It’ll go to the FBI with the NTSB as a investigatory assistant , but it has to have a prosecutable crime. They’d have to show who did it outside of the crew. If no one is indictable, the US attorney will just pass.

GF

FWIW, from the NTSB Egypt Air 990 Final Report https://www.ntsb.gov/investigations/AccidentReports/Reports/AAB0201.pdf beginning at page 40, if 2 of the 3 PCU's have become disconnected due to "yielded or failed shear rivets in a bellcrank assembly" on one side of the horizontal stabilizer, a full nose-down deflection of the elevator can occur (and was reproduced by Boeing) on the failed side. The opposite elevator would still function, but with one side nose down and the other nose up, aerodynamic forces to correct the nose-down attitude were neutral at best.

Relative to the 3591's gradual recovery near the ground, the fully nose-down deflected elevator would experience "blow-down" at higher airspeeds, thus reducing that aerodynamic force and allowing the functioning elevator to become more functional...

The B747 classic had a stab brake and electrical disconnect of the trim system if the control column was moved more than 4 degrees opposite to stab trim movement - I have no idea how the B767 system is configured, but does anyone with B767 experience know if the trim brake may have been engaged due to the full "pull-up" forces on the control column?

From my POV, the flight profile is consistent with one side of the elevator hard-over (nose down) and, after a brief "WTF?!?" moment and natural response of pulling the yolk back, but only after the aircraft was pitched over (not sure how the autothrottle/max power plays into this other than a rapid increase in airspeed) and no immediate ability to change the pitch attitude other than, as in my previous post, decreased aerodynamic nose-down force associated with elevator blow-back as airspeed increased.

Could the automation issue Feith referred to be that in order to correct this situation would have required that all control column back pressure would have to be relieved before attempting to operate the stab trim to arrest the descent? This procedure was necessary, and not intuitive, to correct a "mach tuck" recovery procedure on the classic B747.

I'm going to guess that the moment column input (nose down) was initiated was 18 seconds before the moment of impact with ground. CVR transcript would fill in a lot (all?) of the holes.

Endeavor wrote:

I'm going to guess that the moment column input (nose down) was initiated was 18 seconds before the moment of impact with ground. CVR transcript would fill in a lot (all?) of the holes.

The initial CVR release said "Crew communications consistent with a loss control of the aircraft began approximately 18 seconds prior to the end of the recording although the overall quality of the audio is considered poor, except when using advanced audio filtering. "
What does that mean? I really don't know.

Something like "Are you pushing the yoke? Because I'm not...."?
Followed by "why isn't it ****ing responding?"

In fact just imagining the situation on that flightdeck is giving me goosebumps.
Maybe it's better if I don't know the precise detail?

jetmatt777 wrote:

That's an interesting thought...the NTSB also says that Houston TRACON told the pilots when they turned west to avoid the weather, to descend "rapidly" to 3000'. It could be possible ( I'm not a pilot, so I don't know) that the Pilot or FO input 30000' thus engaging the thrust to full power and a climb? or the opposite, and a low altitude like 300' was selected, causing a dive?

Perhaps consider for a moment the China Airlines 140 scenario. TOGA was inadvertently activated for some unknown reason. A/P firewalled the engines and the airplane entered climbing attitude. The crew was caught off guard, and forced the nose down in an attempt to continue the approach and satisfy the ATC's instruction to descend to 3,000. This seems to be consistent with Greg Feith's insight that the crew was fighting automation?

SierraPacific wrote:

The change in words from the NTSB is huge since it seems like the initial verbiage referenced some sort of pilot suicide/error while now it reads "in response to nose-down elevator deflection" which seems like it was some sort of mechanical problem that started the catastrophic series of events.

Not necessarily. "Nose-down elevator deflection" is also a way of stating the normal and expected action of the elevator in response to a nose-down control column input...So they could still very well be saying "in response to control column input" without actually still saying "in response to control column input".

jasonelantra wrote:

jetmatt777 wrote:

That's an interesting thought...the NTSB also says that Houston TRACON told the pilots when they turned west to avoid the weather, to descend "rapidly" to 3000'. It could be possible ( I'm not a pilot, so I don't know) that the Pilot or FO input 30000' thus engaging the thrust to full power and a climb? or the opposite, and a low altitude like 300' was selected, causing a dive?

Perhaps consider for a moment the China Airlines 140 scenario. TOGA was inadvertently activated for some unknown reason. A/P firewalled the engines and the airplane entered climbing attitude. The crew was caught off guard, and forced the nose down in an attempt to continue the approach and satisfy the ATC's instruction to descend to 3,000. This seems to be consistent with Greg Feith's insight that the crew was fighting automation?

First it wasn’t a TOGA event because they weren’t coupled to the glide slope and pilots don’t have to fight automation — all you have to do is hit the button and disconnect it — you’re supposed to be paying attention.

SheikhDjibouti wrote:

Endeavor wrote:

I'm going to guess that the moment column input (nose down) was initiated was 18 seconds before the moment of impact with ground. CVR transcript would fill in a lot (all?) of the holes.

The initial CVR release said "Crew communications consistent with a loss control of the aircraft began approximately 18 seconds prior to the end of the recording although the overall quality of the audio is considered poor, except when using advanced audio filtering. "
What does that mean? I really don't know.

Something like "Are you pushing the yoke? Because I'm not...."?
Followed by "why isn't it ****ing responding?"

In fact just imagining the situation on that flightdeck is giving me goosebumps.
Maybe it's better if I don't know the precise detail?

It won't be pretty as you can imagine the enormous tension during those 18 seconds. However in terms of identifying the main cause and/or contributing factors, the CVR might provide significant information.

RyanVHS wrote:

The NTSB just posted an update here: https://www.ntsb.gov/investigations/Pag ... MA086.aspx. One thing to note is: “Also, about this time, the FDR data indicated that some small vertical accelerations consistent with the airplane entering turbulence. Shortly after, when the airplane’s indicated airspeed was steady about 230 knots, the engines increased to maximum thrust, and the airplane pitch increased to about 4° nose up and then rapidly pitched nose down to about 49° in response to column input. The stall warning (stick shaker) did not activate.”

Is it possible someone went to change heading and changed airspeed instead? Engines increase power, pushes nose up, pilot overreacts?

hmmm... think about your question. You are asking if experienced pilots confused the throttles with the yoke/rudder?

WPvsMW wrote:

hmmm... think about your question. You are asking if experienced pilots confused the throttles with the yoke/rudder?

Presumably the previous poster is asking about the various windows and knobs on the mode control panel of the autopilot, not the control column and thrust levers. This is indeed an interesting thought, although I can't imagine that if the pilots were in Heading Select mode (as they likely were if circumnavigating weather) they would want to change the value in the heading window unless intending to make an immediate turn...that is, there is no way to "pre-select" a different heading unless they are using a different lateral mode. But yes, if the PF had mistakenly looked at the speed window and spun the speed knob to a much higher value instead of the heading knob, that could cause the autothrottles to respond immediately with an increase in thrust.

After initially reading the NTSB brief, my mind wandered to the idea of the "head-up" illusion that can catch an instrument pilot off guard, if there are no outside references:

From Wikipedia:
The head-up illusion involves a sudden forward linear acceleration during level flight where the pilot perceives the illusion that the nose of the aircraft is pitching up. The pilot's response to this illusion would be to push the yoke or the stick forward to pitch the nose of the aircraft down.

Since this crew was operating in IMC at the time, it's possible that they felt the sudden and pronounced forward acceleration from the engines going to full thrust (this can be quite remarkable in a lightly loaded aircraft at low altitude) and perceived it as a sharp pitch up instead, leading them to aggressively push the column forward because of fear of an imminent nose-high stall. Of course, one would hope that they would have been hunkered down on their instruments instead of simply reacting by feel alone, but you never can truly predict what a pilot is thinking, and what training or technique they revert to, in a moment of panic. It doesn't take much for an abrupt, extreme improper response to create an unusual attitude that's impossible to recover from at such low altitude.

I agree it could be somatogravic illusion, which would make it similar to the Fly Dubai 737

https://www.popularmechanics.com/flight ... -illusion/

WPvsMW wrote:

hmmm... think about your question. You are asking if experienced pilots confused the throttles with the yoke/rudder?

Don't be a condensing prick. It's a valid question for someone possibly not in the know. After all, the speed and heading inputs on the 767 glare shield are three digit numerical inputs with a round knob underneath, and they are right next to each other.

Very interesting and disturbing news from NTSB and, yet, difficult to decipher. Clearly they know more, and we'll have to wait for the full picture to come out. I'd say deflection was clearly commanded by the pilot(s), because otherwise their initial wording would have been different. However, it is unclear why they pushed the control column. There's only three possibilities really, spatial disorientation, accidental push (illness/fall) or suicide.

Accidental push seems IMHO unlikely; they would have been strapped in at that point in the descent, and a third person was in the cockpit to help in case illness caused one of the pilots to not be responsive.

And suicide is IMHO ruled out by earlier NTSB language about fighting control problems.

So, does this leave only spatial disorientation and confusion then as the initiating action? Can others see any other issue that would fit what we've been told?

AirlineCritic wrote:

...

So, does this leave only spatial disorientation and confusion then as the initiating action? Can others see any other issue that would fit what we've been told?

Mechanical failure of some sort? Hardware failure within the elevator control chain?

BlueberryWheats wrote:

WPvsMW wrote:

hmmm... think about your question. You are asking if experienced pilots confused the throttles with the yoke/rudder?

Don't be a condensing prick. It's a valid question for someone possibly not in the know. After all, the speed and heading inputs on the 767 glare shield are three digit numerical inputs with a round knob underneath, and they are right next to each other.

An accidental input is resolved with a push of a button. No way would an experienced crew not have it corrected within a couple of seconds.

TTailedTiger wrote:

BlueberryWheats wrote:

WPvsMW wrote:

hmmm... think about your question. You are asking if experienced pilots confused the throttles with the yoke/rudder?

Don't be a condensing prick. It's a valid question for someone possibly not in the know. After all, the speed and heading inputs on the 767 glare shield are three digit numerical inputs with a round knob underneath, and they are right next to each other.

An accidental input is resolved with a push of a button. No way would an experienced crew not have it corrected within a couple of seconds.

And yet we argue in other threads that that's exactly what experienced crews (my definition, not yours) didn't do - push the button. Bottom line: Mistakes happen.

The NTSB report is filled with clues that are worth picking apart rather than just focusing on it for confirmation of one's own theory, like the nitwits in the media have done.

You learn about a turn directed by ATC which was executed, a quicker-than-usual descent requested by ATC which they were in the process of executing (if you believe Greg Feith, the descent was commanded through the automation), you know about them being in IMC, you likely know who was flying, you learn about some jostling from a little turbulence (which has had effect in the past: from vestibular system illusions to a misperception of stall buffet), followed by max thrust (either from automation or manual input; if Greg Feith is correct and the autothrottles and autopilot were continuously-coupled during the accident sequence, then necessarily from the autothrottles), which in level flight in IMC is known to have somatogravic effect, followed by a pitch up to 4 degrees (apparently from autopilot if Feith is correct) followed by a control column input of a significant nose-down elevator causing a nose-down attitude of more than 40 degrees, increasing airspeed during the dive, followed by a continuous pitch up from the dive to 20 degrees nose-down before impact. The report says there was no stick shaker, which is a clue that the aircraft systems did not detect a stall at the initiation of the accident sequence. There are at least a couple of lines that could be drawn through these clues and facts that would involve a normally-performing aircraft that was unintentionally-miscontrolled. Folks who fly this aircraft every week are likely well-aware of not-infrequent unintentional mistakes made by themselves or their fellow pilots that don't get spoken of but that could, if lined up into a sequence, result in an accident like this. And I think that's who can benefit from reading between the lines of this report, in combination with Feith's comments (or without), and come up with a few viable ideas of how this could have happened, particularly if one discards the nagging, "But how could they have done (or not done) that?" Fact is that humans are imperfect and on rare occasion we do (or don't), which is why stuff like this doesn't happen often.

Anyway, these are all clues as to where NTSB is currently going with this, but like I said yesterday, they are not going to publicly impugn the pilots so soon after the accident, and without having every I dotted and T crossed. But I think we can put intentional self-harm or a major malfunction of the aircraft now very low on the list of potential primary causes of the accident.

This thread is a refreshing diversion from the admittedly more "serious" MAX threads. Seriously, I feel a sense of peace when I come over to this one. Appreciate everyone's contributions.

One question: Would it be theoretically conceivable that wing icing might have played a role in this occurrence?

Hendric

PlanesNTrains wrote:

This thread is a refreshing diversion from the admittedly more "serious" MAX threads. Seriously, I feel a sense of peace when I come over to this one. Appreciate everyone's contributions.

I will second that, seems a lot of the baboons have taken off to greener pastures, and this thread is actually gaining some quality input again. However, what we do know now after the NTSB release really leaves us none the wiser, I would be very interested in exactly what was on the CVR.

The report says they were or encountered turbulence what are the chances of a micro burst the rapid change in air pressure would defiantly screw with the static readings wouldn't they?
is there a such thing as a still pocket or condition where air is sucked away long enough for an aircraft to loose lift?

cdp wrote:

I just followed the link to read the NTSB statement and it appears the wording has changed:
....The airplane then pitched nose down over the next 18 seconds to about 49° in response to nose-down elevator deflection. .....
It seems they don't want to imply (at this stage) that the dive was manually commanded.

yes
this wording does not say anything about HOW it was commanded

was it commanded by the console with manual input from a pilot or was it automatically commanded by a flight control unit

asdf wrote:

cdp wrote:

I just followed the link to read the NTSB statement and it appears the wording has changed:
....The airplane then pitched nose down over the next 18 seconds to about 49° in response to nose-down elevator deflection. .....
It seems they don't want to imply (at this stage) that the dive was manually commanded.

yes
this wording does not say anything about HOW it was commanded

was it commanded by the console with manual input from a pilot or was it automatically commanded by a flight control unit

Very strange that they changed that wording. "Control column input" means just what it says.

They could have said "commanded by the autopilot, or driven by the elevator servo". When the autopilot servo moves a control surface it just drags the column and control wheel along for the ride.

I'd have to look closely at an FDR readout to see if there is anything that would indicate manual input to the control column. My initial guess is no. They would see elevator surface moving and see if that reading correlates correctly to column position. They also would know at the time of elevator deflection if the autopilot was engaged. So, regardless, I'm pretty confident they know for sure if the autopilot system commanded the elevator or not. That first statement that they made implies to me that it wasn't the autopilot. But, mistakes are made and maybe it really was worded incorrectly.

wjcandee wrote:

The NTSB report is filled with clues that are worth picking apart rather than just focusing on it for confirmation of one's own theory, like the nitwits in the media have done.

You learn about a turn directed by ATC which was executed, a quicker-than-usual descent requested by ATC which they were in the process of executing (if you believe Greg Feith, the descent was commanded through the automation), you know about them being in IMC, you likely know who was flying, you learn about some jostling from a little turbulence (which has had effect in the past: from vestibular system illusions to a misperception of stall buffet), followed by max thrust (either from automation or manual input; if Greg Feith is correct and the autothrottles and autopilot were continuously-coupled during the accident sequence, then necessarily from the autothrottles), which in level flight in IMC is known to have somatogravic effect, followed by a pitch up to 4 degrees (apparently from autopilot if Feith is correct) followed by a control column input of a significant nose-down elevator causing a nose-down attitude of more than 40 degrees, increasing airspeed during the dive, followed by a continuous pitch up from the dive to 20 degrees nose-down before impact. The report says there was no stick shaker, which is a clue that the aircraft systems did not detect a stall at the initiation of the accident sequence. There are at least a couple of lines that could be drawn through these clues and facts that would involve a normally-performing aircraft that was unintentionally-miscontrolled. Folks who fly this aircraft every week are likely well-aware of not-infrequent unintentional mistakes made by themselves or their fellow pilots that don't get spoken of but that could, if lined up into a sequence, result in an accident like this. And I think that's who can benefit from reading between the lines of this report, in combination with Feith's comments (or without), and come up with a few viable ideas of how this could have happened, particularly if one discards the nagging, "But how could they have done (or not done) that?" Fact is that humans are imperfect and on rare occasion we do (or don't), which is why stuff like this doesn't happen often.

Anyway, these are all clues as to where NTSB is currently going with this, but like I said yesterday, they are not going to publicly impugn the pilots so soon after the accident, and without having every I dotted and T crossed. But I think we can put intentional self-harm or a major malfunction of the aircraft now very low on the list of potential primary causes of the accident.

Thanks for this.

wjcandee wrote:

The NTSB report is filled with clues that are worth picking apart rather than just focusing on it for confirmation of one's own theory, like the nitwits in the media have done.

You learn about a turn directed by ATC which was executed, a quicker-than-usual descent requested by ATC which they were in the process of executing (if you believe Greg Feith, the descent was commanded through the automation), you know about them being in IMC, you likely know who was flying, you learn about some jostling from a little turbulence (which has had effect in the past: from vestibular system illusions to a misperception of stall buffet), followed by max thrust (either from automation or manual input; if Greg Feith is correct and the autothrottles and autopilot were continuously-coupled during the accident sequence, then necessarily from the autothrottles), which in level flight in IMC is known to have somatogravic effect, followed by a pitch up to 4 degrees (apparently from autopilot if Feith is correct) followed by a control column input of a significant nose-down elevator causing a nose-down attitude of more than 40 degrees, increasing airspeed during the dive, followed by a continuous pitch up from the dive to 20 degrees nose-down before impact. The report says there was no stick shaker, which is a clue that the aircraft systems did not detect a stall at the initiation of the accident sequence. There are at least a couple of lines that could be drawn through these clues and facts that would involve a normally-performing aircraft that was unintentionally-miscontrolled. Folks who fly this aircraft every week are likely well-aware of not-infrequent unintentional mistakes made by themselves or their fellow pilots that don't get spoken of but that could, if lined up into a sequence, result in an accident like this. And I think that's who can benefit from reading between the lines of this report, in combination with Feith's comments (or without), and come up with a few viable ideas of how this could have happened, particularly if one discards the nagging, "But how could they have done (or not done) that?" Fact is that humans are imperfect and on rare occasion we do (or don't), which is why stuff like this doesn't happen often.

Anyway, these are all clues as to where NTSB is currently going with this, but like I said yesterday, they are not going to publicly impugn the pilots so soon after the accident, and without having every I dotted and T crossed. But I think we can put intentional self-harm or a major malfunction of the aircraft now very low on the list of potential primary causes of the accident.

Kudos! Easily the best summary to date.

you learn about some jostling from a little turbulence (which has had effect in the past: from vestibular system illusions to a misperception of stall buffet),

Turbulence affecting the vestibular system - I know all about that, and some. But that's another story. (I'm not talking airsickness!)

Turbulence to upset the status quo, IMC to void nature's reference points, and just possibly one other factor you haven't mentioned. Was the jumpseater a distraction rather than a possible aid? Nothing deliberately malevolent, but how often is the sterile cockpit rule broken? Would describing 767 procedures and instrumentation to the jumpseater count as acceptable? Or the jumpseater himself commenting "wow, that was some downdraught we just had. I nearly fell out of my seat. Hey, now what's happening..."

Pure speculation on my part; just considering a possible scenario.

wjcandee wrote:

The NTSB report is filled with clues that are worth picking apart rather than just focusing on it for confirmation of one's own theory, like the nitwits in the media have done.

You learn about a turn directed by ATC which was executed, a quicker-than-usual descent requested by ATC which they were in the process of executing (if you believe Greg Feith, the descent was commanded through the automation), you know about them being in IMC, you likely know who was flying, you learn about some jostling from a little turbulence (which has had effect in the past: from vestibular system illusions to a misperception of stall buffet), followed by max thrust (either from automation or manual input; if Greg Feith is correct and the autothrottles and autopilot were continuously-coupled during the accident sequence, then necessarily from the autothrottles), which in level flight in IMC is known to have somatogravic effect, followed by a pitch up to 4 degrees (apparently from autopilot if Feith is correct) followed by a control column input of a significant nose-down elevator causing a nose-down attitude of more than 40 degrees, increasing airspeed during the dive, followed by a continuous pitch up from the dive to 20 degrees nose-down before impact. The report says there was no stick shaker, which is a clue that the aircraft systems did not detect a stall at the initiation of the accident sequence. There are at least a couple of lines that could be drawn through these clues and facts that would involve a normally-performing aircraft that was unintentionally-miscontrolled. Folks who fly this aircraft every week are likely well-aware of not-infrequent unintentional mistakes made by themselves or their fellow pilots that don't get spoken of but that could, if lined up into a sequence, result in an accident like this. And I think that's who can benefit from reading between the lines of this report, in combination with Feith's comments (or without), and come up with a few viable ideas of how this could have happened, particularly if one discards the nagging, "But how could they have done (or not done) that?" Fact is that humans are imperfect and on rare occasion we do (or don't), which is why stuff like this doesn't happen often.

Anyway, these are all clues as to where NTSB is currently going with this, but like I said yesterday, they are not going to publicly impugn the pilots so soon after the accident, and without having every I dotted and T crossed. But I think we can put intentional self-harm or a major malfunction of the aircraft now very low on the list of potential primary causes of the accident.

Excellent clearly explained reasoning, thanks!

This is what happens when you have data: real data. Not just tidbits from FR24, cow farmers, and random comments from an airline CEO.

simking wrote:

The report says they were or encountered turbulence what are the chances of a micro burst the rapid change in air pressure would defiantly screw with the static readings wouldn't they?
is there a such thing as a still pocket or condition where air is sucked away long enough for an aircraft to loose lift?

No. Those things don’t exist and that was not the kind of turbulence that would cause this accident. This could very well have been human failing in automation monitoring. If you read between the lines.

TTailedTiger wrote:

BlueberryWheats wrote:

WPvsMW wrote:

hmmm... think about your question. You are asking if experienced pilots confused the throttles with the yoke/rudder?

Don't be a condensing prick. It's a valid question for someone possibly not in the know. After all, the speed and heading inputs on the 767 glare shield are three digit numerical inputs with a round knob underneath, and they are right next to each other.

An accidental input is resolved with a push of a button. No way would an experienced crew not have it corrected within a couple of seconds.

But what if this crew lacked the CRM to function as a team and pair that with performance issues. I’ve watched in evaluation sessions crews make simple mistakes and compound those with choosing solutions that exaggerated the problem.

I will guarantee there isnt airline in service that doesn’t have a pilot they worry about performance wise. Not everyone is Chuck Yeager and many get lots of additional training to get through programs versus the rest of the group. In many cases the unions are to blame for protecting some of the underperforming pilots.

The „column“ part was just deleted from the NTSB report, for whatever reason.

BlueberryWheats wrote:

WPvsMW wrote:

hmmm... think about your question. You are asking if experienced pilots confused the throttles with the yoke/rudder?

Don't be a condensing prick. It's a valid question for someone possibly not in the know. After all, the speed and heading inputs on the 767 glare shield are three digit numerical inputs with a round knob underneath, and they are right next to each other.

That’s what I’m talking about. Didn’t Helios go down because the pilots couldn’t see a caution light on the panel due to lighting or some such? Task saturated, heavy radio traffic, Wx radar, configuring the a/c...grab the wrong knob. I doubt that’s what happened here but I could see that causing a momentary upset

The rewording of the NTSB update is significant in that they SPECIFICALLY changed "column movement" with "elevator deflection". The follow-up statement "The stall warning (stick shaker) did not activate" is not incidental; to my knowledge, other than the control cables (which drive bellcranks that direct the PCU's to deflect the control surfaces), the only other things that directly move the control column are the pilots and the stick shaker/pusher. (There are centering and feel systems, but their design is not to move the control column, per se)

Previous generation large jets (not fly-by-wire) have cables from the control columns to a bellcrank that directs a PCU to deflect the control surface as required through the control column movement. If the control column is moved but the associated PCU does NOT move (or vice versa), it may cause a rivet connecting the bellcrank to the PCU to shear (by design), effectively severing positive control to the PCU as well as feedback from the control surfaces to the control columns. The elevator system design of the B767 seems to have a history in this regard requiring a few AD's. The triple PCU redundancy on each side of the elevator is designed to let two PCU's overpower a single "runaway" PCU; if TWO PCU's are disconnected through rivet shearing, the runaway condition Boeing warned about is possible.

Given the latest clues from the NTSB update and innuendo by Greg Feith, it is very possible that the combination of a rushed but commanded rapid descent, turbulence and momentary level-off (and slight climb when given a clearance to "expedite descent to 3000") may have caused rapidly changing control column movement and autopilot input disagreement so as to precipitate or further aggravate an already compromised PCU control loss leading to a hardover/nose-down situation.

Having seen the hangar wreckage walk-through video and the position of the stab screw (not near either limit nose-up or nose-down), does anyone with B767 experience know if the stab trim would have been available/operable to counteract the elevator nose-down input with full aft pressure on the control column (i.e. a severly out of trim condition)?

430kts is well into overspeed territory at that altitude, is it not?

OldB747Driver wrote:

...
Previous generation large jets (not fly-by-wire) have cables from the control columns to a bellcrank that directs a PCU to deflect the control surface as required through the control column movement. If the control column is moved but the associated PCU does NOT move (or vice versa), it may cause a rivet connecting the bellcrank to the PCU to shear (by design), effectively severing positive control to the PCU as well as feedback from the control surfaces to the control columns.
...

Thank you for that detailed explanation.

galleypower wrote:

The „column“ part was just deleted from the NTSB report, for whatever reason.

My understanding about what happened there is that when the Twitterverse (Twit-verse) heard about a substantial nose-down column input, they exploded with this supposed "confirmation" that it was pilot-initiated self-harm. NTSB I think re-thought whether they wanted essentially to be describing pilot actions at this point in the investigation, and changed it to elevator deflection, which would say basically the same thing but go over the heads of the Twits, and describe control surface movement rather than inputs to controls. The 100 percent for-sure have data on control column movement from that DFDR, so they know how it was moved, but instead decided to describe its effect.

As mentioned above, I think they're in a position to line things up as a series of mistakes following some initiating event, perhaps the turbulence (which would be the point of mentioning it, like the point of mentioning the stick shaker was to hint that there was no actual stall). Remember, they have all the DFDR data now; they could tell us what every control input was through the whole sequence and what every control surface was doing, what the automation was doing, etc. A significant data dump. But instead they determined to disclose this particular information in this release, meaning that they're telling us that this is what they currently think is material. But because they are smart and cautious, what they think is material could change as the investigation proceeds, so they did it this way. But it's a good solid clue as to the direction they are taking.

(Think about that Ameristar accident. A year ago, they did a significant data dump that all but said the pilots did the right thing. They wouldn't say so then, because it wasn't the proper time or place in the investigation to do so. The media kind of got the hint, but seemed to think it significant that they weren't more definitive about the pilots. When the final report came out, it was as clear and complimentary as to the pilots' actions as any I have ever read, as was Sumwalt. The final report clearly showed that not only did the waaaaaaay-experienced PF make the right call, it made clear that if there had been even a couple of seconds' additional hesitation, or interference by the PM, the math showed, in effect, that a lot of people would have died. Guys had their reputations in the balance for a couple of years without a clear exonerating explanation; I think you're gonna see the same level of caution in the other direction about impugning this accident crew until they can make the determination bulletproof, and until they have in hand all the reports and information from all the working groups.)

wetpantsmcgee wrote:

430kts is well into overspeed territory at that altitude, is it not?

Indeed!

Even with engines at idle, a nose-down attitude of 49 degrees would rapidly increase speed; with the max thrust input, 200 knots in a few seconds? Absolutely.

OldB747Driver wrote:

Having seen the hangar wreckage walk-through video and the position of the stab screw (not near either limit nose-up or nose-down), does anyone with B767 experience know if the stab trim would have been available/operable to counteract the elevator nose-down input with full aft pressure on the control column (i.e. a severly out of trim condition)?

I think it is significant that the stab screw is sitting on the floor of a warehouse rather than in a lab. Its position is apparently roughly where it should be for normal flight. There have been no formal or informal recommendations about inspecting shear rivets or frankly anything on the 767 since the accident and since the CVR and DFDR readouts. Experienced mechanics who work on 767s at major US airlines that fly a lot of them have indicated that they have never seen issues on inspection that would implicate the relevant AD. I kind of line those up and form the impression that, at least at the moment, there is nothing to suggest that a relevant mechanical condition of the aircraft has been identified. And given the existence of the AD and the effects of malfunction that it describes and your excellent description of the method of operation, I would imagine that your exact concern would have been an early question asked by the investigators, because it is a very, very good one.

It is such a pleasure to have a contributor to the discussion who understands so well the system operation of the aircraft that he flew for years.

OldB747Driver wrote:

Having seen the hangar wreckage walk-through video and the position of the stab screw (not near either limit nose-up or nose-down),

I have been wanting to return to that issue for some time now.
Given that we now believe the impression given by the video showing an aircraft trying to pull out of the 49deg dive, is correct... shouldn't this jack screw be at whichever extreme end equates to "pulling up"?
The fact it is in the middle suggests either the pilots were happy with a steep dive into the ground, or...….

SheikhDjibouti wrote:

OldB747Driver wrote:

Having seen the hangar wreckage walk-through video and the position of the stab screw (not near either limit nose-up or nose-down),

I have been wanting to return to that issue for some time now.
Given that we now believe the impression given by the video showing an aircraft trying to pull out of the 49deg dive, is correct... shouldn't this jack screw be at whichever extreme end equates to "pulling up"?
The fact it is in the middle suggests either the pilots were happy with a steep dive into the ground, or...….

A screw would be for trim, not the elevator itself - wouldn't it? (Sorry if I'm ignorantly polluting the discussion with this)

Edit: what I mean is that you're describing "pulling up" on the elevator, which is not the same control mechanism:

SomebodyInTLS wrote:

SheikhDjibouti wrote:

OldB747Driver wrote:

Having seen the hangar wreckage walk-through video and the position of the stab screw (not near either limit nose-up or nose-down),

I have been wanting to return to that issue for some time now.
Given that we now believe the impression given by the video showing an aircraft trying to pull out of the 49deg dive, is correct... shouldn't this jack screw be at whichever extreme end equates to "pulling up"?
The fact it is in the middle suggests either the pilots were happy with a steep dive into the ground, or...….

A screw would be for trim, not the elevator itself - wouldn't it? (Sorry if I'm ignorantly polluting the discussion with this)

Edit: what I mean is that you're describing "pulling up" on the elevator, which is not the same control mechanism:

Please shoot away; I'm here to learn all I can (and very occasionally give something back, I hope)

In attempting to be unambiguous, I may have achieved the opposite.

By "pulling up", I mean whatever actions a sane & rational pilot would make in order to rescue a plane from a dive.
That usually means pulling the control column/yoke/sidestick back.
But it could also mean altering the trim settings, particularly if they were out of normal range to start with.
Presumably CRM would normally result in one pilot performing one of these actions, and the other pilot.....

wjcandee wrote:

It is such a pleasure to have a contributor to the discussion who understands so well the system operation of the aircraft that he flew for years.

Likewise, and I appreciate your insight and perspective as well, which motivated me to jump into this conversation.

Has the NTSB issued a statement yet ruling out foul play? I don't mean one where armchair aviation experts try to rule it out based on Twitter posts, but an actually statement by the NTSB where they state foul play was not involved. How long does it usually take for that to be eliminated as a possibility?

I will probably be burned for being insensitive here but it seems like there is alot of mental gymnastics trying to explain how this could have happened with spatial disorientation, not understanding how automation is affecting something and reacting incorrectly etc., and everyone is trying to avoid talking about or rule out suicide due to the cvr ( and im not declaring this happened or denying the possibility of other causes) but it really could simply be that the pilot flying decided he was done and nosed it over. Possibly the other pilots weren’t paying attention ( long day, not their leg) and gave him is opportunity. This can often be a spontaneous act without a declaration. The cvr is easily explainable if the person didn’t want it declared suicide ( insurance). Make a vague statement about control issues and point the plane to the ground. No time for anyone to react and 18 seconds later its over.
I don’t like it but it really is occums razor

wetpantsmcgee wrote:

430kts is well into overspeed territory at that altitude, is it not?

Yes- 350 would be approximately Vmo, or max airspeed, at that altitude. They were in the red part of the airspeed indicator by close to 80 knots. That’s exceptionally overspeeding.

SheikhDjibouti wrote:

Given that we now believe the impression given by the video showing an aircraft trying to pull out of the 49deg dive, is correct... shouldn't this jack screw be at whichever extreme end equates to "pulling up"?
The fact it is in the middle suggests either the pilots were happy with a steep dive into the ground, or...….

My impression is similar to most observers here (and I am in agreement with you, wjcandee, that its presence on the floor indicates its functionality was not suspect) that the jack screw is in a "normal" position and displays no damage that might indicate its ability to function was other than "normal" - but also that it was not utilized to attempt pitch recovery, i.e. regardless of which direction indicates nose-up trim, it does not appear to be at its limit of travel in either direction. Which is what makes me wonder if either of the pilots attempted to use the stab trim in the very brief 18 seconds they had to contemplate what must have been a bewildering scenario aggravated by negative g's and a windscreen filled with the ground; my heart goes out to these three crewmen.

Going back to my experience with different aircraft, in order to prevent a runway stab trim, if a certain amount of pressure were applied to the control column (indicating an severe out-of-trim condition, such as would be expected with a runway stab trim) an independent stab trim brake would prevent the jack screw from moving. In this particular case, if such a system exists on the B767, that safety feature might have prevented the pilots from trimming the nose up due to the very normal reaction of the crew to pull back on the control column in reaction to a rapid pitch down of the aircraft due to an unknown cause.

OldB747Driver wrote:

Going back to my experience with different aircraft, in order to prevent a runway stab trim, if a certain amount of pressure were applied to the control column (indicating an severe out-of-trim condition, such as would be expected with a runway stab trim) an independent stab trim brake would prevent the jack screw from moving. In this particular case, if such a system exists on the B767, that safety feature might have prevented the pilots from trimming the nose up due to the very normal reaction of the crew to pull back on the control column in reaction to a rapid pitch down of the aircraft due to an unknown cause.

I believe the feature you are describing only prevents stab movement opposite control column deflection. If the pilots were pulling back on the control column, the stab trim would still operate in the acft nose up direction, but would be prevented from moving in the acft nose down direction.

IF the excursion started in a trimmed condition,230-ish knots, how could the plane reached 430 knots without some pretty high Gs pulling the nose up unless the autopilot was furiously trimming to keep up with the speed increase?


Gf

OldB747Driver wrote:

The rewording of the NTSB update is significant in that they SPECIFICALLY changed "column movement" with "elevator deflection". The follow-up statement "The stall warning (stick shaker) did not activate" is not incidental; to my knowledge, other than the control cables (which drive bellcranks that direct the PCU's to deflect the control surfaces), the only other things that directly move the control column are the pilots and the stick shaker/pusher. (There are centering and feel systems, but their design is not to move the control column, per se)

Previous generation large jets (not fly-by-wire) have cables from the control columns to a bellcrank that directs a PCU to deflect the control surface as required through the control column movement. If the control column is moved but the associated PCU does NOT move (or vice versa), it may cause a rivet connecting the bellcrank to the PCU to shear (by design), effectively severing positive control to the PCU as well as feedback from the control surfaces to the control columns. The elevator system design of the B767 seems to have a history in this regard requiring a few AD's. The triple PCU redundancy on each side of the elevator is designed to let two PCU's overpower a single "runaway" PCU; if TWO PCU's are disconnected through rivet shearing, the runaway condition Boeing warned about is possible.

Given the latest clues from the NTSB update and innuendo by Greg Feith, it is very possible that the combination of a rushed but commanded rapid descent, turbulence and momentary level-off (and slight climb when given a clearance to "expedite descent to 3000") may have caused rapidly changing control column movement and autopilot input disagreement so as to precipitate or further aggravate an already compromised PCU control loss leading to a hardover/nose-down situation.

Having seen the hangar wreckage walk-through video and the position of the stab screw (not near either limit nose-up or nose-down), does anyone with B767 experience know if the stab trim would have been available/operable to counteract the elevator nose-down input with full aft pressure on the control column (i.e. a severly out of trim condition)?

wjcandee wrote:

The NTSB report is filled with clues that are worth picking apart rather than just focusing on it for confirmation of one's own theory, like the nitwits in the media have done.

You learn about a turn directed by ATC which was executed, a quicker-than-usual descent requested by ATC which they were in the process of executing (if you believe Greg Feith, the descent was commanded through the automation), you know about them being in IMC, you likely know who was flying, you learn about some jostling from a little turbulence (which has had effect in the past: from vestibular system illusions to a misperception of stall buffet), followed by max thrust (either from automation or manual input; if Greg Feith is correct and the autothrottles and autopilot were continuously-coupled during the accident sequence, then necessarily from the autothrottles), which in level flight in IMC is known to have somatogravic effect, followed by a pitch up to 4 degrees (apparently from autopilot if Feith is correct) followed by a control column input of a significant nose-down elevator causing a nose-down attitude of more than 40 degrees, increasing airspeed during the dive, followed by a continuous pitch up from the dive to 20 degrees nose-down before impact. The report says there was no stick shaker, which is a clue that the aircraft systems did not detect a stall at the initiation of the accident sequence. There are at least a couple of lines that could be drawn through these clues and facts that would involve a normally-performing aircraft that was unintentionally-miscontrolled. Folks who fly this aircraft every week are likely well-aware of not-infrequent unintentional mistakes made by themselves or their fellow pilots that don't get spoken of but that could, if lined up into a sequence, result in an accident like this. And I think that's who can benefit from reading between the lines of this report, in combination with Feith's comments (or without), and come up with a few viable ideas of how this could have happened, particularly if one discards the nagging, "But how could they have done (or not done) that?" Fact is that humans are imperfect and on rare occasion we do (or don't), which is why stuff like this doesn't happen often.

Anyway, these are all clues as to where NTSB is currently going with this, but like I said yesterday, they are not going to publicly impugn the pilots so soon after the accident, and without having every I dotted and T crossed. But I think we can put intentional self-harm or a major malfunction of the aircraft now very low on the list of potential primary causes of the accident.

Both of these posts are the most well thought out posts I've seen so far in this thread.

One thing that I think is lost is just how disorienting the whole event would have been. 18 seconds is not very long and getting the nose down to 49 degrees nose down would have taken IMHO at least 3 or 4 seconds...maybe more. I don't have the engineering to know just how severe what I'm about to say would have been, but I would think there would have been significant negative Gs while the 767 was pitching over. These guys while trying to sort things out were likely "hanging in their belts" even though right side up. And I would think experiencing a lot more than just weightlessness.