EatinRamen

As a 737 pilot myself, I cannot believe that the crew was not monitoring their airspeed. This is one of the basic things you are taught when getting your Instrument Rating: "your scan."

I'm sure some people will try to blame Boeing on this one, but in my opinion, all three crewmembers were not paying attention or were not situationally aware. That's the plain, hard truth. Whether the automation is on, or off, it is our jobs as pilots to monitor what the aircraft is doing.

Our company issued us a memo on this very incident today, and it stated the crew let their airpseed drop to 40 knots below Vref. I hardly ever let the airspeed drop below Vref +5 before I'm on the throttles myself.

The crew may not be here to defend themselves, but if they were, I would tell them they have set the pilot community back, in the view of the flying public, because of their lack of professionalism.

Sniper

I'm not speculating - the Dutch NTSB is the source for my info.

The crew noticed an RA discrepancy according to the CVR and also got a audible gear warning @ 1950' according to the CVR. They then descended 1000'. @ 450', they got a shaker. Until the shaker, they made no control changes or automation changes. Those are the facts, and they are undisputed, no?

During that time, the airspeed tape on both PFDs should have indicted a significant speed reduction, the TL's should have physically moved to the idle stop, and the engine instruments should have all indicated the engines spooled down and were then @ idle and unspooled. I do not know if any of these indications occurred in the accident flight, though I believe all of them should have been visible to the crew if the aircraft was functioning as designed.

Like I said, "I just can't see 3 pilots all noticing a RA/altimeter error and choosing to sit on their hands . . .". and I stick by this. I still find it hard to believe that there were not other factors here, given the many indications they should have received, the assumed 'heightened state' that all aviators go through when approaching the ground with full automation, the 3 sets of trained eyes in the flightdeck, and the status of the Captain and the flight as 'training'. For these reasons, I assume there's got to be unknown factors that lead to this accident (like all accidents, the causal 'chain of events' were each individual action alone is not enough, but it is rather the accumulation and order that allows the unlikely scenario to unfold), and await the findings of the Dutch NTSB.

But, as a pilot, regardless of the ultimate findings, I can learn from the event now and apply it to my operations now to make my flightdeck safer. As both 'georgetg' and 'TurnandBurn' have brought up, it's the danger of the 'perfect storm' - autoland w/ autothrottles engaged, and the aircraft has a RA failure, inducing flare mode while in flight with no flare annunciation. Once you notice it, you advance the TL's, take your hand off, and they go back to idle again. Yikes!

shfo

I hope you don't think 3 miles vis, and 10 knots of wind with an 800 ft ceiling and maybe a little light chop is bad weather.

I agree with all of Sniper's points except the flare annunciation. On a Single CH autoland there will be no flare annunciation. I don't see how three pilots could just watch George take them into the ground. Does anyone actually do an autoland without their thumbs resting on the A/P disc and A/T disc?

joepilot

I haven't flown the 737-800, but in the 737-300 you needed to have both autopilots engaged to make an autoland; a single autopilot could be used for coupled approaches to Cat I minimums. Has this changed for the 737-800?

Joe

georgetg

Boeing designed the SINGLE CHANNEL Autoland as an undocumented fail safe feature, and yes, the -300 will do this also.
Obviously it is not certified to do that but it will do it.

Setting aside crew factors, from a Systems standpoint at issue is the FCC logic that lets the AT command RETARD (at an erroneous -7RA), while simultaneously the AP B is at 1900RA tracking the glideslope. This should not happen.

2 simple software design features could have prevented this accident.

1) RA DISAGREE Annunciation (just like SPD, ALT, AOA PITCH & ROLL have)

2) AT RETARD mode for Autoland disarmed until 50ft below Minimums


From a crew standpoint I also can't see flying an AP approach without keeping a thumb on the AP and a thumb on the AT disconnect SW.
They *should* have caught the speed decay and all of the associated signs.

I'm guessing the various systems messages associated with the #1 RA indicating -7 at 1900AGL probably added an element of confusion at an inopportune time.

From a human factors standpoint I wonder if cultural issues prevented the JS from speaking up if he did notice the speed decay.

Also:
As it slowed the AP would have trimmed to the AP nose up limit.
The whole situation might have been unrecoverable due to limited nose down pitch authority at full power and slow speed and full nose up trim.

This wasn't a simple "what a bunch of idiots situation"
Usually there's more to the story...

Cheers
George

B757200ER

The observer was a 'cadet' pilot and did speak up, 4 times. After the 4th time, the Captain said "My Aircraft", but it was too late. Their speed had deteriorated from 144 KIAS to 100 KIAS and they had 17 degrees pitch; the resultant stall produced the 'falling out of the sky' that witnesses saw.

Remember the old adage--if the automation isn't doing what you want it to, "CLICK---CLICK" . FLY the airplane.

Skyone

There just might have been a cultural thing. In this part of the world, management does not want pilots flying aircraft. I can't speak for Turkish, but only other airlines in the ME. A few years ago Gulf Air flew an airplane into the water (A320) doing a visual 360 after messing up the first approach. I believe the captain was handflying the aircraft. Long and short of it, "automation will be used to the full extent available" philosophy and FOM regulations. Here at my airline, you will receive a letter of reprimand if you hand fly the aircraft above certain FLs or disconnect too soon on an approach. The 777 will "snitch" on you. In fact the letter states that we do not fly aircraft anymore but rather monitor and furthur violations of the FOM will result in blah blah blah.

So back to the cultural thing, it is possible that the over reliance on automation sets these type of accidents up. One's scan is no where as good if every approach one does is coupled down to, say, 200ft. Again, I don't know what Turkish Airlines FOM and philosophies are. But it could be an operations mindset that pilots-bad, automation-good and that's how we train.

__________________
S1

olympic

Relying on automation is always a NO NO in my book. I have caught the Flight Director on days being all over the place, Autothrottle not responding quick enough and airspeed dropping below Vref speed, it's just not 100%.

Regarding Visual Approaches, here in Greece we almost always ask for Visuals when flying to the islands with CAVOK. I usually see foreign operators flying into the islands doing FULL APPROACHES with summer conditions, I laugh at times because the TOWER will clear them for the Visual and they usual don't accept it.

shfo

So what do these pilots (or should I call them passengers) do when the automation fails? Or what if they have to land on the 10s at SFO or land on the 7s at LAS or heaven forbid they have to do the Canarsie visual into JFK? I remember hearing a British Airways 747 declare an emergency when their autopilot failed descending into Florida a couple of years ago.

A couple of years ago I rode in the jumpseat of a UAL 777 from ORD-LAX. The FO had just finished OE and handflew from 32L to 380 and from 380 to touchdown on 25L.

moonkey

I am all about hand flying, but that's just too much and way annoying for everybody else, it actually increases the workload on the PM needlessly...