Slashdot Mirror
Second SFO Disaster Avoided Seconds Before Crash
sabri writes "On July 25th, flight EVA28, a Boeing 777 flying from Taiwan to SFO, was on the final approach for runway 28L when they were alerted by ATC that they were only at 600ft above the ground at less than 4NM from the threshold. SFO's tower directed the flight crew to climb immediately and declare missed approach. Assuming they were flying at 140 knots (typical approach speed of a 777), they were less than 2 minutes from the runway and at a 3 degree angle (approx 500ft/min descent), about a minute from impact. This is the same type of aircraft and runway used by the crashed Asiana flight. Similar weather conditions and awfully similar flight path. Is there a structural problem with computer-aided pilot's ability to fly visual approaches?"
Clearly he learned so much from his last flight
"Is there a structural problem with computer-aided pilot's ability to fly visual approaches?"
No, Just Pilot error. The 777 has constantly landed at SFO everyday for years without issue and the cause of the Asiana has been well-documented.
copy paste from a forum poster at the link:
=====
@ Roly final thought
By LW on Sunday, Jul 28th 2013 15:45Z
+++Failure to use all available aids, even during a routine VFR approach is a crew or training issue.+++
=====
Planes don't fly themselves... yet. An experienced and attentive pilot is still necessary, who'd have thought?
In short, no, there is not. There is a problem with the airline putting inexperienced dumbasses in the left seat to save money.
Is there a structural problem with computer-aided pilot's ability to fly visual approaches?
Parse fail. I've even had my 3 cups of coffee and I got nothin'.
is to ban all airplanes. Because of the children.
Time for a 500 post thread saying the same 3 things:
1) I am not a pilot but here is why the pilot was wrong
2) There is a problem with Asian pilots since they weren't loved enough by their mothers
3) Hey don't be racist, Asians are just good at different things than Americans
"that they were only at 600ft above the ground at less than 4NM from the threshold"
And what is an NM? I'll assume it's not nanometers.
Yes, they were below the glidepath, and yes they blew the approach and had to go around: but this is hardly seconds from disaster or even a close thing. 600' at a normal approach speed is not "close" to the ground and 3.8 NM is more than 3 minutes at Vref which is certainly adequate time to respond.
These kinds of things happen and the only reason we're even hearing about this one is that it happened at SFO 28L.
I expected a little less sensationalism and a lot more intelligence from slashdot.
"For about three months in mid-year 2013, the FAA will render the LDA for runway 28L and the ILS for runway 28L out of service (OTS) at San Francisco International Airport (SFO) due to runway construction. The loss of these navigation aids will eliminate the ability for SFO to conduct PRM approaches during simultaneous offset instrument approach (SOIA) operations. The FAA plans to publish RNAV (GPS) PRM procedures prior to this navigation aid shutdown."
The visual approach slope lights have been OTS at SFO for awhile..
The problem is that some countries license pilots that cannot fly the plane if all the computer-aided bells and whistles are not available. This alone wouldn't be so bad but there are enough carriers around that hire anyone who has the legally required paperwork and is willing to do the job at the peanut pay offered. Actual skill at operating the plane beyond what is required "by the book" (use automatics) is optional.
Ban such pilots from operating commercial jets in US (and/or EU) airspace and it will fix the issue in a hurry.
Otherwise we can wait until enough expensive jets have been trashed by these "pilots" and enough people killed - at which point operators that do not change their policies and recruitment standards die out.
Some Asian carriers actually outright ban the pilots from flying the plane manually (automatics _have_ to be used at all times, if available). How on earth do they expect the pilots to cope in an emergency when the toys fail? Or when the toys can't be used (SFO does not have serviceable ILS at the moment so visual approaches are needed)
If we are going to have a news article every time there is a missed approach, Slashdot should be renamed.
Or maybe even 'Kent Parker Wright'
Tune in to KTVU San Francisco. I bet they have the names of the flight crew. Probably the same guys.
Join the Slashcott! Feb 10 thru Feb 17!
At least it wasn't Ho Lee Fuk or We Gwan Di
Even if there was a problem with the auto-pilot, AFAIK it's illegal to land on auto in the continental USA...?
See this take on the problem from Aviation Week:
http://www.aviationweek.com/Article.aspx?id=/article-xml/AW_07_22_2013_p25-597816.xml
Pilots are like anyone else, if they lean on a crutch long enough they forget how to walk. Then if the crutch turns out to have a fault, boom!
Yeah, but ATC had to tell the crew, "hey dude", you are coming in too low. This is good of the tower to give them that help, but it is like your passenger calling out to you that you are about to drive off the road -- it is really the pilot's/driver's responsibility to stay on path. If your passengers are calling out warnings of impending crashes to you, you might want to better look where you are going.
" is there a structural problem with computer-aided pilot's ability to fly visual approaches?"
What the fuck does that even mean? The only people deciding whether something is a "structural problem" should be licensed PEs. I'm not one, and it still irritates me when people misuse engineering analogies to sound smarter than they are.
What you're asking is, "Do commercial pilots flying heavily-automated aircraft suffer from diminished ability to manually fly visual approaches?" And the answer is, quite probably, and the solution is better training.
Is there a structural problem with computer-aided pilot's ability to fly visual approaches?
Not sure of the specifics of this incident (VMC or IMC), but there's no "structural problem" with automation and visual approaches. It is more likely simply an issue of training- about limits of automation and flying a visual flight path.
The automation can be used as a aid during a visual approach, but one must be familiar with how to set up the FMGC/FMC. Training costs money. Sim time is a limited and costly resource and managers are always looking to save a buck. Safety and profit are often opposing metrics.
Have you ever noticed that anybody driving slower than you is an idiot, and anyone going faster than you is a maniac?
So the LDA is OTS at SFO and the FAA published RNAV PRM for SOIA. TTL that ATC stepped in or EVA28 would have been SOL and all passengers DOA.
Is it just my observation, or are there way too many stupid people in the world?
I was under the impression that Otto Pilot hasn't been out there since the days of the 707, and even then, Ted Striker had to land the plane.
In SOVIET RUSSIA... erm...NSA AMERICA, the Internet logs onto YOU!
since the approach to SFO is over the bay.
Wei = "way" and Wie = "wee", if you're going to make race-baiting jokes, at least get the pronunciation right.
if this is supposed to be a new economy, how come they still want my old fashioned money?
Plane arrived in SFO on the 24th, not the 25th, according to TFA.
I was compelled to check because I was in SFO twice on the 25th.
So, basically, the air traffic controller at SFO needs to avoid directing Asian airliners to runway 28L until they can use easy-button landing again.
There is a valid point here about VFR vs IFR experience. But it works both ways...
Pilots who spent minimum time on VFR then went straight to lots of IFR-intensive commercial jet flying tend not to look out of the window so much, and even then don't always correctly interpret what they see.
Pilots who spent, and still (typically as a hobby) spend a lot of time flying VFR can still make mistakes, (a long, wide runway is easy to confuse with a short, narrow one, for example), but generally would be the ones I'd prefer to be with in a tricky VFR approach.
Some of them have really poor IFR skills, however...
My anecdote; flying right seat in a C130 on short finals - all the numbers are good and we're cleared to land.
Me keeping quiet as I was called in due to an emergency, and have no experience on type.
I suggest to the boss that going around might be a good idea - he snaps that "everything looks fine"
That's when I pointed out that something big and heavy was busy rotating off the same runway, and coming our way...
Guess the other guys were not looking out of their windows either...
(They'd screwed up big time; downwind takeoff and on the wrong runway.)
For you racist idiots here - all pilots involved were highly-trained by the military and white Anglo-Saxons :)
I suggest you read this post from a former UAL Pilot and Flight instructor for Asiana:
http://originalforum.justhelicopters.com/DisplayThread.asp?BD=2020564&Page=1&ForumID=23&msgid=2020564&OM=2020564&Return=DisplayThread.asp&D83jsd=True
In short, the culture in SE Asia produces pilots who are well trained to operate an aircraft as a piece of Machinery, however are unable to "Fly".
When asked what happened, the pilot stated "As you Americans say, I fucked up."
This tells me that ATC was engaged in a "hurry up you guys" approach processing. Seems likely that the controller was pushing to maintain spacing and make the best use of the limited runway. This lead to controllers creating difficult flying conditions by not getting aircraft lined up well and entering final approaches far enough out, or not allowing for speed and altitude reductions far enough in advance. This makes it extremely difficult to get a large aircraft into landing configuration on a stabilized approach. As a pilot of a large commercial aircraft, there are a LOT of things to get done and checked before you put the wheels on the ground. This process takes time and space. Apparently ATC wasn't giving pilots enough time and space if there are multiple pilots having to conduct missed approaches for being too low.
ILS approaches take *TIME* to do, even with a radar assisted approach fix. They where NOT in a situation where they would or could do even a quick ILS approach. During the daytime in reasonable visibility, they simply don't do ILS approaches because they are time consuming. You have to find the In this case, nobody was looking at the ILS, ATC was apparently pushing everybody to shoot short visual approaches which where steep and not stabilized. Had they been doing even what I call a "Quickie" ILS (where ATC gives you radar vectors to the outer marker/approach fix) They would have entered their final approach nearly 5 miles out at 3,000 AGL at a minimum. Chances are good they could have reconfigured the aircraft and stabilized the approach given the time required to fly 5 miles at 140 Knots.
So, this accident was pilot error with multiple contributing factors. One of those factors will likely be that ATC was pushing pilots into approaches that where starting too high and fast. Another will be that the pilot didn't recognize the unstable approach, got too low and too slow to recover. He flew his aircraft into the ground. The only question I have is about any unexplained mechanical issues. There are cases where ice in the fuel has caused the failure to throttle up when commanded, leading to landing short. (See BA Flt 38 crash) My guess is that we won't find any mechanical issues, just pilot error with ATC contributing by shortening the approach profiles too much to be safe.
"File to fit, pound to insert, paint to match" - Aircraft Maintenance 101
It indicates that this was routine, and nothing to be alarmed about: http://www.airliners.net/aviation-forums/general_aviation/read.main/5828205/
Apparently they're doing construction on the runway there, according to the comments in the linked story. That would explain a lot of things.
Occasionally living proof of the Ballmer peak.
responding to my own post because I looked at the data.
At 1500 feet they were descending at 2220 feet per minute. I assume this is when ATC freaked out. By 800 feet they were 'only' descending at a rate of 1920 feet per minute. By 600 feet they were still descending at 420 feet per minute. The next measure they were still at 600 feet but ascending at 900 feet per minute. So somewhere between 600 feet and going down and 600 feet and going up, they were below 600 feet. The data resolution is every 15 seconds, so roughly speaking they probably hit 500 feet on the way down.
Assuming the decrease from 2200 FPM to 1920 FPM is the first indication of a correction, it took them 1000' of altitude to correct their rate.
So, based on their initial rate of 2200 FPM and a 500 foot "cushion", it looks like they had 13 seconds "extra" to spare, and at that we need to figure in how much higher the transponder is than the landing gear and figure in wave height. Somebody buy that ATC a beer (after work).
My God, it's Full of Source!
OUTSIDE_IP=$(dig +short my.ip @outsideip.net)
Sorry, but I learned to fly at San Carlos airport (next to Redwood Shores, right adjacent to SFO airport and airspace) so I know a little bit of flying AND the area. I cannot see anything "unsafe" in the approach to SFO. Ofc I don't fly a "heavy", so if a pilot of one of those wants to disagree I'll bow to superior knowledge. But as long as there is no ("heavy") pilot who disagrees I'll say the only thing a LITTLE bit difficult is the approach over water.
However, even that is not an issue, you should have learned an easy way to track the point where you are going to touch down without ANY technical aids (we are talking visual approaches here, and visibility is near perfect in that area almost most of the time, esp. during the day): Keep your head in a position that you can easily remember and fix a point on the runway over a fixed point in front of you inside the airplane. When you look from your fixed head position over the fixed point inside the cockpit to the point on the runway it should not move. If it does (up or down) you are going to over- or under-shoot. That works independent of what the actual sink rate and speed (ergo the angle) is, always.
But then, my very own flight instructor later asked ME to demonstrate when I went on to learn aerobatics (i.e. "real flying") - turned out the "professional" pilots hardly ever do anything but "straight & level". Also, 5000 hours does not seem a lot if most of it is spent not just "straight and level", under computer control, and "at altitude". Only while maneuvering, incl. take off and landing, do you exercise flying skills. I said "flying skills", piloting skills include a lot more of course, from talking to ATC to calculating course, fuel, etc. etc. What those "professionals" seem to lack is good old FLYING SKILLS. It may sound strange from a lowly "small airplane pilot", but when I read that that Air France flight from Brazil went down because the pilots wanted to pull up when the airplane was in a stall (or close) - FOR MINUTES!!! - I really couldn't believe it - with some solid (small airplane!) training every pilot knows that you can never, ever pull UP to get out of trouble unless you have excess speed to trade for.
That doesn't mean I could fly a big airplane (wouldn't even be able to start it I guess), but while it does not matter to anyone that I lack the skills to fly a big airplane it matters to all passengers if the pilots cannot FLY (not "pilot") their airplane. I mean "fly" as in "without computer".
Is there an airline pilot here? I'm curious, what would you say about the FLYING skills of (big airplane) pilots? It seems that in the US the situation isn't bad, that this is an Asian (or Korean?) problem, and as I read it in an aviation forum not necessarily one of culture (at least not any more) but of many variables, including how easy it is for a lot of people to get to fly privately in the US vs. small countries like S.Korea, so that when a S.Korean wants to become a pilot they start from zero and do the training with an eye on the cockpit jobs (ASAP ofc, time is money), so no time/resources to do "fun flying" (like acro, which really, really teaches to fly). Then there's that even if you go into the job with good skills, how much is left after 10 years of mostly computer-aided careful "by the book" flying? How many pilots keep their (low-level) flying skills sharp by flying a small airplane in their spare time, to do "fun stuff" and "unusual attitudes and maneuverer"?
600 meters would be great on that approach! Wait, the altimeter shows feet? Oh shit.
Chewbacon
The Bible is like Wikipedia: written by a bunch of people and verifiable by questionable sources.
No dependency on computers necessary for people to do stupid things. We blame automation but incompetence, failure to follow procedures, complacency, and dysfunctional cultural norms (ethnic or professional) are often a major contributor to disasters.
At a desert oil production station (Gialo) in Libya in the 80's a pilot crashed an LAA F27 passenger plane carrying field workers when he attempted a landing on the old runway instead of the new one.
The day was perfectly clear: not a cloud in the sky. Visibility was extremely good. Virtually no wind.
The old runway had been dusted over with crushed white Saharan calichi which made it fade into the surrounding background of light tan hardpan and sand. To further discourage use of the old runway, loads of rock had been dumped in piles down the centre end to end.
The new runway was inline with and off the end of the old. Brand spanking new asphalt, black as midnight, complete with high contrast runway markings and looking like big black stripe on a pale background. A blindman in a snowstorm could not miss the dang thing; it fairly screamed LAND HERE.
The NOTAMs were updated properly and anyone flying into Gialo would see the runway info as the first item.
Nonetheless, the pilot made his usual approach over the station like he had done many many times previous and did not realize the mistake until he was just about to touch down. Pulled up but a wheel caught a rock pile and he pranged the nose into the old strip. Go figure.
No one was killed and as far as I saw, they all walked off. One fella (a Brit of some flavour) had been sleeping and stepped out saying 'What's all the fuss about?". We turned him about and, gazing at the bent props, crushed nose and broken gear, he said: "Felt like a regular landing to me."
Pictures and official description at
http://aviation-safety.net/database/record.php?id=19869999-1
The explanation given is not exactly correct and likely is the narrative the pilot gave to keep from losing his license/job, or perhaps other consequences given the nature of the regime at the time. The pictures appear to have been taken some time after the accident (rust on panels) and it is likely not the original location. Looks like it was dragged away.
Is anyone really surprised? They can't drive either.
There IS a structural problem with that sentence. :-)
I think it is missing an article.
"Flyin' in just a sweet place,
Never been known to fail..."
I like how the title/summary assumes you know what the fuck SFO means.... (Google says it's San Francisco Airport, good, it wasn't that hard, shouldn't be that hard to just add the "(San Francisco Airport)" to the summary)
Proper journalism, where art thou!? (Yeah, I know...)
-- Counting backwards since 1984!
This is a nice solution to the problem. I'm sorry, history has proven that you are unable to recover from this failed approach, go around now, throttle up, pitch up and stop, you are going to kill people. Maybe culturally this allows them an out, that saves them from having to make the decision to go around.
Personally, if I were the insurance company, I would field test (sim) all the pilots for failure to figure out they are going to smack the ground. Quadruple the premium, if they company let's people that fail, continue flying. It is a business decision.
You'd know that half of Korean culture nowadays is a hardon for Western Culture (Like any K-Drama involving Korean nobility and their decidely european dress uniforms?).
Difference being: All the cyberpunk-esque things like Smart TVs and ubiquitous cell phone data plans that've actually been produced were shown in 'production' form in Korean TV first (I still haven't seen anybody in Western TV using a REAL smart TV to surf the internet for instance.)
Maybe it isn't an Autopilot issue. Has anybody asked yet if these pilots used the same training simulator? It could be that they practiced making this approach wrong.
I won't join Slashcott. OTOH, If Beta goes live, I just won't be back until it's fixed. Sorry Dice.
scientists have reaffirmed that magnetite has absolutley ZERO PRACTICAL value in transistor applications. "No way in hell" was the unatributed quote.
Popular Mechanics have a transcript and analysis of the final 10 minutes.
When I was going to the KDE 4.0 release party at Googleplex in 2010 (my only US trip), the pilot failed to put us down at SFO three times in a row. During the first two attempts, I remember thinking "wow, that was pretty close". He finally had to put us down in Oakland instead, where we had to sit on the runway for quite a while before the customs there got their act together and could process us.
This was a very misty January day though, and I have no idea if this is common at SFO.
Err I mean 2008. Time flies..