Yes, most of us have been trained for that scenario as well. I suspect you think of MCAS which is a slightly different animal. I can see how a system that starts and stops, starts and stops can confuse everything. Which it did, but it doesn’t belong in this discussion.

youre obviously an outstanding sim pilot. Maybe too good...

ManaAdaSystem

it wouldn’t stop the engine Captain. It would raise one to whatever thrust is needed on an aircraft with a short rudder arm. It would also make it roll quickly and unexpectedly. We were vfr when it happens to us and we simultaneously got a “wind shear” caution with audible. I’m not a simulator maverick so I probably suck, however, the double startle factor of a caution with a rolling motion made for an alarming scenario that we thankfully caught. Night in a a storm, it might be kind of hairy. Then again, you trained for it so you tell us Captain.

4runner

Does on also see sequences that vary depending on the other autos that may be trying to compensate. What about pitch. etc.

Your scepticism is understandable, but in view of the joint teams invited to participate, one would still expect a fairly decent report within a respectable time frame.

The autothrottle theory has a 99.9% probability of being BS.
If the authorities are seriously considering it , well.............
Boeing autothrottles do not fail to idle in climb.
You can dispatch without it.
On rare occasions they can fail to come up evenly on hitting TOGA on take-off when a clutch doesn’t engage properly, but in climb , never.
How this theory got legs is a mystery.

https://admiralcloudberg.medium.com/...6-e8bbc6683018

A reminder, albeit on a different aircraft, of how a sluggish engine can lead to a chain of events that do not end well.

The later pitch (not AVIAT.) in this thread seems familiar to me ... and I don't like it.
Having said that and not flying heavy iron:
I can understand not noticing split throttles timely.
I can understand not noticing a rudder input by AP (which can be ruled out due to 2 channels only?)
I cannot understand not noticing a significant aileron input by AP during wings level climb.
Maybe some 737 classic driver can elaborate on that?

Exactly what did I say that triggered you in this way?
You don’t see a difference between a AT cutback at low altitude when in a WS, and this accident which happened during climb at 11000 ft?
Well done getting out of your incident, but an engine failure in climb should never lead to an accident. Both throttles are set to climb thrust and one that retards to idle is less severe. Fact!
I have trained engine failures in a lot of diffrent situations. Climb and cruise are two of them. I’m not an ace, I’m just a well trained pilot.

ATR CLUTCH

The ADSB info shows a modest performance level of the aircraft from quite early on after takeoff. The loss of control is abrupt, as it almost always is, but the lead up to it, if due to an asymmetry in thrust is not abrupt, it has been there for a considerable time. If so, there will be some long hard looks at standards.

SITUATIONAL AWARENESS

China Eastern was an abrupt event, the problem arose and caused a quick follow on to loss of control. The crew are startled and also not necessariy aware of the problem, an SA Level 1 type failure.

CI006 engine failure in cruise resulted in a decay of performance and while the crew were deliberating, the aircraft, well above it's 3 engine ceiling, lost speed, and the APLT lateral control diminished to the point the plane rolled over. an SA Level 3 failure of the engine loss, and an SA Level 1 failure related to the APLT aileron authority and airspeed decay.

Endsley (& Jones) came up with the taxonomies of Situational Awareness (SA) failures as:

• Level 1: failure to correctly perceive the information
• Level 2: failure to comprehend the situation
• Level 3: failure to project the situation into the future

Level 1 failures can be cured in time, with warning systems and good training.
Level 2 failures need well-understood warning systems (Helios type problem) and good crew training
Level 3 failures need soul searching by the selection and training and monitoring systems. They are fundamental to the humans involved and are hard to defend against, and are latent failures within our systems at all times.

Endsley evaluated ASRS reports and found that:

• 76.3% were Level 1 SA errors,
• 20.3% were Level 2, and
• 3.4% were Level 3

For one major airline, I evaluated their accident and incident database and found that 97% of all reports involved SA failures of one or more of the types. The only events that did not involve an SA failure of some sort were birdstrikes, and some mechanical failures that were correctly presented, perceived and responded to. On the hull loss cases, which there were far too many, involved almost always multiple forms of SA failure. Possibly the saddest case, an SA Level 3 failure, had 34 occasions in the last 30 minutes of the CVR where the crew showed they recognized the problem, but didn't recognize the impact on their flight path management and parked a billboard on the top of a hill, killing a couple of hundred people, the crew rather surprised when they saw the goats in the clouds. When a robust, reliable countermeasure for SA Level 3 failures arises, aviation will be a lot safer.

APLT CHANNELS

Apart from the curiosity of the A340-600, the majority if not all of the APLT fitted act as 2 axis autopilots, for flight path management. Only yaw damping is usually functioning by a rudder channel, or a torsion/load relief mode as in the B74 rudder. When configured to autoland, a rudder channel may be active for runway alignment manoeuver and rollout control, and a system such as the B777 and B787 Thrust Asymmetry Computer, TAC, may function. On most the yaw damper has a limited turn coordination function as well, however essentially APLTs function in 2 axis for operations, drivers feet still have a use.

SOPs, ECAM, TCAS, GWPS, and other technological advances have made aviation vastly safer than it was a few decades ago. The only component that cannot be improved is the human mind. While these days it is politically incorrect to state the obvious, humans are NOT born equal. Some have better cognitive abilities than others, and even that can be split into many mental and motorical skillsets that make us up as individuals. I for instance can read very fast, well above average, however my handwriting borders on the illegible even with the greatest attention. Any training program by nature is designed for the average, with a line drawn at some minimum requirements below which the less skilled will fail. However there will always be individuals with marginal performance, and it is always a huge dilemma on where to draw the line. The lesson of the past 50 years is very clear. Rather than trying to improve the human, improve the automation to make the task easier. Of course it has taken all the fun out of the job (for those in the upper skills bracket), but it made, and continues to make things a lot safer.

Andrasz

Yes, it is very clear, maintain piloting skills! Virtually every fatal accident in recent times has been caused by pilots that can't fly aeroplanes when presented with a unusual situation.

aterpster

Jepp. I wonder how much line pilots could cope if unexpectedly faced with aerobatic maneuvering in IMC. As far as I know mainly prevention is teached. Once the ball (wing), is dropped stuff can go south in a hurry with no visual references.

Unusual Attitude training in the sim plus a good, practiced, instrument cross-reference should cope with most thrust-roll scenarios (hypothesising here).

EDLB

First and foremost, they should have not let it get asymmetrical. The acrobatic aspect could have easily been prevented.

A question from someone who is not familiar with A/P-A/T technical data : is this failure scenario only possible with older A/T or are modern ones also all subject to this ? and can this also occur on Airbus where the throttles are normally not moving ? Just curious.

It cannot happen on Airbus even if one thrust goes to idle because yaw damper will automatically try to stop the yaw although with limited authority and the spiral will be prevented by opposite aileron and spoilers. It will do a skidding turn to that side with a moderate bank. This is basic demonstration of OEI in simulator.

As noted in previous posts, AT split is unlikely to generate a yawing moment beyond the aircraft capability. Also, this may require a significant change in thrust; e.g. recommencing climb, or intermediate level off, to result in clutch-slip.

The fin is more effective at high speed, vs thrust decrease with altitude.

The AP might have difficulty, but depending on roll power (no rudder servo), yaw - roll should be contained by the ailerons, thus maintaining heading - with slip.

LoC would require aerodynamic input; i.e. mistrimmed aircraft (no rudder or aileron AP autotrim ?). It is likely that the control indications with AP engaged would prompt the crew to add manual trim - with the AP engaged. Rudder trim might not be the first choice, particularly if the hand-wheel angle offset was large, indicating the amount of roll restoring force held by the ailerons; thus aileron trim might be applied.
All would be well until the AP ran out of authority - possible sudden disconnect, and / or the crew had to fly an out-of-trim aircraft without the normal zero-force wheel centre reference - requiring offset wheel force and position to maintain wings level.

I recall that a 737 suffered an upset, similar circumstances, but due to gross fuel imbalance during cross transfer (the aerodynamic / mass disturbance). The AP failed to fly a LNAV turn so the crew disconnected - surprise, big roll, high hand-wheel forces required to recover.
Searching (web and memory) for incident; probably UK AAIB investigation report - 20yrs ago?

safetypee

"I recall that a 737 suffered an upset, similar circumstances, but due to gross fuel imbalance during cross transfer (the aerodynamic / mass disturbance). The AP failed to fly a LNAV turn so the crew disconnected - surprise, big roll, high hand-wheel forces required to recover.
Searching (web and memory) for incident; probably UK AAIB investigation report - 20yrs ago?"

Sounds like this one:

https://assets.publishing.service.go...pdf_501724.pdf

safetypee

Your comments strike me as coming from one who has not flown the 737. If I’m mistaken, I apologize.

A pilot would not add aileron yaw with the A/P engaged. Asymmetric thrust can absolutely cause a roll that would exceed the ability to maintain heading.

That would be in direct violation of a clear limitation for all B737 models, which prohibits the use of aileron trim with the autopilot engaged.

ATC Watcher

I'm going to respond to several different posts here - although this one is the most pertinent.
Speaking Boeing specifically here - Airbus is different. The older cable controlled engines did sometimes suffer clutch slippage - usually due to high down stream cable loads (such as a seized pulley or a damaged push/pull cable). This could case one throttle to either not move, or to move at a slower rate than the other(s). It will not cause the throttle to unilaterally advance or retard - simply fail to move or move at a slower rate as the clutch slips, so it's primarily a concern with large throttle movements. The transition to FADEC has nearly eliminated this issue - FADEC thrust levers do not use a clutch, they have relatively simple friction device to prevent the thrust levers from uncommanded movement due to dynamic loads - there is almost zero force required to move the FADEC lever position resolvers (the friction devices are set to about 2 lbs. force at the lever). However, it is still possible for a FADEC thrust lever to slip during A/T movement if there is some obstruction - e.g. someone drops a pen or other debris into the thrust lever quadrant.

A good accident investigator 'never says never' - however the causes of the rudder hardover and the Lauda in-flight T/R deployment were identified, and corrective action taken via AD over 20 years ago. Since this aircraft was originally owned by United, it's rather certain that the fixes were incorporated, so it's highly unlikely that one of those was the cause. Furthermore, both would be quickly obvious during the initial review of the FDR data. Since there are still thousands of 737-3/4/500 aircraft out there, there would be urgency in getting that cause public so that checks could be made of other in-service aircraft.

I've re-read the last para several times and I don't understand the message and how it links to this accident

Thrust disparity between engines seems to be what investigators are looking at. I wonder if China Airlines 006 may offer a scenario. From Wiki.

I hate to sound a clever dick, but until the FDR is read and there is evidence of an asymmetric condition, then we are all guessing what would have caused the final dive.

That said, on numerous occasions as an instructor during simulator training I have "arranged" a clutch failure in the 737 Classic simulator at certain phases of flight. Sometimes this would occur as TOGA was selected at the start of the run, and other times while the throttles were idle during a clean descent. Then when power was needed, only one throttle would move. Where the students have been briefed beforehand of the symptoms of clutch failure on one throttle and the recovery actions (simply use the throttle manually) we found some students would quickly take corrective action before the flight path got out of hand.

On the occasions where students had not been alerted to an impending clutch failure during the preflight briefing, there were some who recognised there was a throttle problem and used normal airmanship to rectify any roll or yaw. There were others who seemed to have been so bemused that they failed to apply immediate corrective action until it was too late to recover. Read Post 310 as an example. So much depended on the flying ability of the pilot. It was those pilots who were automation dependant who were the bemused.

If the PF doesn't notice split thrust levers (not "throttles") within a few seconds he is not paying attention. Engine instruments all over the place, thust lever position, yoke suddenly on the sosh but above all and instantly by the seat of the pants.
Why? NGs do not have an a/p yaw channel...the resulting slip should be absolutely unmissable to all but someone asleep. It would feel horrible!

Agree, what does the DFDR latch on these machines relative to on or off warnings. It would seem that even if the DFDR is still under review that certain warnings of such magnitude would already have leaked out. I've done some stuff where even before I could make it to the end of the day critical dfdr warnings had already leaked out.

Re A/T in the news,

My biggest gripe with the 737-5 was the very insiduous disconnect of the A/T - no master warning or caution, unlike the A/P disconnect siren. Only a “click” of the switch flicking off. Very easy to not notice that you’d lost your A/T if you weren’t vigorous with your FMA’s AND by keeping your hands on your thrust levers.

at first mention of an A/T issue, my initial thoughts would be a chain of events such as - automatics disconnect, loss of speed, leading to a stall, leading to a loss of control by improper yoke input eg colgan, AF447 etc.

There's that annoying flashing red light in front of you that should grab the attention.

At night, with the lights setting on bright maybe. In daylight, that attention getter is a waste of space, completely unnoticeable in my experience.

I used to fly the classic and one thing I found was that it was hard to work out if you had the right amount of rudder with an engine failed / idle if the autopilot was making a turn. You use the control column being centralised to know if you have the correct amount of rudder. That’s not possible if the autopilot is rolling into or out of a turn.

It would be easier if the slip/skid ball under the EADI was backlit:rolleyes:

Propellerhead you are correct it’s harder in a turn but the slip indicator will tell what you need to know under these circumstances.

Watched an inexperienced first office in the sim recently trying to get the aircraft in trim during a turn. Struggled slightly but the slip indicator was way off centre and that was the big clue as to wether he had applied too much or not enough rudder.

Back to basics which unfortunately don’t get taught these days.

Was this on a NG? On the CL the slip indicator is hard to look at especially in the sim. It's under the EADI but the only reference is the control wheel, which should be kept level.

I don't even know if the slip indicator on the CL simulator is there just as a fitting or if it could be referenced reliably.

Yes it was an NG sim. The slip indicator is a bar under the pointer on the pfd.

Very clear and obvious assuming of course you know to look at it.

This combined with control wheel as you suggest reinforces in trim condition.

Banana Joe

By regulation it should operate correctly

Bloggsie. How right you are but the actuaries don't seem to care.

I think you mean chief executives don't care.

SLF here. I'm trying to wrap my head around discussions here; are posters suggesting this accident might be similar to a single engine loss of power in a light twin, with speed below Vmc and a consequent roll and ....