Hi PJ2. I am assuming this statement was not directed at my previous posting but as a general caution to those who would analyze accidents. I have a copy of Mr. Dekker's field guide in my library. I have seen two accidents long ago that caused me to examine my thinking and I understand well how the cockpit view differs from the external view.

My conclusions with regard to AF447's crew's performance resulted from an analysis about what must happen for success, and then examining the record for evidence of these critical performance activities (and not finding them).

Dekker's new book " " is more likely to be highly relevant to the underlying causes of the AF447 accident. What has me highly concerned is that the same causes are very likely to be found active in other airlines to varying degrees due to the similar operating environment. Guess I know where I should use that Amazon gift certificate I received awhile ago :}.

My statement with regard to backdriving the sidestick should be interpreted as meaning that I believe it is technologically practical to do so based on my slightly outdated understanding of automation devices. It should not be interpreted as a recommendation. I would only suggest it if it was practical to lay a hand on the second stick and perceive the other side's inputs effectively by tactile means. That may require some research.

The PFD is highly cluttered already, and I too question the effectiveness of changing the color of portions of the display to signify important changes in operating mode. A stressed pilot is very likely to miss that type of cue.

ALF, It is a pleasure to have your input again on this thread.

That I think we already know. They were a part of the "Drift into Failure". Where the blame lands is for the courts to decide. However, blame is not relevant to accident investigations in most cases.

Machinbird;

Re, "Hi PJ2. I am assuming this statement was not directed at my previous posting but as a general caution to those who would analyze accidents."

You assume correctly, and my apologies!...not directed your way at all, (as I know you know your stuff), but placed as part of the discussion I wished to have on the topic so that others might take up the conversation as well. I think it is a very timely discussion, given that the report is on its way. The quote I used is exactly as it appears in Dekker's book, including the "(you think)" phrase, (Ch 5/pg39) and it was my unfortunate oversight that did not clarify the quote.
Precisely.

There is no hindsight in examining the record of what is. Hindsight bias is an interpretive gesture, not an investigative act.

Your comments on the SS were taken as intended - as entries into the dialogue and not as recommendations or even suggestions.
Yes, we already know that it is possible for whatever reasons, to miss serious, emergency warnings designed to grab and focus attention. In my view I think this in itself is worth a full examination - why did this occur?

We know already that, first, comprehension through thorough training, and trust, which follows belief in the airplane and its systems and instrumentation, is paramount to safe flight. As many have said, we have here an accident like no other.

The record shows a lack of crew coordination and understanding of the airplane and of high altitude flight. Considering why, is an investigative act. I think the best example on record for highlighting this shift in thinking from thinking about how crews should behave to why they behaved the way they did, is still Diane Vaughn's book because she talks about her hindsight bias first and then shows how her persistent investigative work and open mind made such biases visible and provided the foundations for the shift in perception.
Indeed it is...also, great cartoon.
The following statement is a general observation and is NOT to be understood as referring specifically to any regulator or airline here. The statement carries no meaning other than what it says. That clear, I want to observe that we already know that too comfortable or cozy a relationship between regulator and private corporation is unhealthy for both, and also for the end-users of and in both organizations. It has occurred in the past and has resulted in incidents and accidents. The regulator must be free to enforce or support as necessary but to retain authority of standards. That was my initial reaction to SMS and as far as I'm concerned the jury is still out on the self-regulation of airline safety...another thread entirely.

No other tool that fully visible control columns can better enhance crew coordination - It is all about naturally sharing first hand information - A crew needs sharing, not hiding.

In the case of AF447, I believe a distinction should be made between phases 2 and 3. After the airplane stalled, seeing the control pulled to the back stop with the nose up might have helped to focus the PNF's awareness (and even more that of the captain) on the problem that the airplane was stalled. IMHO in phase 2 knowing the movements of the longitudinal control would not have added anything to the PNF's understanding of what was happening.

from PJ2, quoting Machinbird

Quote:
My conclusions with regard to AF447's crew's performance resulted from an analysis about what must happen for success, and then examining the record for evidence of these critical performance activities (and not finding them).


from PJ2: "Precisely"

agreed.... That sustains a reasonable interpretation of a common metric, to wit:

"What did they know, and when did they know it." Which is a presumption requiring evidence to support, save for some linkage with actual commentary from the pilots....

To fine tune, "What was displayed, available; and what actions can be associated with its acquisition, comprehension, and 'action taken'.

So I will go out on a vulnerable scout, not all three pilots were untrained in probblematic high altitude flight, if we entertain that the two not "inputting" honestly were prevented due to poor design from sussing the control inputs of the PF.

alf, a question? Subject to humor, is hindsight bias a conclusion in itself subject to bias on the part of the judge?

The foresight discussion: "What did they know, and when did they know it"? Is susceptible to unprofessional critique, but I expect many in the Public will have all the bases covered.

I submit that this investigation has the potential, due new technology, to include widespread exposure. That will be good, and it will be bad...

Columbia, and Challenger before it, laid the groundwork for widespread exposure of the stupid things that are done by those previously protected by the culture from disclosure.

We will have to see how everyone reacts. The balance sheet will be written through time, but for me, the worn out buddy sytem and the conspiratorial secrecy that endangers the lives of the public is about to be tested, and there will be many things that are distasteful to know. The good, I believe, will far outweigh the dying rattle of a culture that may be held responsible for needless death, and that by greed.

C-star for dummies?
 

PJ2 says that neutral stick is easy to place, in the mind, by feel.... I believe.

So.... 1) The PF does not know where neutral is, or, 2) he does not know the NOSE position, or, 3) he thinks that minor corrections are needed with the stick back of neutral, similarly, when it is forward of neutral.

Would a Bus driver take the time to interpret for us the stick placements that killed them? Is it reasonable to comment on a partner's use of the Stick, or is it too 'personal'?

Another question that I think has not been addressed: Why is the pilot so stick busy? The aircraft is very large, very gentle in responses, and he looks like he flies a helicopter in a windstorm.....

At the very start of phase 2, the stick is already between neutral and 3/4 back. From my personal experience, I just cannot remember any control column 3/4 back at FL350, except maybe for an approach to STALL simulation ...
To witness such movement is simply shocking.

C-star for dummies
 

HN39 an excellent presentation of how the AF447 crew lost all their energy before the stall. Not as a concentrated pull up, but as an additive process with a nose up bias.

Another indication of a FUBAR scan.

Machinbird;

Thanks, I was having some trepidations about posting.

Machinbird

"but as an additive process..." Could I ask for more? By additive, do you mean he piles on NU on NU? Increasing, without regard to result? You don't mean added to any process of a/c, eg THS? Or other action of which he was unaware?

Would you hazard a possible reason?

I ask because I find it impossible, given the evidence, that any second Pilot would allow the PF to eat paste whilst the a/c climbed as it did, according to the released evidence....

Consider this rhetorical if you like.

I think this "additive process" is in here somewhere.

With momentary audio stall warnings at 2:10:10 & 2:10:13, that graph itself would 'sell' me on an 'intrusive' PFD "PLI" for Alternate & Direct Law.

a brocken rod-spring in the SS was and is seldom, but not impossible to lose the neutral point

HN39;

Agree with Machinbird's comments. Thank you for posting this.

The graph shows clearly, the "time-spent-in-NU-territory" of the control input, and the resultant pitch trend.

Lyman;
How do we know beforehand, that a culture is a buddy system and not an honest system? What are the metrics, beforehand? When do we know that people are "being protected" vice doing their job? How is "conspiratorial secrecy" recognized from "proprietary information" which, in a competitve political economy, is the lifeblood of all organizations? How do we parse and subsequently judge action in different discourses, the most well-known example being the "discussion" between Morton-Thiokol engineers (engineering discourse) and NASA managers at Houston (management/political/economic discourse), regarding the Challenger launch decision. Are the discourses "translatable" into a common language or do we have power politics as the arbiter?

The PF does not know where neutral is
 

implausible with dual cartridges:

http://i474.photobucket.com/albums/r...999/ArtRod.jpg

HN39, okay, thanks very much for the correction and the information. We let go of the stick in the sim exercise but that was in phase 1.

PJ2.

How de we know beforehand? We don't, and that is the plan. Unless the domain is so witless or uncaring, no care is taken to protect eachother.

Oversight, Investigation, Audit, Inventory. These are costly, and so at the very least, profit deters their implementation. I was involved in an ad hoc investigation of seemingly above board process, the mission and its ennablers were clear to see. The laziness with which the chain of incompetence and criminality was covered up was breathtaking.

In aviation, the burden is borne by the common man, through taxes, and the product is consistent with most ventures assigned to the public 'servant'.

Government work.... "You want me to go the Moon aboard a device built by the lowest bidder?"

I have great respect for the FAA, EASA, and BEA, though that may raise an eyebrow or two. By and large, these are competent, and professional organizations. Asx they are public, Politics rears its drooling jaws, and leadership is awarded to whomever purchased more lunches or green fees for the Appointing wonk....

By and large, most of the sheer stupidity and underhandedness is found out, but generally because someone left the lights on....

The larger the consolidated process, the more power, and the more chance for hiding the truth.

Converesely, the smallest organizations tend to be the more corrupt, the consenting few make up the majority of the workforce, and so are well disposed to self protection.

Fukushima is a beacon and clarion. I see ample room for this accident (447) to accomplish some dark corner illumination, you?

Thanks for the action,

A33Zab: Yes, the Stick was fine, I personally am broaching the topic of feel, and PJ2 says Neutral is easy to suss.... So.. He establishes a climb as neutral, and builds and builds....Que paso?

PJ2... "Neutral". Someone needs to at least offer a possibility why PF chose to climb, and appeared not once to be satisfied.....With ADI, and a nosy PNF?

Duff ADI? not the first time, if at all.....What was PNF using then ISIS? Why was PF unaware, demonstrably, of his PITCH, or altitude? And since the climb was NEVER (assimilated) here is the cause, the precise cause. Now we hone in on the possibilities...... If someone, a neutral, is allowed to hear the CVR from 20 seconds either side of 2:10:05, I'll say fine. Complete, tested, and authenticated.

Come on Lyman, BEA playing games?

Lyman;
I am counting on the BEA Human Factors Group for this. Many here have posited serious and plausible theories to account for the sustained NU input. Early on, the Airborne Express DC8 accident was introduced as a possible explanation - the natural tendency to pull back if going down; - the DC8's control column was held in the full-NU position all the way to impact, as was the case with the Northwest Airlines B727 accident. The memorized UAS drill has been discussed, with sub-theories attending to a general inexperience in high-altitude manual handling of transport aircraft throughout the industry possibly pointing to a correct but over-controlled input by the PF and tolerance by the PNF while the PF settled the response down. Inadvertent, unintended NU input while responding to the right roll has been discussed.
In examining the U.S. political economy and events which led to October 2008 and the numerous boom-bust cycles from the Nineteenth Century on, yes, I agree with all your views. Dekker and others have written about "what's common" between aviation, medical and economic "accidents" and I think there is fertile ground for decades of research from which ad-hoc societal change will emerge. The masses will eat cake for only so long, so to speak, but, except for very specific areas of concurrence when it comes to why organizational accidents occur, we are way off topic. I re-emphasize that I think Vaughn's book is the reference work for highlighting the shift from assessements of amoral calculation to the normalization of deviance, which, I believe, is far more prevalent as an operative factor in organizational behaviours, at least in high-risk enterprises, than the former.

@lyman:
 

Ok, and that's exactly reason to investigate the seat/armrest position:

Has there been further consideration of the ACARS messages and are they now fully understood. Will the final report deal with this?

for the caracteristic of the SS the stronger spring must have space for the first 6.5 mm, (your nice picture is not exact in this detail) if the weaker spring was broken then the neutral aerea is between +/- 6 deg of handle deflektion.....

there are two dual cartridges so the risk will be seldom x seldom...... what is very seldom....

grity, For springs not operating in the yield range, the lifetime is very long, in the millions of operations. For a dual failure under this condition, the probabilities suggest that the first failure must not be observed, and the system continues to be operated with a single cartridge providing feel. If this case is valid, then the probabilities of two broken springs on the same flight move into the possible range.

We have to ask the 'Bus maintainers and pilots, the likelihood of a single failure not being written up within a reasonable period of time ~ 6 months max.

Pilots might like a softer feel to the spring cartridge and might not write it up.

Maintainers have periodic checks that they do on everything. The question is not how often this is checked, but how often is it checked using a method capable of reliably detecting a partial spring failure.

I know of no Airbus pilot that would take any chance at all when considering aircraft control. Unless an Airbus pilot is totally inept, he/she is acutely aware of the tenuous link we have to life and won't take any chances. I would put the chances of a pilot ignoring a soft stick at exactly zero.

:eek:

You got that right. There's medication available now. :}

Time 2 10 07

• VS is not zero but minus 500 fpm
• The altitude shows 200 feet below the assigned flight level
• The attitude is below the horizon

Every reason for the PF to initially pull.

Time 2 10 08
Bizarely AP has already disconnected but FD are still displaying, probably for a nose up command.

Time 2 10 17
After dissapearing, they are back, but for a few seconds only, in ALT CRZ*
The altitude is very close to 35000 feet so the FD must indicate a sharp nose down command if the VS is still at a positive 4000 fpm.

Time 2 10 26
Again FD are back, is it for a 10 seconds period, but it must be in VS mode this time.

Before studying anything in that direction, we need to see the Vertical mode trace and analyze how and if the FD may have influenced the PF inputs.

CONF iture
An altitude 200 feet below the assigned level is not in itself major reason for concern and it doesn't look like the PF's reaction to this situation was appropriate. I am also interested as to where you derive your nose down attitude from. There is no evidence that the FD was feeding the pilot false information and you are just introducing another red herring.

BEA IR3 page 88
02:10:00 Pitch attitude decreases from 1.8° to 0° in 3 seconds.
Also visible in the FDR readout on BEA IR3 page 111

You shouldn´t be that fast with red herring.


In FL350 a 0° pitch can be considered a nose down attitude, as the aircraft won´t maintain level flight anymore. The crew had to react, that it ended in an unexplainable overrreaction is another thing.

Even better on page 42.

HN39,

You said C * :
This means that the feedback of the control law changes at a speed close to 200 KT (not mentioned by AIRBUS in our reports, to my knowledge).
To lower speeds the pitch rate is used in the feedback; at high speeds feedback is using G. :p
When the speed is correct, with 3 correct ADR, the smoothing of the two feedback goes well.

But in the AF447, speed selected for that smoothing is either false, or switching from one value to the average of two dissimilar, then the feedback leaps and bounds in the loop! :E
Therefore we must expect APC / PIO (aircraft pilot coupling / pilot induced oscillation) !
1. The initial PIO pointed to by Machinbird may be born of such a discontinuity. :}
2. Other APC have been born then at the option of pitot failures (curve of the three speeds is very discontinuous) creating "non-surmountable" (CVR) instabilities in pitch and loss of control. :E
3. The very rapid oscillations of the parameters from 02:11:40 may perhaps be explained as well. We never got to overcome the possible output of the stall and we never discussed these oscillations that trouble me since I've seen these oscillations.:{

A slight loss in indicated altitude seems to accompany a UAS event.

From the Air Caraïbe A330 UAS Report written by Hugh Houang, January 12, 2008
Translation:
At 22:22:59, a very rapid decrease of CAS, Mach and altitude (Mach correction) is recorded. These parameters respectively decrease to 85kts from 273kts, M.26 from M.80 and 34.700ft from 35.000ft.
. . . .

At 22:24:25, the CAS raises from 111kts to 275kts, Mach returns to M.80 and altitude abruptly raises from 34.200ft to 34.500ft.

My knowledge of these things is limited to the bits I picked off this thread. My understanding is that throughout the speed range the feedback is the response of the airplane in terms of pitch rate and Nz. At low speed the pitch rate is dominant, and at high speeds the Nz is dominant, but both components are present all the time. The response of the airplane in terms of q and Nz is not affected by errors in the measured airspeed.

It was established some time back that the 367 foot difference in recorded altitude was an anomaly. I buy that, but remember the crew had no knowledge of that.

I'll leave it to others to review the data. For the record, the PITCH was accurate, at '0', the altitude was displayed as 200 feet low, and the speed? The Roll correction was met with 'twitchy' response, the airmass was turbulent, the cockpit noise was unknown (sic), etc. Weather, Dark, Surprise, mind how you go.

I'll simply remind everyone that you were not there, and I venture to say that the upset found its beginning in the initial handoff. Look at some of the actual (inertial) vertical data points, remember the conditions, and try to avoid judgment of the outcome based on the initial record......


re: IA What he said, V V V V V V

PJ2;

The 300 ft drop in indicated altitude is explained in IR#2, para. 1.6.11.6 Consequences of a drop in the measured total pressure (p.49). At cruise AoA the pressure at the static port is slightly higher than the true ambient pressure. The measured static pressure is therefore corrected by a term that is proportional to airspeed-squared (or Mach^2). That correction is essentially lost when the pitot pressure drops to a low value.

The TAT/SAT/Mach anomaly is demonstrated very clearly in this ATSB FDR traces graphic of an UAS event to VH-EBA on 29 October 2009.

http://oi48.tinypic.com/2urtro8.jpg

Click on the image to see full size graphic.

please do not set the possibility of any mechanical mistake to exactly zero, as long as no one know "why he had pulled in this way"

mayby it is 1/100000 but truly not zero

grity, Tex was not addressing the likelihood of actual failure, merely in his opinion, he thought an Airbus pilot, if he knew, would write it up, immediately, every time.....

HN39;

Thanks very much for the information - much appreciated.

CONF iture,

From a pilot's p.o.v I think the theory that the reduction in indicated altitude as an initiating source for the immediate pull on the stick by the PF is reasonable in and of itself, however it does not account for the destabilization of the flight path and rapid loss of control that follows. As for the theory of two pilots slavishly following their flight directors to the exclusion of all the raw data being presented, we may posit what we wish but it is not possible to draw any conclusions either way.

No conclusion at this time, but considering both AD + one OEB + the data, one must consider in priority what were indicating the FD when displayed. At this time I see far too little in the 3 IR on that subject … Also, I am still very much concerned we still have to be presented the vertical mode trace ?

Regarding the initial maneuver, it is not only the reduction in indicated altitude, but also a negative VS + an unusual low pitch for a cruising FL.

Blindly following the FD is not a thing to do, but I am afraid signs here could let us think that’s possibly what happened.

@PJ2

however it does not account for the destabilization of the flight path and rapid loss of control that follows.

No. I have virtually always used it is a beginning, however. To connect the initial manual responses to what happened, may not be so inexplicable, with patience, and an objective and sincere attempt to put one's self in the cockpit.

I have. I am not ATPL, but a pilot, and what I feel when I get to Bonin's seat scares the !!!! out of me... I can make it real. I have sat mesmerized while some one on the right tells me something I need to do, and I nodded, and did not do it.... I have gotten behind, in turbulence and inadvertent IMC in a VFR ship, and nearly crapped my pants. I praise God for the quality of the aircraft I have flown, and the forgiving nature built into most a/c.