-09 17:27:08]

Arabear · Wed Jul 10, 2013 12:36 am

Quoting PW100 (Reply 75):
Let a crew fly an fully visual manual SFO 28L approach (no ILS, no automation), and you will find that they will have no problem whatsoever doing just that. Let them do it a hundred times, and they will be a hundred times successful - some landings would be better than others, but they succeeeded all 100 times. Now let them do it a million times (remember virtual experiment, thought exercise), and you're pretty much guaranteed to see several, if not dozens of crashes. Now, repeat that same thought experiment with a 1000 different crews, and you will most likely see that every crew will crash more than once. Some crews will have dozens of crashes, some will have just a couple.

As a statistician, I'm not sure I follow the probability logic here. I'm not sure I can follow the logic that 100% success in 100 trials necessarily leads to less than 100% success in 1MM trials -- unless you are assuming that crashes are probabilistically random (which they're not: this is not a flip of a coin or "roll" of a million-sided die). Indeed 100% success in 100 trials is equally good evidence that you might expect 100% success in all subsequent trials (or maybe you have a coin with two heads or two tails

Maybe historical data on non-ILS landings and/or non-ILS airports, weighted for improvements like CRM and in other factors across time and controlling for non-contributory factors, is what is needed to make some sort of "crash rate per million due to no ILS."

It also feels as if the overall probability of an accident of any sort is assumed to be much higher than it actually is in this example.

For comparison, there are over 7,000,000 landings every year just in the USA (granted many with ILS) and there hasn't been a crash since 2009 (although crashes are now so rare in the developed world that they present a probability problem in and of themselves). In the past 5 years, there were 2 years with no hull losses in the US.

B777fan · Wed Jul 10, 2013 12:44 am

Quoting mayor (Reply 7):
Excuse this old ramp rat for a stupid sounding question, but if the ILS was indeed, out, what difference would it make if the wx was severe/clear?

Somebody has probably already responded, but the answer is NONE!

With three pilots in the cockpit, this is beyond belief. This is worse than the three pilots in the Air France cockpit mushing their plane into the ocean from 40,000 feet. At least in the Air France case, they initially had a real, if minor problem that they couldn't manage to solve.

(edited for number of pilots in the cockpit)

[Edited 2013-07-09 18:20:54]

Viscount724 · Wed Jul 10, 2013 12:46 am

Quoting B777fan (Reply 119):
With four pilots in the cockpit, this is beyond belief.

Three pilots were in the cockpit during the landing. The relief captain was in the passenger cabin.

Klaus · Wed Jul 10, 2013 12:58 am

Quoting Starlionblue (Reply 116):
The glide slope is a bog-standard 3 degrees.

For all runways in all directions and for all approach paths?

Further information from the NTSB press briefing:
http://www.c-span.org/Events/NTSB-Br...a-Airline-Plane-Crash/10737440372/

Further post-crash cockpit observations by the NTSB:

• autothrottle switches in the ARMED position

• Flight Director ON for the right seat

• Flight Director OFF for the left seat

• All 3 fire handles were EXTENDED (detached right engine had a fire due to a ruptured oil tank apparently spilling on hot surfaces - no statement yet whether any of the fire bottles were actually discharged; both engines and the APU detached from the fuselage)

• Flaps set to 30

• Speedbrake lever was DOWN

Preliminary information from pilot interviews according to the NTSB:

• PF/FO/trainee: left seat (10 legs on the 777, about 30 hours on type; Rated on: 737, 747, A320 (captain and instructor), 777)

• PNF/PIC/instructor: right seat (3000h 777, 1000h PIC, prior 10y korean air force experience)

• both active pilots flew together for the first time

• Relief FO was in a cockpit jumpseat

• Relief captain was seated in the cabin

PNF recollection of events as reported by the NTSB:

• were somewhat high at 4000'

• set vertical speed mode 1500'/min

• realized at about 500' that they were low

• saw 3 red 1 white PAPI

• told PF to pull back

• had set speed at 137kn

• assumed that autothrottle maintained speed

• between 500' and 200' they had some lateral deviation

• trying to correct

• at 200': 4 red PAPI, airspeed in the hatched area on the speed tape

• recognized the autothrottle was not maintaining speed

• established go-around attitude

• went to push the throttles forward but the PF had already done that at this point

• after impact the aircraft "ballooned" and yawed left

• went into 360° spin

• FO received hospital treatment due to a cracked rib

• other pilots apparently not injured

aaexecplat · Wed Jul 10, 2013 12:59 am

Quoting mach92 (Reply 117):
From a United pilot friend.

Wow. Just wow.

coronado · Wed Jul 10, 2013 1:01 am

Quoting mach92 (Reply 117):
One of the first things I learned was that the pilots kept a web-site and reported on every training session. I don't think this was officially sanctioned by the company, but after one or two simulator periods, a database was building on me (and everyone else) that told them exactly how I ran the sessions, what to expect on checks, and what to look out for.

I wonder if ''cheat'' sessions for pilots are still taking place in Asia. A bit disconcerting to say the least.

katekebo · Wed Jul 10, 2013 1:12 am

So if autothrottle was armed and the speed was set to 137 knots, any explanation to why did the autothrottle fail to maintain speed? Any pilots knowledgeable in 777 could shed some light on this issue?

Mir · Wed Jul 10, 2013 1:13 am

Quoting nm2582 (Reply 88):
Would there be any benefit to altering the stick shaker logic, such that it uses some kind of "look ahead" algorithm? If you take current airspeed, attitude, and configuration (flap setting, power setting, gear up/down, etc.) you can predict many seconds into the future when the aircraft would be on the threshold of stall. You also know the current power setting and how long the engine takes to spool up from the present power setting; so you can determine how much warning is needed to accommodate the engine spool up time. Essentially, the logic would be: "if engine spool-up time plus some safety factor is greater than the time to stall in current config, then activate the stick shaker".

The fly in that ointment is that you'd have no way to predict what the crew might do with the controls, and that's going to be a big part of how close you are to stall because it quickly changes the configuration of the airplane. If the crew is five seconds away from stall, but then pitches up and are almost immediately two seconds away from stall, it still might be too late once the warning is triggered.

Quoting prebennorholm (Reply 111):
But there is one more similarity: Both flights were training flights with the instructor as PNF. In both cases the PNF didn't do their job - they didn't watch the speed. Way too late did they realize that they bled off their Vref down to (near) stalling speed.

It raises the question, at least for me: Is it too much to combine the two jobs, instructor and PNF? Too much workload when things go abnormal?

Were both instructors fully occupied monitoring the performance of their pupils, while forgetting their job as PNF?

I would think that an essential part of monitoring their pupils' performance would be monitoring how they're flying the plane.

-Mir

Klaus · Wed Jul 10, 2013 1:14 am

Quoting mach92 (Reply 117):
Subject: Low-down on Korean pilots

There's obviously some disgruntlement showing in this report, but even admitting for a certain amount of skew there might still be cause for another hard look at the safety culture there, also in combination with the new information from the NTSB briefing above.

Quoting mach92 (Reply 117):
You just can't change 3000 years of culture.

At least not easily or quickly, and likely not at all if airline management isn't fully behind it.

orbital · Wed Jul 10, 2013 1:14 am

I have a couple of thoughts to offer.

The NTSB has not ruled out a malfunction of an aircraft system. In particular, the request for recent 777 approach data from the FAA, coupled with the rather precisely worded statement that "the auto-throttle was armed" indicates there is at least sufficient cause to analyze what the control system in question was actually doing, and why. I imagine there to be an army of engineers from the NTSB and all relevant manufacturers studying this closer than anyone has let on. It is also clear that the relevant pilots (the two at the front) fully believed the A/T was managing the speed, and they did actually recognize the equipment wasn't working, albeit too late.

Now it is up to the NTSB to determine if the pilots made a collective error as to the A/T usage was correct or not, or if there was a systemic failure of the control system that is yet unknown and certainly unexplained.

As to the caliber of Korean pilots, there any number of people who believe there is a cultural influence as to the cockpit management, or proficiency. There are legitimate questions to be asked as to underlying failures in the piloting. But I view most of the comments thus far as simply racist. The fact of the matter is that there is no ruling on this accident. The pilots did not recognize a failure of the A/T, due to user error or system error, for roughly a minute of the descent, but the evidence is conclusive that they DID recognize it. Just too late. And what is the point of cockpit automation? To reduce workload. Reducing workload inherently means looking away from a gauge for a few seconds.

Think about that. The Air France pilots had a lot more time to diagnose what was wrong, and never did (much worse conditions, granted).

When the NTSB squarely blames pilot error, then we can go into the training practices, deficient English, and hierarchical society factors that impact cockpit management. Until then, it is just garbage to even be posting emails from buddies stereotyping all korean pilots, as though it is relevant to this particular crash at this particular stage of the investigation. There are truths to the post I am referring to, I am sure. But those truths can't be applied to these pilots at this stage.

tockeyhockey · Wed Jul 10, 2013 1:20 am

the general consensus seems to be that over-reliance on automation was a contributing factor to this crash. i guess i agree. but i also wonder if there simply isn't an alternative to over-reliance on automation at this point in aviation history. you have thousands of jets in the sky at any moment across the world, and you have the highest demands for safety of any form of transport (or frankly, any activity) in human history. swimming is more dangerous than flying. walking is probably more dangerous than flying.

and why is that? because of incredibly high levels of automation in the cockpit. to say that all the pilots in the world should be stick and rudder experts above all else at this point is simply nostalgic nonsense, especially when managing the highly complex data systems in the cockpit is a job in and of itself.

i remember the pilots comments after the engine explosion on the qantas a380 a while back. he wasn't flying the plane after that. he was managing an incredibly complex system of warnings, beeps, and data processesing on a massive mainframe computer. but if he hadn't been adept at that task, he might have lost all souls.

that's just my two cents on the seemingly prevalent group-think that's going on here about a reliance on automation. it's not as bad a thing as people say it is. and furthermore, it's inevitable and necessary.

rlx01 · Wed Jul 10, 2013 1:22 am

Quoting aaexecplat (Reply 122):

Yeah, meanwhile, ICAO rates Korea as the country with the highest aviation safety standards, including its pilot training standards, in the world.

So... who do we believe? Or maybe we stop focusing on how these guys were Korean, since it's irrelevant.

No one's questioning the nationality of DHC-3 pilots.

hivue · Wed Jul 10, 2013 1:32 am

Quoting orbital (Reply 127):
they did actually recognize the equipment wasn't working

There's no evidence so far of any equipment malfunction.

Quoting orbital (Reply 127):
the relevant pilots (the two at the front) fully believed the A/T was managing the speed

If they'd bothered to look at the airspeed indicator they would have lost their faith pretty quickly.

hivue · Wed Jul 10, 2013 1:35 am

Quoting tockeyhockey (Reply 128):
i remember the pilots comments after the engine explosion on the qantas a380 a while back. he wasn't flying the plane after that.

IIRC significant airmanship was required for the actual landing.

sunrisevalley · Wed Jul 10, 2013 1:36 am

Quoting D L X (Reply 48):
Reports said that many survivors had "road rash" that indicates they slid along the runway, thus having been ejected from the plane

Or being physically dragged to safety...

aaexecplat · Wed Jul 10, 2013 1:37 am

Quoting rlx01 (Reply 129):
Yeah, meanwhile, ICAO rates Korea as the country with the highest aviation safety standards, including its pilot training standards, in the world.

So... who do we believe? Or maybe we stop focusing on how these guys were Korean, since it's irrelevant.

No one's questioning the nationality of DHC-3 pilots.

I don't care one lick about ICAO ratings. IF the story is indeed true that the UA pilot tells, then ask yourself if you want someone like the crew he failed to pilot a plane your family sits on.

I have high expectations from guys piloting panes seating hundreds of pax. Not being able to shoot VFR approaches in CAVOK is unacceptable to me. And IF true, would have me avoid OZ and KE for sure.

Klaus · Wed Jul 10, 2013 1:40 am

Quoting orbital (Reply 127):
The NTSB has not ruled out a malfunction of an aircraft system. In particular, the request for recent 777 approach data from the FAA, coupled with the rather precisely worded statement that "the auto-throttle was armed" indicates there is at least sufficient cause to analyze what the control system in question was actually doing, and why.

The point of interest appears to be exactly which modes and situations permit the autothrottle on the 777 to be armed but at the same time still not active!

Quoting orbital (Reply 127):
There are legitimate questions to be asked as to underlying failures in the piloting. But I view most of the comments thus far as simply racist.

I'd be cautious about that judgment. If the above quoted mail is actually based on factual experience and is not willfully distorted due to racist preoccupations of its original author, it may need looking into – but given that there had been an earlier suspension threat against the two korean airlines on grounds of a lacking safety culture, it seems highly likely that this suspension threat will be revisited in light of this investigation, and I doubt that it would be due to racist prejudices by the FAA or EASA.

Quoting orbital (Reply 127):
The fact of the matter is that there is no ruling on this accident.

Indeed: Not yet.

I need to point out that my post on the NTSB briefing above is a non-sanctioned partial transcript of advance information from the NTSB which is not validated by the other sources yet (such as the FDR and CVR).

And that we all are just speculating here, even with some of us having substantial background knowledge on the matters (which does not apply to myself – I'm just an interested civilian!).

Quoting orbital (Reply 127):
The pilots did not recognize a failure of the A/T, due to user error or system error, for roughly a minute of the descent, but the evidence is conclusive that they DID recognize it. Just too late. And what is the point of cockpit automation? To reduce workload. Reducing workload inherently means looking away from a gauge for a few seconds.

Not in every critical phase of flight – and final approach certainly is a highly critical one where attention must be prioritized very tightly.

Not that this is conclusive yet, but they might have had a fatal combination of mistaken assumptions about autothrottle behaviour very close to the ground on final approach and being too preoccupied with their lateral correction to realize their mistake in time.

Quoting orbital (Reply 127):
When the NTSB squarely blames pilot error, then we can go into the training practices, deficient English, and hierarchical society factors that impact cockpit management. Until then, it is just garbage to even be posting emails from buddies stereotyping all korean pilots, as though it is relevant to this particular crash at this particular stage of the investigation. There are truths to the post I am referring to, I am sure. But those truths can't be applied to these pilots at this stage.

The anger is palpable in that quoted mail, but the claims seem to be primarily based on verifiable or falsifiable(!) facts, not on racist stereotypes.

These claims can be tested. And it may be necessary to do just that.

Quoting tockeyhockey (Reply 128):
the general consensus seems to be that over-reliance on automation was a contributing factor to this crash.

Quoting tockeyhockey (Reply 128):
that's just my two cents on the seemingly prevalent group-think that's going on here about a reliance on automation. it's not as bad a thing as people say it is. and furthermore, it's inevitable and necessary.

There is a fine but crucial distinction:

The problem is not primarily an overreliance on automation, but a false reliance on automation.

Meaning that it is not a problem relying on automation which you understand correctly, including its capabilities, limitations and failure modes.

But it is a big problem relying on the automation supposedly doing things which it actually doesn't provide.

Anybody working with potentially dangerous automatic systems has a supreme responsibility to know as fully and as precisely as feasible what that automatic system can actually do, what it can't do, how it can fail and how one can cope with the expected failure scenarios.

Only having a fuzzy imagination about these matters is utterly inexcusable when the lives of people are at stake.

And at that point the manufacturers of such automatic systems come into play, because they must clearly and exhaustively document the behaviour of these systems, permitting full training of the operating personnel.

If one of the automatic systems fails or is mishandled which I've specified, designed or implemented, people may just be annoyed or inconvenienced. I have a pretty good understanding of how vastly more extensive and expensive my own work would have to be to satisfy aviation standards – and still from my own experience how difficult it can be to train people on the proper use of complex combinations of layered and interrelated automatic systems.

But this challenge needs to be met by those people being trained as well.

[Edited 2013-07-09 18:57:17]

Viscount724 · Wed Jul 10, 2013 1:57 am

Quoting rlx01 (Reply 129):
Quoting aaexecplat (Reply 122):

Yeah, meanwhile, ICAO rates Korea as the country with the highest aviation safety standards, including its pilot training standards, in the world.

So... who do we believe? Or maybe we stop focusing on how these guys were Korean, since it's irrelevant.

However, it's interesting that KE is the only carrier to have written off 5 747s in accidents, including the 742 shot down in 1983 after straying off course into Soviet airspace. Total of 515 fatalities in those accidents (one was non-fatal, a 744 landing accident at Seoul in 1998.)

Not sure, but KE may also be the only major carrier to have had 2 scheduled flights shot down. Also the 707 that went hundreds of miles off course between CDG and ANC in 1978 and was shot down when it strayed over Soviet territory. Fortunately it was able to make a successful forced landing on a frozen lake in northern Russia but 2 passengers were killed from the missile strike.

[Edited 2013-07-09 19:01:03]

Aaron747 · Wed Jul 10, 2013 2:03 am

Quoting katekebo (Reply 124):
So if autothrottle was armed and the speed was set to 137 knots, any explanation to why did the autothrottle fail to maintain speed? Any pilots knowledgeable in 777 could shed some light on this issue?

Based on readily available commentary from 777 pilots, there are some combinations of hold modes and MCP settings that will inhibit active autothrottle.

There is also this in the 777 FOM:

Note: During a descent in VNAV SPD, the autothrottle may activate in HOLD mode and will not support stall protection

Regardless, the fact remains the PFD is there for a reason.

[Edited 2013-07-09 19:04:13]

quiet1 · Wed Jul 10, 2013 2:12 am

Quoting Klaus (Reply 98):
Related to this: I have not heard of the pilots participating in the rescue efforts. I'm fully aware that one of the pilots was himself in need of assistance which his colleagues probably provided and all of them were also likely traumatized to some extent. I certainly would not insinuate any post-crash mishandling of the situation on their part without actual evidence. But are there any reports about the pilots after the fact?

It appears at least two of the pilots were active in the evacuation. Quotes from an article about Cabin Director, Ms Lee's account of the crash and the evacuation:

"In the process, a slide on the right side of the plane inflated into the plane during the crash landing, and Lee and the captain helped a fellow flight attendant injured when buried under the slide."

"Although a fire ignited in the back of the plane, Lee and the captain contained the fire with fire extinguishers and evacuated the remaining passengers."

"After evacuating all the passengers, Lee along with the co-pilot and the injured flight attendant were the last to leave the plane."

source: http://english.khan.co.kr/khan_art_v...?artid=201307091648227&code=710100

rlx01 · Wed Jul 10, 2013 2:14 am

Quoting Viscount724 (Reply 135):
However, it's interesting that KE is the only carrier to have written off 5 747s in accidents, including the 742 shot down in 1983 after straying off course into Soviet airspace. Total of 515 fatalities in those accidents (one was non-fatal, a 744 landing accident at Seoul in 1998.)

I believe the entire system was revamped after KE801. Which actually bares some similarities to OZ214.

I don't think you can really blame them for KE007.

Klaus · Wed Jul 10, 2013 2:16 am

Quoting quiet1 (Reply 137):
It appears at least two of the pilots were active in the evacuation.

Ah, good. That sounds pretty much as it should be.

ltbewr · Wed Jul 10, 2013 2:24 am

Quoting CODC10 (Reply 107):
Quoting Klaus (Reply 98):That's not a fabrication, it was written by one of the flight crew on UAL885 on 06JUL and posted to a private, internal UAL message board in which identities are verified and posts are made under real names. Unfortunately, this was duplicated almost in its entirety and shared throughout the web via email, which explains how the Curaçao Chronicle may have obtained it. The version posted was not totally scrubbed, either, as an internal ID appears on the first line of the email.I would take that analysis as accurate, personally.

One issue with this statement is if it came from a Pilot/FO and didn't go through UA corporate and legal clearance, then the author could face being fired or other discipline for violating corporate rules as to talking to the press as to your work without official ok from the employer. That means someone is lying who they are or is going to be in trouble.

Quoting Klaus (Reply 121):
• both active pilots flew together for the first time

Quoting Klaus (Reply 121):
• assumed that autothrottle maintained speed

To me these are red flags that stands out from the initial review comments. By having a mismatch of pilots, you raise issues of coordination, nuances of understanding of each other, possible errors. On the other point I highlighted, remember never assume, there should have been cross-checks to make sure the setting was right for such a critical component.

7BOEING7 · Wed Jul 10, 2013 2:24 am

Quoting tockeyhockey (Reply 128):
because of incredibly high levels of automation in the cockpit. to say that all the pilots in the world should be stick and rudder experts above all else at this point is simply nostalgic nonsense, especially when managing the highly complex data systems in the cockpit is a job in and of itself.

Automation doesn't cover everything and the day you want a good stick and rudder guy is the day the s**t hits the fan. It wasn't the automation that allowed Al Haynes and his crew to make a "landing" in an Iowa cornfield (without even a stick and rudder) or the Qantas crew to get their severely wounded A380 back on the ground safely. Flight crews with excellent "button pushing skills" but poor stick and rudder would have lost both those airplanes.

[Edited 2013-07-09 19:25:26]

BEG2IAH · Wed Jul 10, 2013 2:28 am

Quoting Klaus (Reply 134):
The anger is palpable in that quoted mail, but the claims seem to be primarily based on verifiable or falsifiable(!) facts, not on racist stereotypes.

I would strongly disagree. I studied with Korean, Chinese, Thai, and Japanese colleagues in a grad program here in the US. The below description is absolutely true and it applies to all Korean students I studied with. I shared my office with two of them and I can assure you that this UA captain is being politically correct. It's even worse.

Quoting mach92 (Reply 117):
First off, their educational system emphasizes ROTE memorization from the first day of school as little kids. As you know, that is the lowest form of learning and they act like robots. They are also taught to NEVER challenge authority and in spite of the flight training heavily emphasizing CRM/CLR, it still exists either on the surface or very subtly. You just can't change 3000 years of culture.

Another very worrisome practice that all of us noticed is a very high level of cheating and using older colleagues' notes and help. Then when exams came, we passed while they massively failed as they had hard time adjusting to modified exam questions they haven't seen earlier. They all went back home after graduating and not a single one stayed in the US.

I expect this will be a factor that contributed to the accident and I hope the NTSB gets a chance to talk to this UA captain.

bristolflyer · Wed Jul 10, 2013 2:35 am

I heard today that they could see 3 red Papi lights. Is that acceptable if all other parameters (ie speed) are as they should be? Or is 2 and 2 the only allowable showing?

Aesma · Wed Jul 10, 2013 2:47 am

Quoting D L X (Reply 77):
You don't need to transmit the data in the cameras the same way you don't need to transmit the data on a CVR/FDR. Go get the data when you need it, which is maybe once every 40 years at SFO.

But then what if the plane crashes ON your camera ? Do you make it resistant like a black box ? And you also have to regularly check that it is working, manually. Whereas if you have a data cable both problems disappear.

I don't think it's a bad idea in itself, it could be implemented in a couple decades, you would install them at the same time you would do other major work (resurfacing the runway for example, or redoing the lighting, etc.). I'm pretty sure it would end up being more useful to track people entering airport property illegally or things like that, though.

Quoting fpetrutiu (Reply 99):
http://www.youtube.com/watch?v=JlMdMTy8JLg

here is a recreation:

http://www.youtube.com/watch?v=VqmrRFeYzBI

This was Tarom ROT381 attempting to land at Paris Orly airport in 1994.

Scary, I didn't know of this incident ! Here is the BEA report in English, fascinating read : http://www.bea.aero/docspa/1994/yr-a940924a/htm/yr-a940924a.htm

dirtyfrankd · Wed Jul 10, 2013 2:49 am

http://www.liveatc.net/listen.php

This is an ATC recording of the KSFO Tower between ~11:00 am PST and 11:30 am PST on Saturday 7/6...OZ 214 crash took place at approximately 11:28 pm PST.

Key Messages:

@ 17mins 48secs - KSFO tower instructs UA 885 (744 which witnessed OZ 214 crash) to hold short of Rwy 28L

@ 18mins 35secs - OO 6263 radios KSFO tower requesting clearance to land Rwy 28R

@ 18mins 40secs - KSFO tower gives OO 6263 clearance to land on Rwy 28R but cautions wake turbulence from 757 that just landed on 28R and cautions "heavy 777" off to left hand side (not sure if referring to OZ 214)

@ 18mins 56secs - Unknown Heavy (cannot make out callsign) radios KSFO tower stating they are on final for Rwy 28L

@ 20mins 07secs - OZ 214 Heavy radios KSFO tower stating they are on 7 mile final for 28L

@ 21mins 13secs - KSFO tower gives OZ 214 Heavy clearance to land on Rwy 28L

@ 21mins 26secs - UA 885 radios KSFO tower stating they need a few more minutes (if you read the UA 885 F/O's e-mail that was published earlier in this thread...he states they were dealing with a HAZMAT cargo issue and a weight and balance issue at the time)

ACCIDENT SEEMS TO OCCUR SOMETIME BETWEEN 21mins 50secs and 22mins 00Ssecs...listen to the flurry of radio messages after that

Aesma · Wed Jul 10, 2013 2:52 am

Quoting bristolflyer (Reply 143):
I heard today that they could see 3 red Papi lights. Is that acceptable if all other parameters (ie speed) are as they should be? Or is 2 and 2 the only allowable showing?

Well if you see that you have to act immediately to correct it, it's not horrible yet. In this case they went from three red to four red, says it all.

I know that in a small piston plane you will not abort your landing for that, but in a large transport at 500 feet when everything should have been stabilized, I'm guessing a go around at that point wouldn't have been superfluous.

Aesma · Wed Jul 10, 2013 2:58 am

Quoting Klaus (Reply 121):
• saw 3 red 1 white PAPI

• told PF to pull back

• had set speed at 137kn

• assumed that autothrottle maintained speed

Thanks for posting that, as I lack time to follow most links/watch videos.

I quoted the really crucial points/moments in my opinion.

It proves that all the talk here by some about automation and precisely the autothrottle were spot on, I didn't intervene myself on that subject since I really don't know much about how it works, just that many here rant that Boeing does it better with their moving throttles.

locoflyer · Wed Jul 10, 2013 3:00 am

Quoting mach92 (Reply 117):

Not wishing to cause offence, I spent almost 2 years on 747's with an Asian carrier and left after the 5th serious incident involving incompetence and lack of CRM that management refused to take seriously following my reports.

The incidents were ignored by the company and an investigation was not carried out.

Not wishing to come accross as racist (and I do not believe I am), I am convinced the issues stem from cultural issues that beset the majority of Asian educated/trained aviators.

Those of Asian decent who were born and raised in a western country do not have the same issues and in the most part display a far better attitude and mind set to flying and CRM.

In reference to hand flying an approach and landing, I can most definitely state that in my experience the airline I was with would have bent far more 747's if it wasn't for the operation of autoland, they just don't seem to get it.

This accident, on the basis of evidence provided at this time, clearly shows that positive CRM was not being displayed.

To me, the 3 important facts of:
1. the approach was not what I class as stable at any time
2. a missed approach was not called before (preferable in this case) or at minimums
3. the PNF/Trainer was not monitoring his trainee effectively and failed to take positive action when it was clear the aircraft was not on profile
are indicative of the root cause of the crash regardless of all the other speculation provided on this forum.

It is my professional opinion, unless the NTSB can provide evidence to the contrary, the responsibility for the crash clearly falls at the feet of the flight crew, and if you cannot land an aircraft in VMC without the assistance of electronic devices (A/T, ILS etc), you should not be piloting an aircraft.

wjcandee · Wed Jul 10, 2013 3:10 am

Quoting dirtyfrankd (Reply 145):
"heavy 777" off to left hand side (not sure if referring to OZ 214)

It was not. Listen to the tape carefully and you would know.

Klaus · Wed Jul 10, 2013 3:16 am

Quoting 7BOEING7 (Reply 141):
It wasn't the automation that allowed Al Haynes and his crew to make a "landing" in an Iowa cornfield (without even a stick and rudder)

Better automatic support might have helped saving additional lives if it had been available. As far as I remember the circumstances, they had major workload issues in their situation even with an additional pilot happening to be at hand, forcing them to improvise more than they might have needed to with additional useful system support. In a crash with multiple fatalities it is difficult to argue against well-designed technical improvements if those would have been available or even feasible back then.

None of that takes anything away from them managing to save many more lives than anybody could have realistically expected, and they have most likely fully earned every bit of admiration (no offense intended – it's just not my call to judge).

But I just don't believe that they wouldn't have given everything to get any additional help they could have had – human, automated or otherwise.

Quoting 7BOEING7 (Reply 141):
or the Qantas crew to get their severely wounded A380 back on the ground safely. Flight crews with excellent "button pushing skills" but poor stick and rudder would have lost both those airplanes.

That is pretty much directly wrong as far as I can tell.

Not entirely dissimilar to the Iowa crash landing, they were faced with a severely damaged aircraft but they had a multitude of automatic support systems which they triaged and utilized to the best of their combined capabilities to avoid even a single fatality. As far as I'm aware making it back all alive was very largely due to them having a very good understanding of the systems they had left and of how to use those to get through it all alive.

As far as I recall the sequence of events, traditional "stick and rudder" flying was not the primary reason why they survived, even though a good understanding of flying basics was still very important to get the damaged airplane down in (mostly) one piece. But the FADECs in the still operating engines and the Airbus FBW systems supported them even in that.

The two events can't be compared directly across the decades and different circumstances (the plane in the older event was much more severely crippled), but technological progress likely does play some role in how they differ.

I'm not a pilot but a (non-aviation) IT specialist, so my biases naturally differ from yours, but do you really argue for removal or neglect of automated systems whose introduction is at the very least very clearly correlated with a massive increase in overall aviation safety?

I wouldn't dream of dismissing the importance of fundamental flying skills, but I hope I've just misunderstood your argumentation as dismissing technical advances such as automated support systems summarily.

[Edited 2013-07-09 20:29:41]

Who is online

Popular Searches On Airliners.net