Final Report on B737 severe upset and Loss of control by F/O

[Centaurus]

http://www.mlit.go.jp/jtsb/eng-air_report/JA16AN.pdf

Above is the link to the Final report on an All Nippon B737-700 JA16AN nosedive from Upset at 41,000 ft on 6 September 2011. Nightime.

The 100-page document also highlights the 38-year-old co-pilot's delayed and confused response, which investigators attributed to gaps in training, undue reliance on automation and seeming anxiety about quickly letting the captain back into the cockpit. According to the report, "excessive dependence on autopilot" exacerbated "lack of full awareness about the need to monitor" flight controls. The co-pilot couldn't recall the stick shaker's activation, it said.

The report said the co-pilot failed to recognize there was a problem for 17 seconds, and then alternately pushed forward and pulled back on the controls. The captain, returning from a bathroom break, was locked out of the cockpit while the plane nose-dived and executed back-to-back rolls in opposite directions. The maneuvers lasted about 90 seconds, though passengers may not have fully realized what was happening because it was dark outside.

With the large number of the Boeing 737 Series operating in Australia, pilots should find time to read this lengthy report. The captain had left the cockpit for a toilet visit leaving the F/O alone up front. He had never been alone up front before, which considering what happened next is probably understandable. To put it bluntly he didn't have a clue. Yet he was certified as competent to be second in command. The starting point for the gyrations of the 737 through the F/O's attempts at recovery was he inadvertently actuated the rudder trim instead of the cockpit door unlocking switch. Angles of bank reached 130 degrees which is close to inverted. This was at night initially on autopilot until the autopilot gave up the ghost and said "All yours" Included in the report was comment that the operator did not include simulator training for high altitude unusual attitudes and stall recovery. Many operators don't either, considering it unnecessary.

[Capn Bloggs]

I had an FO recently who said "but you'd never find yourself there, would you?" as we were dicussing extreme UAs. The point is, you just never know when, how or why...

[ampclamp]

Angles of bank reached 130 degrees which is close to inverted. :uhoh:
He had never been alone up front before,:mad:

How do we find out if someone has no idea until it hits the fan?

[Oakape]

How do we find out if someone has no idea until it hits the fan?

The Check & Training system is supposed to be the filter to weed these problems out. They should never get cleared to line at that level of competence. If they manage to slip through that filter, the on going recurrent training should have exposed the problem. Then it should be either adequate re-training or, if they are unable to reach a satisfactory standard after a suitable period of re-training, the unfortunate situation of letting them go.

Unfortunately the check & training departments of many airlines have been watered down so much that this is what we can now expect into the future. The 'PC' brigade, the 'unfair dismissal' brigade & the 'near enough is good enough' brigade all have to share some of the responsibility. Airline flying is a serious business with serious consequences if you get it wrong, which is true to some extent of all forms of aviation. However, some seem to feel that this is not the case & we can all be a little casual in our approach to it. The old term 'Aero Club with jets' springs to mind.

[Icarus2001]

How do we find out if someone has no idea until it hits the fan?Is that a serious question? Really?

You train and check for these things in a simulator. Simple. We do UA's EVERY sim and the IRT REQUIRES them to be done each year.

[deadcut]

I'm surprised pilots in Asia don't stack their planes more often.

[ampclamp]

Icarus, of course it was a serious question.

Despite what you say there was some guy up front that had serious issues controlling his aircraft.

There are some seriously flawed checks being carried out if this guy is deemed competent.

[Sue Ridgepipe]

There are some seriously flawed checks being carried out if this guy is deemed competent.
Using that logic, you could say the same about JQ after someone ended up in alpha floor trying to conduct a simple change of level.

[waren9]

you sure could.

how many alpha floors have jq had now? 3?

[Tee Emm]

The report states that during UA "training" in the simulator the pilots being trained were told to close their eyes while the instructor or testing officer put the aircraft into a UA. What is the purpose of that? Pilots undergoing initial instrument flying training in (say) Cessna 172 wear an IFR hood so that in theory they cannot sneak a look outside to see the horizon. In addition the instructor gets them to look down and close their eyes for the same reason - to prevent cheating.
But in a simulator where IMC is programmed, the candidate cannot cheat by looking outside. So why is it that some simulator instructors still tell pilots to close their eyes and look down while the instructor flings the simulator around and says "Handing Over, Bloggs - sort out that lot"

[Capn Bloggs]

But in a simulator where IMC is programmed, the candidate cannot cheat by looking outside. So why is it that some simulator instructors still tell pilots to close their eyes and look down while the instructor flings the simulator around and says "Handing Over, Bloggs - sort out that lot"
To test that the pilot can actually work out what the current situation is then recover. If he can't, then he needs more training.

There may not be an outside horizon, but watching the ADI as the simmo puts you in a UA is effectively the same and makes the whole exercise a waste of time.

[Australopithecus]

High altitude, and more importantly, high mach number, upsets are more critical that the low altitude stuff that you see first as a trainee pilot, and later reinforced in stall training.

Angle of attack is critical at the margins of the envelope, but there seems to be a cognitive gap in the average line pilot's appreciation of the values involved. On my own type there has been a well documented fatal stall "accident", yet no line pilot I have conversed with could guess the stall alpha at FL370, Mach 0.82. (It is 4°, by the way, a number that I could scarcely credit when I first heard it.)

Basic understanding of low and high speed buffet, mach effects and other high speed aerodynamics seem to have fallen out of favour, replaced by the false promise of centre-of-the-envelope optimism. You should always aim for the centre, but train for the region outside of the limits.

Many modern aircraft have a device that displays actual aircraft path relative to the longitudinal and horizontal axes of the aircraft. It is called by different names,(FPV, Bird) but it does display the horizontal drift and the gamma. If you know gamma and theta, you can instantly detect alpha, unless of course you allow gamma to get off the scale, as happened in AF447.

My user name pretty much shouts "dinosaur", but at least we dinosaurs knew that we were in need of help, so we poured over "Handling the Big Jets", "Fly the Wing" and "Aerodynamics for Naval Aviators" to name but three classics.

Finally: the 737 rudder trim knob has caused a crash before this attempt. It is not impossible to grasp it by mistake when groping for the cockpit door switch. The tactile clues inherent in slightly different knobs are obviously insufficient for all users. Time for a relocation of the door lock switch to a position not adjacent to the rudder trim.

[Gas Bags]

"you sure could.

how many alpha floors have jq had now? 3?"






How many times have JQ put an aircraft into a golf course???

[Lord Spandex Masher]

Finally: the 737 rudder trim knob has caused a crash before this attempt. It is not impossible to grasp it by mistake when groping for the cockpit door switch. The tactile clues inherent in slightly different knobs are obviously insufficient for all users. Time for a relocation of the door lock switch to a position not adjacent to the rudder trim.

No! Look at the :mad:ing thing before you turn it.

[Australopithecus]

Lord Spandex Masher. Generally that's a useful idea, the looking at things. But not all stressed people perform all the usual steps, as in this case. A very good design would eliminate such areas of tactile ambiguity.

The pilot seat occupier in this case needed quick assistance, not quick and unexpected yaw. I'll bet he didn't have an extra three seconds to look for the knob given the gyrations of the aeroplane.

[By George]

Apart from better high altitude training, one thing that has always puzzled me, is why no angle of attack indicator? I must admit a 4 degree critical alpha is a surprise but so was the height loss in a stalling session just completed in a sim renewal. More education in this area seems to be happening at last, but its been a long time coming.

[Oakape]

The pilot seat occupier in this case needed quick assistance, not quick and unexpected yaw. I'll bet he didn't have an extra three seconds to look for the knob given the gyrations of the aeroplane.

It was the activation of the wrong switch that caused the 'gyrations'. He shouldn't have been that stressed sitting there monitoring the aircraft that he couldn't look at the switch before activating it. If he was, he should not have been on the flight deck in the first place.

[Australopithecus]

Yes, I know. But how many times, honestly, have you seen dumb things done on a flight deck out of well-intentioned but flawed actions?

I have had 737 pilots turn off the B hydraulics instead of engine anti-ice, I have had them do the rudder trim trick, I have had them start to retract the flaps instead of the gear. (Exciting! Fun for all ages!) I am sure that I have pulled some bone-headed maneouvers of my own too.

The Helios Airlines 737 pressurisation crash was due to the commonality of the cabin pressure and take-off warning horn.

My point in all of this is that good cockpit ergonomic design tries to eliminate ambiguity. Thats why gear levers became wheel shaped and flap handles look like airfoils. You can be a he-man aviator all you want. I was that way too, once. Flight Directors? Autopilots? Pffft Kid Stuff! Eventually it sunk into my thick, primitive skull that I was fallible, as are the guys with whom I fly. Hence my comments about the switch design.

[Oakape]

Yeah mate, we have all seen dumb things done on the flight deck. The smart ones learn from it & pass the lesson on to others in an attempt to avoid a repeat. The dumb ones wonder what the hell happened.

The 737 accident you are referring to is the Helios one, not Turkish Airlines.

And if you take a look at the photos in the report you will see that there was significant tactile difference between the two switches in question. Similar to the difference between the flap & gear handles. It would seem however, that there are differences in how various airlines position theses panels in the pedestal. Perhaps there is room for improvement there.

Accuracy is important in this business, hence the comments in my other post.

[Mister Warning]

It's safer to take some FO's to the bathroom with you....

[lemel]

It's safer to take some FO's to the bathroom with you....

Hahahaha! That made me laugh (I am an FO but its still funny… and true in some cases)

[Tee Emm]

To test that the pilot can actually work out what the current situation is then recover. If he can't, then he needs more training.

There may not be an outside horizon, but watching the ADI as the simmo puts you in a UA is effectively the same and makes the whole exercise a waste of time I guess it all comes out to personal opinion. In other words there is no right answer. One could also argue that in a fair dinkum UA it is most unlikely that the pilot would just happen to have his head looking down at his knees with his eyes tightly shut. It is probable that in most of the Loss of Control accidents the pilot could see what was happening a few seconds before it developed into a UA. But for whatever reason, was unable to take prompt appropriate recovery action.

On the contrary, his eyes would have been wide open on stalks, he would be frightened fartless and maybe hands on the controls over-controlling and even turning or pushing the control column the wrong way. Read the Loss of Control in IMC accident reports - it's all there and no shortage of them. Of course, all this is speculation but the point is he sure wouldn't have his eyes wide shut when the stuff hit the fan.

That being so, what is the point in the simulator of telling the bloke to look down and shut his eyes when in real life it would never happen like that? I believe the real reason some simulator instructors do the heads down and shut the eyes caper before a UA, is a throw back to their own experience in general aviation when they had this happen to themselves when training for their initial instrument ratings under the hood in a Cessna. Rightly or wrongly they simply carried that training over to a simulator.. nothing more, nothing less. Just another myth

[bdcer]

Well, in the sim the whole exercise is expected. Closing your eyes also helps create more of a 'startle' factor.

[lemel]

What are you folks on about? You can't honestly tell me that you stare at the EADI for the entire duration of the flight. I think its fair to assume that you take your eyes away from the instrument panel in the cruise quite regularly to do tasks such as fuel logs, read the paper, read the company manual, eat food, etc.

So whats to say that while your not looking at the instruments and you are performing other tasks that the aircraft doesn't enter a UA while your in IMC. Then you would be recovering from the same situation that was simulated in the SIM. Heck, you could be the PM during climb out, look away to do a position report while the other guy is hand flying and by the time you get back to monitoring he/she may be in a UA.

Its an important skill to have - that is to identify a UA on instruments and then initiate the correct recovery.

[boaccomet4]

I am just wondering how many seat cover changes they had to accomplish before the next flight or should it be fright!!

[Judd]

look away to do a position report while the other guy is hand flying and by the time you get back to monitoring he/she may be in a UA.


Which may be why a number of the major airlines around the world, especially the Middle East and Asia discourage any hand flying as they don't trust their pilots not to get into a UA :E

[MASTEMA]

Deadcu*t what sort of exposure have you had to flying in Asia (or the rest of the world) and flying with Asians?

[deadcut]

Am I wrong?

[Old Fella]

Maybe I am missing something, but for the position and type of control switch knobs to be brought into this equation seems to me to be a nonsense. They are positioned aproximately 20 cms apart and are entirely different in shape and function. As I read the report the Rudder Trim control is spring loaded to a central position and requires significant torque to hold it in either direction to make the circuit to achieve Left or Right Trim movement. Regardless, what happened to "Identify before Moving" any switch?

[Australopithecus]

Old Fella, I agree that the generally accepted way to be a pilot includes thoughtful contemplation and due caution. That is not the point. What is the point is that humans are fallible, and it is a worthy pursuit to try to eliminate as many ways of making dumb mistakes as possible.

The industry has moved so very far from where we were forty years ago. At every step of the way there was a naysayer decrying the need for GPWS, EGPWS, enhanced radar, windshear warning, HGS, RNP, GPS, EFIS (for crying out loud) etc etc. What is left is sweeping up the remaining crumbs of ergonomic improvements.

The fact that there have been more than one 737 inadvertent rudder trim command incidents indicates that there is room for improvement in either the aircraft or the operator.

[Lord Spandex Masher]

You can't engineer out all of stupidity though.

I used to operate a new type which had a PA specific button and a radio transmit button separated by as far as possible in the flight deck, completely different areas, one in line of sight and one not, one used with the left hand and one with the right. I still managed one PA on the radio.

How far do you need to go?!

[Australopithecus]

Ergonomics and human factors are not my area of alleged expertise. At this incremental, marginal improvement level there are nuances best left to the design experts. Often a perceived improvement in something that is already almost right results in unintended consequences rearing their ugly heads.

The rudder trim knob is already on its second iteration...I don't think that there are too many variations on cylindrical knobs that would fit. The door release mechanism doesn't need to be a rotary knob, but it does need to be something that cannot be triggered reflexively. The door release switch gets used three times every flight at least...maybe it could be something really unique in a really unique location.

The 737 cockpit is, in the words of an Australian Airlines predescessor, an "ergonomic slum". The grandfathering of a mid-1960's design (which has some commonalities with 1930's aircraft that I have flown) poses a seperate ethical conundrum, but the least they could have done is embrace the present, if not the future, last time around. (I am aware of the contsraints in the scope of improvements in each generation, too. Having flown every variant from the -100 to the -800 I think that the range is too broad for one type certificate.)


As an aside: i am familiar with an early DC-8 accident in which spoilers were inadvertently selected close to the ground with landing flap. The resultant high sink rate ended with a high G impact, a go-around with a disintegrating aeroplane which then caught fire and went in like a lawn dart, killing everyone.

The post-hoc fix was a placard which basically said "do not select speed brakes with landing flaps". The pundits of the day observed that the placard could just as easily have read "do not crash this aeroplane".

Interlocks and guarded switches and the manifold other measures have all been incorporated to protect us and our charges from the tired, distracted, sub-optimum versions of ourselves. I do not think that this is a good time to stop that particular pursuit.

[swh]

In Australia, even the most serious events have not resulted in a loss of control, and were
generally well managed by flight crew. About 70 per cent of stall warnings reported to the ATSB
were genuine indications that the aircraft was approaching the point of stall if the flight crew did
not reduce the aircraft’s angle of attack. The remaining 30 per cent of occurrences were related to
stall warning system problems, although very few of those reported resulted in false stick shaker
activation. Most of the 169 stall warnings reported to the ATSB in the last 5 years that resulted in
genuine stall warning events (usually stick shaker activations) were momentary in duration (lasting
for 2 seconds or less). There were no occurrences reported in this period where a stall occurred,
and no occurrences where the stick pusher activated to prevent a stall occurring. As might be
expected, stall warnings happened in situations where the stall speed increased (due to a
particular (and often unexpected) combination of environmental conditions and flight profile), and
the buffer between the stall speed and the aircraft’s airspeed reduced. The majority of reported
stall warnings (81 per cent) were associated with aircraft tracks in the vicinity of thunderstorms or
other turbulent regions of air, and the greatest proportion of these occurred when the aircraft was
operating at an airspeed close to (or below) the minimum for the current configuration (VRef), at a
bank angle greater than 20°, or when there were sudden and rapid changes in pitch angle or
airspeed.

http://www.atsb.gov.au/media/4359010/ar-2012-172_final.pdf