Yeah, this is what I fear might be involved in this crash. In a chain of tiny, otherwise very minor issues (momentary distraction, normal fatigue, ILS out, etc.) it might take only a subtle Crew Resource Management issue to ruin your whole day.
TOTAL HYPOTHETICAL FOLLOWS:
Suppose, for a moment, that Asiana Airlines has its CRM in really good shape across the board. But then one day there’s the “perfect storm” of human nature problems: a specific Captain happens to be just a tiny bit too puffed-up and overly authoritative, and he’s paired with a First Officer that happens to be just one iota too deferential.
The Captain’s dog just died. He spills coffee in his lap 30 minutes before landing and he’s in a foul mood… The First Officer, on the evening before the flight, had been upbraided by his father for talking out of turn and not showing his papa the proper respect.
Landing time. Captain is PIC. He’s coming in just a bit low and a bit slow. The FO wants to say something but the plane’s not that low… it’s not going *that *slow… and the FO “knows” the Captain is uber-experienced and doesn’t need some pissant junior pilot’s reminders of how to land a plane in perfect conditions.
The ground (or water) is coming up fast, the runway’s looking too far away, the air speed is uncomfortably low. Hey… no problem! Captain could do this in his sleep. Hmmm… wonder where I’ll go for dinner tonight. Maybe that place up in… CRUNCH!!!
The above “just-so story” is not meant to be any kind of approximation or guess about what happened on Asiana 214. I’m just pointing out how if everything lines up just right (wrong), CRM can bite you in the tail even when you’re sure that CRM is not an issue (tragic pun intended).
**NOTE: ** CRM might have IN NO WAY been a factor in yesterday’s crash. Any mention that is was is mere supposition by me and/or others.
Question: In a standard “go-around,” are you supposed to pull back on the yoke (nose up)? I assume so (along with full thrust). If so, there must be some limit on how soon before expected touchdown you can initiate a go-around, lest your tail drop so much it clips terrain, right?
We just heard from the NTSB that a go-around was called for 1.5 seconds before impact. Even with the low and slow conditions, might it have been better to try and land – too late for a go-around, period?
Despite something that the narrator is saying, I could see no last-minute pulling up or correction in that video. It just came in too low too fast and clipped the jetty edge, then spun round arse-first before coming to a stop. Thankfully, as has been said, no cartwheeling or tumbling.
Generally speaking, no. I have done many a go-around where the aircraft subsequently touched down after power was applied. This is not a big deal usually as you should be over the runway by the time you are that low.
Now, if you are that low that far out on this runway or a similarly laid out design like runway 22 at LaGuardia then you can be in a lot of trouble.
For that matter, tail strikes on commercial airliners are relatively common as well but they usually happen on the runway and the passengers may never even know it happened. But, hitting a seawall with the tail is a whole different animal.
Okay, now that we’ve done that, where can I get some of this “glideslope”? Is it expensive? What’s the high like? How long does it last? Does it have any other street names?
Whoever. Sorry, but that’s about the tenth time today I’ve seen the wrong case for that particular pronoun.
A note: It’s great that the TM is such a large and diverse group. For almost anything that comes up on this board, there are more than a few professionals and even very competent amateurs who can give excellent commentary. For instance, thanks to all the pilots weighing in on this thread!
Well, that and the fact the lights weren’t on. Both PAPI and VASI are actual lights that are quite visible even on a sunny day. Not seeing them is a strong indication they’re not working.
Correct. You are supposed to increase power then lift the nose. As another pilot already noted, it’s not unheard of for the wheels to touch during a go-around before the airplane gains altitude again. Usually it’s not a problem. Usually.
In this case? Apparently so. There is a point at which you go from “landing” to “accident”. If you come in low enough for the gear to hit the seawall yeah, you’ve probably reached a point of no-return on that one. An airliner is too big, too heavy, and too fast to do well on an off-runway landing.
It actually came in low and slow as evidenced by the high nose attitude. Just before it hits the ground the nose gets raised up higher and the aircraft actually sinks faster. The NTSB have put some information out, but I haven’t been able to get it directly yet, I think it is on Twitter :rolleyes:
Ah yes, here. I’m not normally a luddite or anything but that seems like one of the worst methods of disseminating information via the internet.
Anyway, it seems that the target speed was 137 knots, the throttles were at idle and they got too slow. The stick shaker activated 4 seconds prior to impact and a call to go around was made 1.5 seconds before impact*. When go-around power was commanded, the engines responded as normal. No information yet on which pilot was flying or which pilot made the go-around command.
Questions of interest for me will be:
[ul]
[li]Why were the throttles at idle for so long?[/li]
Aircraft like this normally have the auto-throttle (A/T) engaged and the A/T would be expected to hold the selected speed. Was the A/T engaged? If so why didn’t it do what the pilots expected it to (mode confusion)? If not, why not? My guess is that it was engaged but not in an appropriate mode and the crew didn’t notice it wasn’t doing what they expected it to until it was too late. Some kind of system failure is also possible.
There can also be some strange outcomes due to the automatics doing exactly what they are designed to do but the flight crew not fully understanding the system and therefore not being ready for what happens. As an example, in the type I fly if the airspeed drops to the minimum speed for the flap setting, the autopilot will revert to the level change mode. If an altitude above your current altitude is set in the altitude select window, the throttles will command maximum climb power and the aeroplane will climb to that altitude at the speed set in the speed window, but if a LOWER altitude is set in the window, the throttles will go to idle and the aircraft will descend to the lower altitude. This is pretty much the opposite of what you want to happen when you are already at low altitude and you get too slow. I’m not suggesting that something like this happened to Asiana 214, only that automatics can catch you out if you don’t fully understand them.
[li]Why wasn’t a go-around initiated much sooner?[/li]Airlines have stable approach criteria. For us, we need to be aligned with the runway with landing flap and gear down, on slope (not more than three reds or three whites on the PAPI), vertical speed not greater than 1000 feet per minute, and airspeed not more than the selected approach speed +10/-5. If any of these conditions don’t exist at 500 feet then we have 100 feet to fix it, if it is not fixed by 400 feet we must go around. If any of those conditions are exceeded once we’ve passed the 500 foot “gate” then we must also go around.
I suspect the answers to that question will highlight some CRM factors and perhaps company/country cultural issues.
[/ul]
The scenario put forward by JCFindley of being high on approach and then essentially over correcting is quite possible. Add to that a little bit of confusion over what the auto-throttles are doing (or not doing) and you have the makings of an accident.
Further updates from what I’ve said earlier. The PAPIs were working. The NOTAM about inactive PAPIs was issued after Asiana 214 collided with them :smack:. The PAPI had been inactive on occasion over the last few weeks (as per local pilot reports) but not on this day. So the crew had some external slope guidance. It may still be the case that they were a bit out of their comfort zone with no functional ILS, we shall see.
*From one pilot to the other, not to ATC.
Your point that a glide-slope can and should have been set up on the FMS is valid but the GPS will not give a better glide-slope than the ILS. The ILS glide-slope gives a fixed reference in space relative to a fixed point on the Earth, a GPS glide-slope generally uses barometric altitude information and has the same flaws as an aircraft barometric altimeter. When the weather is bad, an ILS approach is the only approach that can take you all the way down to the runway. GPS approaches typically have a minimum altitude of around 400 feet. That’s not to say you can’t use the GPS generated glide-slope for guidance, but it is not an ILS.
[Quote=Monty]
Monty PIC = Pilot in Command? Pilot in Control?
[/quote]
To add to the good information you already got. The PiC usually means the pilot in command, that is the captain and they are the pilot responsible for the flight. The captain and first officer normally share the flying duties though, so on any particular leg the PF, or pilot flying, could be the captain or the first officer. In a well run cockpit when the first officer is the PF they will typically make the routine decisions in relation to that sector but the captain is still ultimately in charge and will call the shots if the something abnormal happens or if they are not comfortable with the decisions the first officer is making. In this way first officers gradually gain decision making experience and skills so that the eventual transition to captain is a non-event.
Some captains are better at doing this than others.
I’ll finish up with a bit of commentary on automatics. Automatics are excellent for reducing the cockpit workload and are generally very safe. They have become so good though, that crews become very reliant on them and when you fly in and out of major airports every day you can get lulled into a routine of SID - STAR - ILS over and over again. That is, a programmed standard instrument departure through to a programmed standard arrival that sequences to an ILS for the appropriate runway. The whole flight from just after lift off to just before landing (and in some cases till after landing) is flown by the autopilot with the pilots merely monitoring the autopilot’s performance and managing speeds and altitudes along the way. In this environment you are no longer a pilot but a systems/people manager and it is very easy to lose your physical flying skills that got you into that seat in the first place.
Something a lot of pilots do to combat the loss of skills is to fly more of the departure and final approach than necessary themselves. That way they can keep their skills up to scratch. But imagine if your handling skills weren’t all that flash to start with. You might not have a lot of confidence in your own ability to fly the aeroplane smoothly. You might have got through your big jet training by engaging the autopilot at the earliest opportunity and disengaging as late as possible. You end up being the pilot who most needs the practice and the one least likely to get any (conversely the ones who do a lot of hand flying probably don’t really need to practice at all, they just like flying). Then when a day pops up where the automatics either aren’t available or aren’t doing what they should, you don’t have the skills to fix the problem, and you probably never did have them, but you got through the system because the automatics are so good and covered your flaws. Unfortunately these pilots exist in the industry and you have to trust that the airline’s training systems recognise them and either weed them out or correct their flaws.
The opposite can also be true. There are pilots who fly really well when it’s their hands and feet connected directly to the flight controls and throttles but are hopeless at managing the automatics.
CBS Radio just said that early analysis of the CVR shows that the PIC on the landing was doing his first landing at SFO, and that they knew they were too low and slow, and tried to correct.
Offtopic, but one (a smaller one carrying only 10 people) just crashed at an airport in Alaska now. Weird. They say these things always happen in threes…this makes it hard not to put a little faith in it.
In another site, someone claimed that simply increasing power a lot and quickly will tend to lift the nose anyway, especially in the 777 which has two big, “touchy” engines. The place where terrain was hit (the sea wall) might have been the same in any case, but the purported extra nose-up angle (from sudden throttle increase plus, probably, starting to pull up the yoke a bit soon) might have been enough to make it that the TAIL hit the sea wall rather than the landing gear; and this MAY have made the difference between a very hard belly landing (with no fatalities, probably) and the more serious crash which in fact occurred. In other words, maybe the go-around attempt was a bad idea (that late in the game). But, as has been pointed out here, some sort of accident would almost surely have occurred regardless.
Engines mounted below the drag line will cause a pitch up when power is increased (almost all aircraft are like this), however I believe the B777 has fly by wire that prevents those types of secondary effects from power changes.
The plane was capable of creating it’s own glide slope with the GPS and I’m assuming the airlines use WAAS enabled GPS’s which are are a more accurate version of GPS.
It’s important to stay on the glide slope for a couple of reasons.
It creates a stabilized approach which sets up the plane for a good landing. You do the same approach as practiced and the plane acts predictably. If a pilot is not on the glide slope and crap happens then it’s that much harder to correct when outside the normal range of operation which may have happened here. (not saying it was).
Staying on the glide slope helps the plane following behind because a higher/steeper approach will place wing vortices in it’s path.
And PIC stands for Pilot in Command. It is not necessarily the person flying. A senior pilot could be training another pilot and letting them land. The PIC would ultimately be responsible if the other pilot hosed it up. The PIC would be expected to take over control of the aircraft. In this case, flying low and slow would require action before it got out of hand.
I guess I’m relying on systems I’ve seen demo’d but a GPS is capable of extremely accurate placement in relation to anything programmed into the database. I don’t understand what you mean by barometric altimeter as applied to a GPS. as for accuracy, I once used an old Garmin to find a hidden grass strip by programming a waypoint 1 mile out and setting the descent rate to match up to the end of the runway. I was frickin amazed how accurate it was. I was +/- a few feet of the centerline.
What doesn’t make sense in this crash is the airspeed involved. I can’t imagine a commercial pilot making such a huge mistake and there were 2 pilots.
Apologies if I missed this above, but BBC TV is reporting that one of the two deaths was actually caused by an emergency-response vehicle running over her. Ouch!
