Asiana crash finding implicates Boeing somewhat, but our view is this is a pilot issue, plain and simple

The National Transportation Safety Board (NTSB) yesterday lay the Probable Cause of the crash of the Asiana Airlines Boeing 777 last year at San Francisco as pilot error, but in the process implicated a “complex” auto-throttle system as a contributing factor.

Not surprisingly, Boeing disagreed (“respectfully” so), nothing that 55m flights in the 777 had occurred without incident.

While we are not at all surprised at Boeing’s position (we would have been surprised had it been otherwise), we side with Boeing on this one.

We will grant that perhaps the auto-throttle system might be tweaked to make a safe airplane and safe system even better, incorporating an aural warning when necessary. And perhaps the training procedures could be made better and more clear. But in the end, it remains the responsibility of the cockpit crew to monitor instruments and speak up when things aren’t as they should be.

In this case, the flight was also under visual flight rules (VFR). So, the pilots should have been:

  • Looking out the window to visualize the approach;
  • Monitoring the airspeed and altitude instruments, among others; and
  • Speaking up. Most importantly, the third pilot did see something was wrong but didn’t say anything. Culturally, it is common in Asia for subordinates to defer to the Captain. But it’s lousy Cockpit Resource Management.

If the pilots had been doing their job, the plane almost certainly would not have crashed. The auto-throttle may have led the pilots to a false sense of security, but in the end they didn’t fly the airplane.

That’s was caused the crash.

24 Comments on “Asiana crash finding implicates Boeing somewhat, but our view is this is a pilot issue, plain and simple

  1. … for subordinates not defer to the Captain..

    Think you meant to say subordinates TO defer to the Captain ???

  2. Given that Asiana implemented the NTSB crew training recommendations before they were formally released adds weight to your argument that pilot issues were the cause. Yes, one could add training and fatigue, and even culture, to the contributory factors, but the fact remains no one was “flying” the plane during descent. This should be a wake-up call to all pilots, not just 777 autopilot pilots.

  3. I agree that the pilots are at fault and the auto-throttle should not have been implicated as it was. It’s their responsibility to make sure they understand how the aircraft systems work and plainly, they didn’t. It is telling that not everyone on the NTSB board wanted to recommend a special safety review. Perhaps the whole brouhaha over the battery last year is causing the NTSB to take a harder line with Boeing.

    “The SCR recommendation was a point of contention among board members, two of whom voted for keeping the language in the final report and two of whom wanted it removed. A tie vote means the language remains, although members can file dissenting opinions as part of the final report.”

  4. We learned the hard way during my USAF career. No mater what is happening (emergency, system malfunction, etc.) FLY THE AIRPLANE. In a crew airplane, like the B-777, someone should always be flying, and others can figure out whatever needs attention.
    In the case of OZ-214, someone thought they were flying the airplane, when in reality no one was. The same happened aboard AF-447.
    An example of the crew doing the right thing, one person doing all the flying while others tackled the emergency was QF-32.

    • Wrong, the crew member would not release the stick! If he had the aircraft would have recoved on its own, get your facts correct please.

  5. It’s one of those problems that come up too often. The airplanes are so well automated that pilots are rarely called upon to use their basic airmanship. So when the pitot tubes ice up or the ILS is out of service they are caught flat-footed.

  6. In addition apparently seams that the flying captain did not adjusted in height its seat that prior the approach was occupied by a taller crew member.

  7. Whoever it was on the NTSB group that decided to “split the baby” and apportion some contributory fault to Boeing’s system need to be removed from the NTSB immediately.

    You, and everyone else without a “dog” in the fight all say the same thing: it was the pilots’ responsibility to fly the plane, in absolutely great conditions. If they didn’t understand the system and failed to pay attention, I don’t see how it is the fault of the manufacturer that the pilots failed Rule#1 in flying: AVIATE. Fly the darn plane, and not sit back as if you are the helmsmen on a Federation starship and let Starbase control take command of your ship and dock you on the star base. We may be close to Star Trek automation, but in the end, it is the pilots who ultimately failed to do their basic job. They can’t sit back and look out the window at SF Bay or the Golden Gate Bridge while the plane is doing whatever it wants (or what they commanded it to do)

    Robert Sumwalt has some explaining to do.

    • Good HMI design requires that it be intuitive _and_ follow some discernible logic.
      ( With a lot of exercise you can forex drive a car with inverted steering controls but it will increase risk in unexpected situations. )
      Some/?Most? of the people now clamoring for pilot error attribution were pretty vocal
      about allegedly basic faults in Airbus FBW layout. And they are wrong in both cases IMHO.

      One can’t have it both ways in a fanbased selection 😉
      A bit of self esteem rightsizing and reviewing the interface can only bring improvements.

      Then my impression is that the crew was intensely focused on “flying” the check captain. Seemingly a cultural problem that needs fixing.

      Is flying so undemanding today that there is not enough pressure around
      to “fuse” a crew into a working unit?
      The AF447 crew was seemingly divided by personal issues,
      OZ-214 crew was detracted by “codling” the check captain.

      That is the commonality blinking out of AF447 and OZ-214.

  8. First and foremost as a professional pilot at a US Major there is no doubt that this accident is the direct result of an incompetent crew who couldn’t manage to safely accomplish something every private pilot should be able to do. A visual approach without automation in CAVU weather. Their failure to monitor the situation and failure to act when it became apparent all was not well is inexcusable. Regardless of how complex the auto throttle system is you simply MUST pay attention in this business.

    That being said if the 777 auto throttle is designed in such a way as to create traps, and the system is poorly or improperly documented such that a pilot who has been through the training program and reviewed the provided materials isn’t likely to really understand the system then that is a problem. And its a problem that likely does rise to the level of a “contributing factor”. And it is appropriate for the NTSB to call out the need for changes in both the function and documentation of that system if they are likely to lead to confusion.

    A similar example is the A320’s Dual Engine failure with fuel remaining checklist. That checklist was designed for an event happening at 30,000′ and as the NTSB found it is utterly useless in a low altitude situation like that faced by Sully and Skiles. As a result the NTSB recommended that the FAA ” Require manufacturers of turbine-powered aircraft to develop a checklist and procedure for a dual-engine failure occurring at a low altitude.” Would such a checklist have changed the outcome for USAir 1549? Probably not. But the crew certainly deserved a checklist that would give them appropriate guidance. Sadly 5+ years removed from that event at least on the A320 the checklist remains unaltered.

    But back to Asiana, as I stated earlier the accident is the crews fault plain and simple. But 777 crews do deserve an auto throttle system that doesn’t feature hidden traps and who’s functions are properly documented. Hopefully Boeing and crews flying its aircraft learn lessons from this one that make us all safer.

    • The checklist for ditching was changed according at least one important point:
      closing of vents was moved from nearly the end of the list to an earlier position.
      The partly sinking of the A320 in the Hudson River was caused by open vents.

      • I double checked my copy of the A320 QRH and selecting the ditching push button is still to occur at 2000′ AGL on both the ditching checklist (which only applies with engines running) and the dual engine failure checklist. And on the dual engine failure checklist it’s still on page 5 of 5.

        And it’s worth noting that in the case of 1549 it rally didn’t matter whether or not the valves were closed since the impact tore the skin on the bottom of the aircraft leading to rapid flooding anyway. Hence the NTSB’s call to review standards for ditching certification.

  9. I thoroughly disagree with Neutron73: It is not the function of a professional board to limit its evaluation or advise or instruction to a single issue if multiple issues are involved. I think that those board members who did not wish the added recommendation to be made should resign, especially and notably if their concern was to avoid causing distress to Boeing. Surely it is patently obvious that in an age of ever increasing automation, that issues relating to its function must be made public in the same way as failures of an air-crew are.

  10. If it’s a pilot issue plane and simple, then we have a large pilot issue.

    And when things go south, you do not need a system deciding to turn itself off under a layer of confusion.
    The FAA 787 test inspector sited and questioned this same issue on its certification.

    The best gross example of what should be foreseen consequences I know of of was a BA Boeing 747-400 passenger departing Johannesburg. On liftoff the slats retracted, high, hot and heavy and the slats are gone. The pilots limped over Jo-Berg on the ragged edge of stall.

    The cause was the thrust reverser cowl rattled the slightest bit open. The position switch changed state, the slats then retracted.

    The design was to retract the slats to protect the engine from ingesting debris on LANDING.

    What was missing was interlocks to the other thrust revers cowls and squat switch interlock.

    The system worked as Boeing intended. It sure was not what you want, it almost caused an aircraft to crash (its been de-activated).

    AF447: 3 experienced pilots cannot figure out they are in a stall, let alone the Flying pilot put it there to start with for unknown reasons.
    SW Airlines: Landed at the wrong airport
    Dreamlifter: Landed at the wrong Airport
    UPS: Crashed at Birmingham, below glide patch.
    Asiana: 3 pilots, two experienced on type and one lots of flying can’t fly an approach.

    It’s a system issue. Pilots are an end product of the system.

    We now know that many pilots when subject to unexpected situations, lock up (startle reaction per the current speak, freeze up or panic by old school )

    Training has been by rote and wasting training time on normal things (take offs and landings). Tell me an Alaska Airline pilot does not get all he takeoff and landing they need

    Do they auto land? Then the test should be a manual approach in difficult conditions and unexpectedly. On approach, glide slope goes out.

    Upset training is now being implemented

    But from the current results, it looks like a too high percentage of pilots cannot handle unusual (or in many case minor situations). When your airplane dot is saying you are not at your airport and you head down anyway there are major system issues.

    Boeing implements auto throttle one way, Airbus another. Boeings move, Airbus does not, Boeing quits when you think it is on, Airbus does not.

    Each of those system should be run through a non mfg assessment and then the same standards should be on all aircraft across all mfgs. Each mfgs approach has had faults. Human factors should be used, not what we did in the past that has not worked (stall, stall stall or the equivalanat is most often ignored, if its not working, find and install an attention getter that does)

    Design to the worst pilots you allow to fly, not the best.

    The sometimes I feel like a nut and sometimes I don’t is the way holes are gotten into a system and the last line of protection goes away.

    Note: Sullenburger also turned on his APU which was NOT in the manual either for a twin engine loss. that meant he had full control, not partial. That is a thinking pilot but few are that good.

  11. I recall the USN introducing “comic book” instruction manuals at one stage which if memory serves me was for certain engine room procedures.
    Heaven forbid we should have them in a flight manual, but don’t forget the KISS principle.
    Many pilots do not have English as their first language after all.

  12. If a pilot cannot fly the airplane without the help of auto throttle, what is he/her there for? If you have gouges telling you that you are a few miles below minimum safe flying speed on that type of aircraft and you don’t notice it or do something about it, (on landing no less) then why are you there?

    • If a pilot cannot fly the airplane without the help of auto throttle, what is he/her there for?

      That statement misses what happened.
      This is about activated automation working in unexpected ways.
      It is not about being unable to fly without automation.

      • In this case the auto-throttle worked exactly as expected. The problem was neither pilot seemed at all familiar with how the auto-throttle actually worked, so they ended up being surprised.

  13. To me it looks like some limit needs to be placed on an automatic system’s ability to switch itself on and off. AF447 is similar in that the Envelope protection was automatically switched off when the plane went into Alternate law, and the PF was obviously unaware, even ignoring 54?? seconds of stall warnings. OZ-214 the auto throttle was switched off by changing the flight mode, pilot wasn’t aware of it. We need a system where pilots MUST switch these systems off, or acknowledge a message that they are off. Just a proposal?

    • Martin,

      With respect to AF447 with the loss of airspeed data the system had no choice but to switch to Alternate law. Without airspeed inputs the normal system logic that prevents stalls simply won’t function. This was properly annunciated on the ECAM and the crew would have been fully aware of it. Had the crew done what they should have and followed the unreliable airspeed checklist they would have been fine. Instead for reasons that we will never understand the PF decided while at high altitude and near the edge of the airplanes maximum achievable height given their weight to pull the stick back and hold it there. An action which had only one possible outcome and that was to stall the airplane. And yes even prior to AF447 Airbus training has emphasized that yes you can stall the airplane in Alternate law.

      As for 214 it was the pilot who switched the auto throttle off not some random system logic. He did so after selecting an autopilot mode (Flight level change when the selected altitude was above him) which caused the auto throttle system to command climb thrust. However, by disengaging it the way he did he placed the throttles into a hold mode versus a speed mode where they would and did remain at idle power. Now I’m not a 777 guy and it seems there are possibly some traps in the system logic which can snare the unaware (and this crew had a duty to be aware and pay attention and it was their failure to do is which caused the accident) but again the system entered a mode that was commanded by the pilot so it’s pretty clear that requiring some sort of acknowledgement wouldn’t have made a difference.

      • I understand the French psychiatric profession are still trying to figure out what the AF guy was thinking, but it looks to me like somehow he “forgot” there was no envelope protection. What I am trying to say is, in Airbus’ case, they need to find a way to make guys respond, and realise what’s happening. Esp in long haul when they are not fully aware of the situation.

  14. Not to be pedantic Scott, but this was not a VFR flight, it was a “visual segment” of an IFR flight, as happens everyday on a lot of IFR flights, not least in the US. An IFR flight can end with either an instrument approach or a visual approach. It is still an IFR flight. Most operators (at least here in EASA land) will not allow a flight to cancel its IFR clearance and fly VFR. I am quite sure same applies to Asian operators.

  15. While I have not seen the full NTSB report yet, this strikes me as similar to the crash of a 737 at Resolute Bay in the High Arctic of Canada three years ago. That report was published this year (

    The report concludes thta the most probable cause was the First Air Captain had inadvertently disengaged the autopilot’s coupled approach mode and did not grasp that the aricraft was no longer tracking the ILS despite several statements by the First Officer.

    The FO did not use standard terminology and was not aggressive enough, but basically the Captain was not really listening. It’s a puzzle to me why he did not notice loss of the “captured” segment in the approach-progress display.

    (Beware that the presence of a temporary military control tower (semi-operational) is a red herring – but the recorded radar data was very helpful to investigators. The unexplained large errors of the unslaved compass system was a secondary aspect that perhaps added to crew workload. While many theories were hotly debated in online aviation fora, to my knowledge no one came close to what actually happened.)

Leave a Reply

Your email address will not be published. Required fields are marked *