The Asiana 214 investigation has proven to be every bit as interesting and disturbing as I’d predicted.
Most of the reporting and commentary has been focused on the pilot’s interaction with — and understanding of — the aircraft’s automation system. It seems clear they were having trouble getting the aircraft to do what they wanted during the approach into San Francisco.
You won’t hear pilots bragging about this at cocktail parties, but “what’s it doing now?” is uttered far too often on the flight deck. I myself have been puzzled about why the airplane didn’t do what I thought I asked it to do. Usually it’s a programming issue, but not always.
The most recent issue of NASA’s Callback publication, issue 407, details the story of four professional flight crews who had automation confusion issues similar to that experienced by the Asiana crew. So this isn’t exactly uncommon.
Either way, pressing the wrong button is not a criminal offense.
“Cleared for the Visual.” Gulp!
What is criminal is putting a captain on the flight deck of a passenger airliner when he’s unable to comfortably hand-fly it, because when the electrons aren’t flowing the way you want ’em to, flying the airplane by hand is often the best course of action… not to mention the most fun, too.
Well, most of the time anyway.
The Asiana Airlines training captain who crashed a Boeing 777 at San Francisco International Airport in July was anxious about the visual approach, which he described as “very stressful,” according to investigators.
Capt. Lee Kang Kuk, an eight-year employee of Asiana on his first extended trip flying the 777, also told investigators he was confused about the operation of the airplane’s automation controls, according to a report released by the National Transportation Safety Board on Wednesday as the board held a hearing into the crash.
The 777’s speed dropped dangerously low on the approach, made with assistance of the PAPI lights but without vertical guidance from the ILS glideslope, which was out of service at the time. Both Asiana 214 pilots said they were unsure about the automation mode with respect to the autothrottles, which should have been engaged on the approach. Instead, the autothrottles were set to idle, according to investigators.
The training captain stated it was “very difficult to perform a visual approach with a heavy airplane,” according to the safety board summary of an interview with the pilot. Asked whether he was concerned about his ability to perform the visual approach, he said, “very concerned, yeah.”
An automation interaction problem — the so-called “FLCH trap” — I can understand. But inability to comfortably fly a visual approach? On the surface, that’s a major head-scratcher. When you dig a little deeper, however, it makes perfect sense.
The Key to Good IFR: More VFR
I don’t know how Asiana does it, but many foreign airlines hire their pilots “ab initio”, meaning they are trained by the airline as airline pilots from day one. They have no exposure to pleasure flying, aerobatics, or gliders because the concept of “general aviation” does not exist in most countries. Ab initio airline pilots receive only the minimum required VFR experience. As soon as they venture into instrument flying, the VFR world is left behind forever. They have no use for it! Or so they think.
I’d imagine many of them never fly under visual flight rules again for the rest of their lives. It’s sad. And it’s no wonder some of them are uncomfortable with the thought of flying a visual approach!
It’s not as if the weather was poor, the runway short, or the airfield surrounded by high terrain. There were no issues with density altitude, runway slope or width, or anything else. San Francisco International’s runway 28R is nearly 12,000 feet long. I’ve landed on it many times myself. The weather was clear, winds calm, and the airport is unmistakably large.
Sure, the controllers do tend to keep arriving aircraft quite high. But even from 10,000 feet on a tight downwind, it’s not rocket science to start slowing the airplane and adding drag. Unless you’re asleep at the wheel, you know what’s coming. And even if you don’t, you can ask. The controllers speak English, too. A visual approach in those conditions shouldn’t scare the pilot-in-command of any aircraft. In fact, if there’s an easier way to land an airplane, I’m not sure what it is.
Kids Can Do It — Why Can’t We?
To put this in perspective, consider a glider. It has no engine, and therefore cannot abort a landing attempt. Once you begin an approach to the runway, you are going to land, period. These aircraft have no instruments, no electronic guidance, and they fly in and out of airports without any visual landing aids whatsoever. The landing areas tend to be short, narrow, and rough. And here in the U.S., students as young as fourteen years old can fly them solo. Fourteen! They’re just kids, and apparently even with virtually no flight time, they have no trouble getting comfortable with something that a highly experienced major airline captain felt very uneasy attempting.
This begs the question of how Captain Kuk became so uncomfortable with a simple visual approach. I’d estimate that 75% of all approaches are visuals. I’d be shocked if Kuk hadn’t flown literally hundreds of them. As a scheduled airline pilot, he was required to undergo recurrent training every six months, and had been doing that for eight years.
So how did this level of discomfort with basic visual flying escape the schoolhouse? If Kuk’s training is anything like what we undergo in the Gulfstream, he may rarely have ever flown that kind of visual procedure in the simulator. Mostly what gets simulated are low-visibility conditions. The assumption that it’d almost be “cheating” to have visual references outside the aircraft might not have been correct. Visual approaches in the sim are typically combined with other anomalies: no-flap scenarios, windshear simulations, landing gear blow-downs, etc. But not the typical slam-dunk from a harried controller.
One wonders how many other airline pilots pale at the thought of flying a visual approach (or as the VFR pilots among us call it: landing). I know most airlines no longer allow circle-to-land procedures, but even the neophyte instrument pilot has to perform them to acceptable standards before being issued an instrument rating, and that’s infinitely more demanding than a visual approach. Instead of practicing an ILS PRM at San Francisco, perhaps we should be vectored in on one of those famously high downwinds and cleared for a visual approach from two miles up. Maybe we should train a little more like we fly.
And while we’re at it, taking a hint from that fourteen year old kid who just soloed a beat up Schweizer glider might not be so bad, either. Get out of the glass palace and into an actual airplane where there’s nothing to do except fly by looking out the window.
I SAY START WITH A PIPER J-3 AND GRADUALLY MOVE UP TO C-150. C172, C182. ANY PERFORMANCE TWIN AND THEN MAYBE, A GULFSTREAM. DO THESE APPROACHES UNTIL HE CAN DO THEM WITH HIS EYES SHUT. THEN MAYBE WE WOULDN”T SEE INCIDENTS SUCH AS THIS.
That’s kind of how a typical progression works. I’d imagine Captain Kuk started in a Bonanza or something similar and then flew a twin for that rating, and then jumped into a Boeing-class airliner. On the other hand, he could have been a military veteran. Who knows. I think that’ll be one of the interesting bits in the accident report.
Total flight time can be a deceptive metric. You’d think a guy with 10,000 hours would be very experienced, but if he spent years as a cruise pilot, that might not be the case. As the saying goes, you can fly ten thousand hours, or fly the same hour ten thousand times.
As the saying goes, you can fly ten thousand hours, or fly the same hour ten thousand times. I like this saying. Thanks Ra
….or on autopilot for 98% of the time, leaving 200 hours of “real” flying.
Admittedly, the “feel” of an airplane becomes less apparent as the size of the aircraft increases but given a clear view out the window and an airspeed indicator there is nothing else needed to safely land any airplane, large or small. That is IF you are looking out the window and monitoring your airspeed. Neither was the case in this accident.
Even if he wasn’t cognizant of the decaying airspeed, there should have been verbal call-outs by the non-flying pilot. Given that the right seater was an instructor, you’d expect near-perfect adherence to standard operating procedures of that kind.
Like most accidents, this one seems to have many links in the chain. Or holes in the cheese, if you prefer that imagery. 🙂
We called it, “Click, Click-Clicl,Click”. When the automation isn’t getting it done, turn it off and fly the airplane. Automation dependency isn’t anything new. Flying the B-767 in the 80’s we soon learned that full up automation can get you into trouble quickly and that maintaining your hand flying skills were essential. Why these lessons have to be relearned is beyond me.
Perhaps the airline pilots of the 1980s came up through the ranks of general aviation VFR flying doing things like pipeline patrol, check flying, banner towing, instructing, charter/corporate flying in piston twins, turboprops, and so on. Quite a different experience from the ab initio airline pilot abroad!
I was Military. Flew fighters. You learn stick and rudder or die. The airline I flew for hired fighter pilots because they knew that fact. The really good pilots I flew with loved to hand fly the airplane, GA or Military.
>>The really good pilots I flew with loved to hand fly the airplane, GA or Military.
That much has not changed! At least, not in my little slice of the industry.
I had my first flight instructing job in 1978. Then I was a part 135 charter pilot for nine years. We operated off dirt runways in C-206s and flew single pilot IFR in twin cessnas. I have flown for a regional airline since 1987. At the regional I have flown everything from Beech 99s to Fokker 70s to CRJ900s. All of the pilots I fly with love to hand fly ac. Personally I do not known any US airline pilots who would have had a problem with the visual approach into SFO just because the ILS was out, including our younger pilots.
Unfortunately in Europe as you said, pilots go from student to airline F/O in 200 hours. Those are very intense airline/aircraft specific training, and company SOPs are very robust. Furthermore most are excellent cadets, but It is not enough. The scary part is that these F/Os will become captains @ around 3000 hours (3-4 years) and 23 years old. In the near future they will be PIC at around 1500 hours as that is the legal minimum. Airlines are aware of this, that is why I presume are teaching their pilots to fly full automation all the time.
Authorities know that and are very permissive (big lobby maybe?).
You may ask yourself how can a cadet ever become a captain since he doesnt ever get the possibility to fly as a PIC (hours required to “unfreeze an ATPL)? Well they invented PICUS ( pilot in command under supervision) which means the fly in a PF role with a captain that must attest their ability of command. In my airline they told us we should sign them off as soon as they are released from line training!!!!
Another unsafe setup is that some airlines are firing F/O s who are not able or willing to become captains in that short time frame. Experienced captains are leaving asap and senior F/Os are not even being interviewed as they are not contributing much to the self sponsored training as it has become the norm at most airlines. There are enough cadets willing to pay their way to a job, and at the end a senior F/O will do the same job as young cadet, same as a senior captain will do the same job as a 3000 hour one ( or so does the management think).
It is all very legal and until now it worked but I believe it started a very very bad trend.
I met a pilot some years back, who flew DC-9’s for Ozark Airlines. He told me that, after TWA bought out Ozark, they would often have ex-Ozark pilots flying with long-time TWA pilots. You could tell the two pilot groups apart by their flying style, he said. When a TWA pilot started to feel overloaded, he’d reach over and turn on the autopilot. When an ex-Ozark pilot felt overloaded, he’d turn the autopilot off.
Well, that’s what he told me, anyway . . .
It’s interesting that speed and altitude restrictions, neither of which were particularly aggressive on this flight, are considered major sources of workload and stress. Their training environment should more accurately replicate the workload of flying into the airports which they serve. If their pilots were training visuals in the sim with accurate ATC, it’s reasonable to infer that they would eventually become more comfortable with this.
Ask a US-based CRJ regional captain or FO who flies in the US if they find visuals with speed restrictions to be stressful and see they respond. They do more operations in the terminal environment than crews flying with the majors because of the shorter legs.
I agree that a better foundation in VFR operations is the way to go, along with more relevant sim training. V1 cuts are great, and they’re important, but how about training for the case you’re actually going to encounter on a daily basis rather than once every million flights hours?
I guess I always felt the opposite. The operations that I did hundreds to times out on the line weren’t the things I wanted to practice in the sim. The sim was for practing the things that very seldom happen but could kill you if you weren’t proficient in them. Losing an engine at rotation or anytime below a 1000 ft. Go arounds,wind shear etc. Getting slam dunked by ATC was a way of life. Visuals are what you do most of the time anyway.
>>Visuals are what you do most of the time anyway.
So true! Although the visuals into SFO can be a little bit ridiculous. I have — more than once — been like 9,000 feet above the field on a tight downwind and kept up there until somewhere between the San Mateo and Dumbarton. That borders on cruel and unusual punishment, even by slam-dunk standards.
Bet you handled it and didn’t complain you were over stressed. And for sure you didn’t land short of the runway.
…but apparently this isn’t the case in Asia. Pilots there rarely do visual approaches and don’t have many VFR hours, so they should do more visual stuff in the sim.
Good question, Keith. The regional guys definitely take the cake when it comes to legs flown per day. They’re also flying to smaller fields with fewer precision approach systems, so they get to do the purely visual thing more often than someone flying a 777.
As you mentioned, PilotEdge would be a great resource for duplicating this kind of approach. And it provides a unique opportunity for the pilot: to simply say “unable”.
My bet is that he’s flown hundreds of visual approaches . . . with the ILS tuned in. Without the FD guidance that he was used to having on 99% of approaches, he got mentally overloaded, and he probably hadn’t bothered to look at a PAPI/VASI since training. Having never flown a jet (let alone a transport category), I don’t want to judge too harshly, but it seems like they realized absurdly late that things were going south—probably because there was no glideslope needle creeping up the dots.
I just read the CVR transcript from the NTSB docket and it doesn’t read like he was terribly anxious and stressed. The conversation looks pretty normal right up until the go-around was attempted. When he arrested the descent, the power he expected the autothrottles to give him never materialized. Missing that crucial element is what doomed the flight. As you noted, it sounds like when things went south they did so very quickly.
I WONDER WHAT HAPPENED TO THE PRE-LANDING CHECK LIST? WAS IT PRINTED IN KOREAN OR ENGLISH? JUST WONDERING.
Way back when.. When the Air Force was flying the B-36. pilots were required to fly a J-3 after so many hours, just to keep their depth perception sharp. I never thought much about that until my own eyes failed when I was 78 years old. Cataract surgery corrected that problem two years ago.
Just in case some of you observed that I am eighty (80) years old and still flying, you are correct. But, I do it with a safety pilot/instructor. I am old, not dumb.
Kudos my friend …
Capt. Rapp, another excellent post. You keep raising the bar.
I haven’t found much, actually nothing, written about the Asiana crash from the cultural perspective. I’m hoping hyper political correctness will not corrupt this investigation. Most accounts I’ve read seem to agree that the pf was the least experienced in the 777 but the most senior in the crew. This presented a real conundrum to the others and not one that is readily understood by American flight crews steeped in CRM for their entire careers. Culture may have made it nearly impossible for the others to speak up, let alone to step up, even in the face of impending doom. We can discuss best practices and training, abinitio or otherwise but I believe this crew was crippled by the overwhelming denial and inexperience of the senior pilot and the crushing weight of cultural respect by the rest of the “talent” in the cockpit. I will never travel on an Asian crewed airliner.
Thanks for the kind comment Roger.
The crew experience was summarized in one of the docket reports. Of the four people on the flight deck, three were captains and one was an FO. Ironically the only one who seemed to know what was happening was the FO! He tried several times to warn them about the decaying airspeed. It’s recorded on the CVR transcript.
Anyway, of the three captains, the pilot flying had the lowest time in every single category: time in type, PIC time, total time, last 30/60/90 days, etc. Yet he was hired before any of the others. It’s also interesting to note that he was the only “ab initio” pilot aboard. The other three were former South Korean Air Force pilots.
If there were cultural factors at play, I hope the NTSB does not shy away from discussing them in the report. It would be a real disservice to gloss over anything which played a material role in the accident.
While I somewhat disagree with Roger, I have to say that USA has some of the lowest standards and thresholds compared to other GA Private through Commercial certifications. Learning to Fly in Israel, Singapore, or South Africa, all of which I have experience with is a lot more stringent than that of meeting standards in USA.
Ron – This post reminded me of a conversation I had two weeks ago with an ATP (the school) student pilot with 400 hours. He was a foreigner on a visa and mentioned to me that he never even went for the 100 burger. All his flights had been purely dual or training . one of my friends who is a 747 captain for PAL now got into 2nd officer in a 727 after 212 hours. I wonder how he hand flies a 172
On paper you’re right, there are fewer hoops to jump through to gain certification in the U.S. than nearly anywhere else. A couple of years ago I wrote about the ridiculous requirements for obtaining an instrument rating in Europe, for example.
But there’s a lot of learning that happens outside the classroom. That’s what a vibrant GA community contributes to airline, military, and other forms of professional flying.
Regarding the 747 captain, many people over the years — most of them airline pilots — have told me there’s nothing scarier than an airline pilot flying a Skyhawk. Would you believe my very first student was an Aeroflot captain who wanted to do some GA flying here in the U.S.? He flared at 50 feet, just like I had been told to expect.
Yup. I had the same experience. An American Airlines captain went flying in my 172 with me and asked if he could land. I said sure. Wouldn’t you know the flare started about 50 feet as I yelled nose down nose down. and yes after 2 or 3 more landings he got much better!
One commenter makes a very important point: the Captain had flown many “visual” approaches but most (probably all except in the simulator) were backed up with flight directors in the approach mode with an ILS providing guidance, and the autothrottles providing speed control. He was actually uncomfortable not having an ILS.
I fly a Boeing 777, and when conditions are favorable, fly takeoffs and landings with flight directors and autothrottles all purposely turned off. We all know that skills need to be practiced to remain effective. I am sorry to report that many of my F/Os ask “can we do that” when I brief a raw data and no automation takeoff since few heavy aircraft airline pilots do them anymore. Most airline pilots hand fly the airplane but almost always have flight director and ILS guidance. Yes, a Boeing 777 can easily be flown by visual “pitch and power” like a C-172 or DC-3 … and should be occasionally to keep those skills effective!
The Air France stall accident in the Airbus from South America is another example of loss of basic skills in airline pilots that fly highly automated aircraft. I did fully developed deep stalls in the Boeing 777 simulator recently and all were recoverable within 2000′ feet of altitude.
Interesting comments all around. I am retired, major airline-DC3 to 747- 1956 vintage, Piper Cub start. From my vantage I am compelled to point out my horror at a 777 left seat man flying a visual without tracking his airspeed regardless of where, when, into what airport, with what machine and under what culture, auto failure, call-out from F/O’s, yelling from a flight attendant, whistles blowing, claxons warnings, voice stall warning, stick shakers, ATC directives, wind, snow, rain, ice, fire, bullets, VASI, engine failure, hijacking…..on and on. Airspeed is like breathing: miss it and you die. The current crop of pilots don’t know how to fly. Period. In my time of training,our airline (old fart alert here) training started with visual on all machines, then IFR, then simulator with multiple failures, then the real world. If you could not handle visual you never made it to the next step. Automation has progressed to the point where there are fewer pilots and more computer operators and game machine players sitting in the pointy part. A friend retired to become a check pilot for an Asian carrier; he tells me the idea of doing a visual approach for any of the pilots in his care is akin to asking them to walk into the terminal naked. Most can’t do it with any finesse, if at all. We have come a long way with automation, to our peril. Well nigh time the regulators start to regulate and the Boeing’s, Airbus and Bombardier builders not deliver into the hands of the incapable. The Bible says on Page one, Chapter one, Verses one, and I quote: Aviate, Navigate, Communicate or ye shall suffer the brimstone of hell. All else in all the rest of the book is nonsense. Cheers.
That’s what I’m afraid of: a systemic discomfort with purely visual flying rather than a one-off outlier among an otherwise capable fleet of pilots.
Let them tow gliders for a couple of months. Maintaining perfect airspeeds during tow and in the pattern flight becomes second nature. After a while you get cocky:” let’s see, I’m going to put the right wheel on the near end of the numbers and keep it down for a while”. Descending works well in circling flight keeping a 60 deg bank angle loading up the wing while maintaining enough manifold pressure to not shock-cool the old Lycoming.
After a stormy approach into O’Hare I asked the copilot who was thanking the paxes at the exit if he had hand-flown this difficult final. He looked away for a second, indicating the he was going to by lying and said: well, part of it.
Towing gliders is another great way to gain manual flying skill. That’d actually be a fantastic operation for an airline: run a glider school. The folks flying the gliders AND the ones in the tow planes would be developing/maintaining valuable skills — and having a hell of a lot of fun in the process!
Asiana pilots can’t fly VFR in an airliner because they’re not taught to. I have letters from two US pilot instructors working for the Asiana training department. Both noted repeatedly that the Asiana pilot trainees or line pilots could not fly VFR without automation. They informed the management at Asiana and were ignored. Eventually both quit because they couldn’t in good conscience continue to train unsafe pilots.
The San Francisco accident was inevitable and totally predictable.
Ab initio works OK as long as you check all the blocks. The box for VFR approaches without automation in Airbus and Boeing products was never checked.
While it’s easy to point at thus crews weakness and say it’s an Asian problem as a instructor for US airline I see the same problem, just look at UPS just two weeks or so after
The SFO ordeal. On a second note if we truly do care about safety why is it a international arrivals airport has the ILS out if service for for over a month?
Good point. This is probably not limited to Asiana.
As for the ILS, I think it was out if service because they were moving the threshhold or doing some kind of construction.
Even something as simple as having equipment and men working near the ILS can cause it to be notamed OTS because the presence of those objects can affect the signal bring transmitted by the localizer and/or glide slope.
I transitioned from the B-727 to the B-757..B-767 late in my career. I flew the B-767 ER for the most part on International
routes. I would hand fly up to initial cruise altitude and for awhile, then hook up the “Magic”, Normal and recommended procedure
was to clean up, climb out to 2500 ft. then let the “magic” go to work. I once had a FO ask me “I notice you hand fly a lot, Why”. My
answer was,” because I can and I enjoy it”. .
I have been a 737-800NG Captain for the last 12 years. I frequently fly sans magic all the way to cruise, then for much of the arrival to touchdown (we don’t do autolands in ours, just Cat III HUD approaches flown manually). My FOs also ask me the same thing, and I answer the same way, “Because I can, and I enjoy it”. I remember once as an FO turning off the magic in the 777 on arrival to London, while flying a holding pattern. The Captain only half jokingly asked me if “it was legal?” I fly my Cessna 182 the same way, mostly ’cause the autopilot in it sucks.
Years ago, Uncle Sam paid for my commercial ticket. I flew most of the course in a Cessna 182 that was certified for instrument training. One day while scanning the panel I noticed a “WING LEVELER” Switch. I asked the instructor about it and he said, “Don’t use, it is no good.” I didn’t know exactly what he meant but the first time I went on a solo cross country at 7500 feet, I was bored and decided to see just what he had meant. I am thankful I had that much altitude. When I flipped the switch, the airplane rolled 90 degrees to the left, almost on it’s back. Turning the wheel was like driving a Mack Truck without power steering. I recalled that the Wing Leveler was the culprit and managed to turn it off. With that I was able to resume normal flight. What I needed was a shot of Black Jack #7 but decided to wait until I was seated at my desk at home while I wrote, “DO NOT FLIP SWITCHES UNTIL YOU KNOW WHAT THEY DO! five hundred times. I never mentioned this to my instructor but should have and the circuit breaker should have been pulled.
Great article Ron! It really points out a gap that I think has been widening over the years – you still have to know how to fly in spite of all the automation.
I think it is starting to migrate even into the GA world. For example, someone in a TAA could fly IFR point to point and never scan instruments or anything. The airplane can do it all except for the flare. If those electrons stop flowing, watch out.
In the case of Asiana, I completely agree with your assessment. He had the deadly combination of not having the skills to get the job done and the complete lack of SA to mitigate the situation he created.
As a professional pilot, routine operations are a given and we earn our pay when something occurs that’s abnormal. He did neither.
Here’s what I don’t understand: if a poor train driver schmock misses a signal or fails to brake in front of a curve he will eventually be indited on negligent homicide charges – and rightly so. Why are these Asiana pilots not in jail or free on a high bond? Their passengers are just as dead and incompetence will bite you in a lot of ways. Standards of responsibility do not seem to apply to these pilots but I may be touching a very sensitive point here.
Train wrecks are usually caused by engineers who are sleeping, on drugs or texting or some other distraction that contributes to their negligence. As inept as these pilots appear to be, I don’t think they were doing anything other than flying very poorly.