Need a new thread
 

Trust me, I have not been coordinating posts with PJ2. The nice words from PJ2 and Turbine and some others are great, but sooner or later we have to discuss the very personal opinions of the pilots here with respect to crew coord, warning systems and training.

Correct me if I have some things wrong:

- Some turbulence, but nothing real violent. No SLF's bouncing off the top of the cabin or spilled drinks.

- Airspeed goes tango uniform for some reason, regardless of whether it was icing or a sensor problem by the electrical system.

- System tells crew that airspeed is unreliable or has failed.

- Pilot states that the jet is in an alternate flight control mode and states something like "I've got the plane".

- Pilot sees a need for aileron/roll control and then also pulls up to "x" degrees on the ADI.

- Aircraft climbs and loses velocity ( indicated, true, inertial, whatever).

- warning systems give confusing data

- aircraft momentum allows the jet to drop below an adequate dynamic pressure over the wings to stay outta stall. The "protections" are not working as the AoA inputs are not helping, nor the confusing warning indications.

- aircraft enters a stall, but is still recoverable

- pilot does not appreciate the fact that his inputs are not working like the Airbus public relations brochures claim

- some systems come back on line but the jet is not responding as the PR documents claim, nor all the training in the sim

- aircraft descends for 3 minutes in a stalled condition while crew can't figure out what is happening

- crew never pushes forward long enough, if at all, to unload the wings and gain a decent amount of air molecules over those thingies that provide lift.


So I see a finding of crew error and contributing factors like sensor failure and confusing warning indications and crew training.

Anybody disagree?

can comment, opine on another thread, but we have beat this horse for over two years.

I'm quite happy with that summary gums, seems entirely fair to me.

Yes I've read every post.

Yes I'm SLF.

I'm supprised the manufacturer hasn't adopted "airspeed" in its "unreliable speed" procedure description. Lets face it, ground speed has limited application, i.e. navigational, along with determining windspeed, and taxiing. So to that end, "Airspeed" is the number that counts from V1 in the takeoff roll until passing a similar point in the landing rollout, and it follows that if it becomes "Unreliable", UAS is the acronym of choice.

Reverting back to Robert's comment, "We've lost the 'speeds'" and Bonin's far from unequivocal response, leads me to suspect that Bonin was focused on an earlier discussion with the Capt in which his desire to go higher was rejected on account of the REC MAX limitations due to high OAT. Well, he got his way and went higher, and as we now know the aircraft would have survived at FL380 if it had been "flown" there.

I use "flown" in a sarcastic sense, as the FDR/CVR transcripts (that we have) indicate little evidence of the true meaning of the word.

Passages in the CVR transcript where the words, "You're going up" and "I'm going down?", and similar variations makes one wonder if these two pilots were actually in the same aircraft! It has been stated often in these threads that we do not "know" what Bonin saw on his PFD, but as there were no anomalies on the the PFD's prior to 02:10:05 it is reasonable to expect the same prevailed for the rest of the flight. What we do not "know" and never will, is how Bonin interpreted what he saw. We do "know" what he did. Responsibility for the safe conduct of the flight and what happened that night lies with the Captain, and decisions he made prior to taking his rest period are all relevant.

Hopefully the BEA will have managed to extract some data from the recovered QAR that will provide closure to this matter.

The BEA's factual Final Report will of course lead to the "blame game", where the roll that either of the pilots played in this accident will be sheeted home to rest with those with the deepest pockets. A rather cynical outcome to an accident that I believe ultimately revolves around "human factors" of which all three pilots in one way or another were involved.

gums, good effort; a reasonable summary and findings.
However, to avoid controversy with the use of the ambiguous and often emotive word ‘error’, consider inadequate human performance. Use of this term would provide opportunity to explain (hypothesise) why the human performance was inadequate for the situation. This also avoids the easy blame and train conclusion/recommendation.

Computer trouble re: quote function.

gums, I agree with everything except the part where you state the pilot acknowledges the a/c is in "some alternate control law", and states "something like" 'I've got the controls'. It didn't happen. No one acknowledged ALTLAW at all except PM when he said "Lost the speeds, Alternate Law." This includes the species of ALTLAW (2). That was eleven seconds post a/p loss, and 9 seconds after the Pilot flying stated: "I have the controls". PF did not mention Law degradation, and seemingly performed none of the (rather loose) UAS protocol. How do we know either knew lost AS was the problem until PM states so? We don't know if PF EVER knew that, he did not acknowledge: his only 'speed' reference was to "some crazy speed". (Overspeed?) The working theory prior to PM "Lost Speeds" could be anything, but likely NORMALLAW, no change was announced, and none acknowledged. Show us where the UAS was sussed after loss of a/p with master caution and cavalry charge and before PNF actually speaks it? (Eleven seconds, enough time to seal their fate?)

It is a matter of opinion only whether the STALL was recoverable. It may never be known, including when it "almost did."

This phrasing stands out as really unusual for pilots flying an aircraft. I'm wondering how much of this is from translation out of French, and how much is just plain FUBAR?
I would expect to hear the initial advice from PM to sound something like,"You are 300' high and still climbing." It needs a reference to the altitude assignment.

I'm wondering if PM was trying to avoid further unsettling a guy who was already badly rattled?

That is correct Lyman. PF's initial lateral control inputs in the BEA 3rd report indicates that he was still using the 'pulsing' type technique of lateral control (That we have seen in a number of YouTube videos of Airbus pilot stick technique) after the switch to alternate law. He may have heard PM's comment about Alternate law, but it never carried down to how he was trying to fly the aircraft. Pulsing would be extremely inappropriate in roll direct.

Machinbird

I might misunderstand "pulsing" as you employ it. If he was 'pulsing' in Roll Direct, wouldn't the traces be "serrated" to show input/reverse of input as the a/c stuttered in bank angle? A 'choppy' excursion, not a smooth one? The ailerons would be deflecting, stop, then recommence deflection, creating the exaggerated banking we see?

The stick would return to center after each pulse, yes? That stops the bank in ROLL DIRECT? In the traces, I see banking that is smoothe, once commenced, albeit overdone. It may be that the graph has insufficient resolution.

Lyman, take a look at this excerpt from the BEA report.
Look at the yellow-amber colored stick move display. Do you see the head and shoulders pattern? Or the Big head, little head pattern. That is what stick pulsing looks like. The other side of the zero deflection line, it is inverted, but it is still there.http://home.comcast.net/%7Eshademake...teralMoves.jpg

Machinbird

I do indeed. My point is that the a/c behaved smoothly in roll, as the actual bank angle shows. His stick is not smoothe, but as has been pointed out, it adjusted rather quickly to DIRECT. AB pilots, as can be seen on YOUTUBE, mix mayonnaise on a regular basis. Looking out the front whilst they do, the a/c is on rails, most impressive. To me, it is difficult to trash their stick movement, given the adequate results we see.

Let me ask a question. The a/c rolled through 8.4 degrees from 0 in the time it took PF to "I have the controls". That is the roll he sussed, I assume from his FD. The a/p dropped at 05, and at 07, it was at 8.4 degrees right. He rolled to the left to 6 degrees in two seconds, per the trace by BEA. The a/c then rolled right sixteen degrees in two seconds, before he arrested, and rolled back to the left. This excursion was 17 degrees. We need to subtract some time for the roll to diminish, then reverse.

I'm getting at roll rate. Rapid, No? The a/c's attitude is not known prior to a/p loss (at least on the graph you show).

I estimate seventeen degrees in 1.5 seconds subtracting for lag. That feels quick for a heavy a/c. Does that establish "Twitchy"? I am not sure how "pulse" is risky in that regime, the a/c won't react quickly enough, (as the actual bank shows)? Aileron overloading? Could one or more ailerons have been damaged to establish an out of rig condition that resuted in the chronic roll to the right?

Nonetheless, and it is asking for an opinion only, rolling that quickly felt like ROLL NORMALLAW? How could he have missed it? The PNF was exasperated enough to say: "Watch your Lateral......" to which PF made no response?

Thanks for your response, and your patience with my questions.

Pilot error or pilot inadequacy
 

Good point PJ, and I am guilty of lapsing into the standard litany.

From a personal view as a PLF, I would prefer "error", implying that the pilot made a mistake versus pee poor training or ability to begin with. I want my crew to be at least "adequate", heh heh.

Make no mistake, the 'bus flies very close to some limits of the performance envelope when at high altitude, and the comments about not climbing due to weight and such show this. Until I went thru a lot of the manuals and such, I had no idea of how close to the mach limit at the alt of AF447. So a simple loss of speed indications is cause for very smooth and precise control inputs, FBW or not. Same for control of the engines, i.e. get rid of the auto stuff and keep the things where they were when the problem happened. That aspect of the throttles may come up in BEA's final report.

I'll return to the lurk mode unless I see something startling or a personal attack. @ Okie...... yeah, I could have done well in the heavies, but didn't want the responsibility. In a single seater, all you have to do is save yourself and you don't take 200 folks with you if you screw up or the jet is really FUBAR.

Lyman, for an aircraft the size of the A330, the initial roll and its successors was very rapid. The amount of aileron applied ensured continuation of a rapid roll rate.

There is no reason to suspect any malfunction of the airframe structure, control surfaces, or hydraulic systems in the observed behavior.

Rolling that quickly did not feel like Normal Law, that is why PF had difficulty in moderating his inputs. He had apparently never flown the aircraft in that environment. His initial control input might have been acceptable for Normal Law, but for Alt 2, it was way too large. Once he set the aircraft to flopping its wings, he didn't know how to stop it. What he did was accelerate (but not moderate) his inputs to try to get ahead of the oscillation. What resulted was a decreasing amplitude oscillation that flipped phases on him at least twice. I posted a chart of the roll oscillation back in December. During the data analysis, it became visible that the interval between half-roll oscillations was decreasing from the initial period.
http://www.pprune.org/6908719-post682.html

Is the maximum roll rate of about 11°/s shown on the AF447 traces really so rapid?

In normal law (FCOM 1.27.20) the roll rate requested by the pilot is proportional to the side stick deflection, with a maximum of 15°/s when the sidestick is at its stop.

In direct law (FCOM 1.27.30) the maximum roll rate is approximately 20 to 25°/s, depending on speed and configuration. Spoilers 2, 3 and 6 are inhibited, except in case of some additional failures affecting the lateral control.

Hi HazelNuts39,

All PF's previous experience in Normal Law in roll was mostly useless when faced with Alt Law.

In Normal Law, a small deflection of the SS to the left would have immediately commanded sufficient aileron (up to max deflection if necessary) to arrest the roll to the right and provide a slow roll left.

In Alternate Law a small deflection to the Left would have merely reduced his roll rate to the right. He would have needed more SS deflection to the left than he was used to. Relaxing the SS would have allowed the aircraft to roll right (rather than reduce the roll rate to zero in Normal Law). He has no aileron trim to reset the balance point, and probably only ever used the rudder trim when he was last One Engine Inoperative (in the sim).

He had to relearn Alternate Law handling characteristics quickly. How much he had to learn would depend on the last time he had practised the procedure at Altitude. When was that I wonder?

Once overloaded, he lost the plot with pitch attitude.

Isn't it the opposite?
I'm not sure that he was 'overloaded'. Why would a slow roll to the right or 8° of bank put so much stress on him?

Hi HazelNuts39,
He had probably never handled the aircraft at FL 350 in Alt Law before. In Normal Law, a small ss deflection left would have commanded sufficient aileron to roll it towards wings level, then if he let go of the ss the roll rate would have been zero.

If you examine his ss deflections in Machinbird's post, he was struggling to find the roll rate neutral balance point of aileron deflection. When he relaxed on the ss, the aircraft continued to roll. Simply holding wings level was taking more of his concentration.

Let me drop in for one remark to HN39 quotes:

The cause for trouble is not the roll rate or the bank itself, it´s the rollrate and bank near the maximum of the flight and speed envelope. Using 20°/s roll rate at 5.000 feet might be sporty, using them at FL 350 is imho insane.

Agreed. That's why Annex 3 of Interim Report #3 recommends making "small corrections".

yy
 

It is comforting to consider the lack of excessive g in 447's maneuverings post a/p drop. What may be hiding in the sensed "feel" of less than ultimate g on the a/c are the stresses at the various control surfaces. This is a heavy. The controls are large enough to produce lively excursions in PITCH and ROLL at all altitudes.

I don't know the M/E's here, but I am not one, and am interested in loads, stress, and response; how these parameters can produce damage to the airframe, where local loads cannot be borne. Rapid cycling can damage hinges and skin.

I am not trying to reinvigorate the "Lost VS" kerfuffel, nor the other arguments.
I don't propose the a/c broke, only that it is possible, and that there are hints of the possibility in the data. There are slim chances of proof anyway, to my knowledge, each surface operates in anonymity, though I am sure some doubts can be borne in discussion.

It is true you can lose the largest part of a surface control without generating a single ECAM message :

http://i45.servimg.com/u/f45/11/75/17/84/310_ru10.png

Here it was the rudder but what about one or more ailerons on AF447 ?

The first clue of a possible impairment happened in autoflight. The V/select reported as 5000 fpm descent. If accurate (the a/c did not remain in this config, only commanded, rejected), the a/c was countering a 50 knot climb, or compensating for a chronic misalignment of THS and/or elevators.

As to RUDDER, I believe the pin is actuated within the fuselage; any "position" would be Pin Only, not necessarily to include the Rudder itself. So RTLU reports the drive portion position. Of course, the Rudder is assumed to be attached thereto.

HazelNuts39. At static deflection, yes. In cycling, the loads (calculated) can be much greater? The airplane's stresses are "independent" of a transiting aileron, yes? Are you assuming Still Air? In turbulence (again, calculated) the stress can multiply, of course.

And that is assuming a full complement of control surfaces at less than max deflection. At cruise, fewer surfaces are utilized, and at less deflection (small inputs). If all surfaces are not available, the ones remaining to accomplish the maneuver must be at max. (Spoilers locked out, eg).

Lyman, before you go traipsing too far off into engineering La La Land, let me point out that the aircraft control systems were functioning as designed later in the flight, that the control gains were designed to be appropriate for that portion of the flight envelope and the aircraft was well inside its envelope, and the decision by the engineers was sensible since in Alternate 2 law, the aircraft did not require using as many spoilers based on actual flight tests results and calculations of spoiler effect on wing bending moments.

There is plenty of evidence to explain the performance of the aircraft as the effect of turbulence and pilot control inputs.

Take a look at the Cooper-Harper pilot rating scale located here:fig66
If you were to ask PF right then about flying qualities, he would have been in the 8 or 9 area, but in retrospect we can assign it a 10 (because he lost the plot in pitch)
(rudderrudderrat has the idea.)

Right about now, I anticipate that all the experienced Airbus drivers will have their hackles up and snorting a loud "Bull Sh&t." "I've flown in those conditions and you just have to be gentle with it." And you would be 100% correct. The problem is that PF had never been trained for that corner of the flight envelope and started off on the wrong foot with it. To be specific, he used an improper control strategy. That was the beginning of the end for AF447. There was time to save the day later, but by then, the crew was too rattled to work their way out of the situation.

At least, that is my take on the HF side of this accident.

Lost information like AS
 

Hi,

CONF iture

It is technically feasible (and not costly) to detect this type of failure (on the fly :)).

In new designs and existing (easy retrofit). Sensing aerodynamic pressure "reflected" on actuators. No load (or light one) means less surfaces like your example.

Curiosity: info on this incident, please.

The beginning of the end and remaining chances
 

Hi,

Machinbird:

Quote:

---

To be specific, he used an improper control strategy. That was the beginning of the end for AF447. There was time to save the day later,

---

:ok:

With a better man machine interface this HF "disaster" could be reverted?

Large stimuli to a System near limits
 

Hi,

RetiredF4

Quote:

---

Using 20°/s roll rate at 5.000 feet might be sporty, using them at FL 350 is imho insane.

---

Important point assertively mentioned!

And prior to any other "pitch stimuli"

Showing a completely inadequate handling technique.

With the enormous responsibility yet in the "correction loop" as one crucial "element".

Other than possible wrong indications (due AS anomalies) the displayed bank angle were reliable (and redundant)

GIGO
 

Hi,

Garbage in is not allowed by IT personnel wisely because will generate Garbage out.

Question:

Considering PF is "inserted in the loop" when A/P (and A/THR) drops why PF of AF447 told about "crazy speeds"?

The System presents to the crew, Garbage (AS)?

If IT system is not fed with garbage, why erratic and conflicting data could be fed to "organic processors"?

This is safe during the transients? For example just after Law degraded.

This (garbage to the crew) promotes "graceful degradation" of the "Effective System" (system + PF) or could promote (briefly) accelerated degradation?

No redundance?
 

Hi,

A33Zab

Why in series?

Look, even at high level it's not that difficult, controls are soggy not firm :sad: but hey they still work.

Straight and Level Two as I taught it for years = Power plus Attitude plus Trim = Trimmed Performance.

Never mind complicating the issue by programming in benign failure, all the crew needed to do, even if the autopilot dumped the problem in their laps, was fly the S&L attitude and leave the ****ing power where it was.

That they weren't trained to do so, or that Airbus/AF encouraged a mindset that the software will save you, is the real problem.

Mac, If you will recall, the statement about crazy speeds was likely not the result of the instrument displays, but from the noise of the turbulent flow around the aircraft (due to extremely high AOA). Actual airspeed indications were largely Invalid due to the extremely high AOA, or extremely low-for the same reason.

New thread
 

Hi,

gums

As i understand, when PF is "inserted in the loop" due A/C anomalies, "crew coord, warning systems and training" becomes much more critical. Man machine interface must be very good (and simple up to a certain limit).

THS silently moving, SS of other pilot not viewable, etc. sounds like a threat to the "CRM involving the automatism features of the plane"

Mac.

PS

"CRM" in the sense of Management of all resources (including the auto ones) like THS.

Was it not the sound of the crystalline ice shower? Was it sound at all? Didn't he just "feel like we have some crazy speeds". We are at a linguistic disadvantage,
French is a challenge for me.

EG. If climbing and airmass is supplying some of the energy, rate of climb can be inconsistently sensed? Climbing "too quickly" for the expected g?

A33Zab can follow up if he wishes, but I think I can answer this.

First, there are two heaters in the pitot tube, one for the pickup barrel of the tube, the other for the watertrap area of the tube. The reliability of the pitot tube heater elements is very high. The current is monitored by the control computer. You want to know if there is a heater malfunction in either heating element. When in series, if either fails open, you know about it in a timely manner with a single current sensor. Since you have redundant pitot tubes, a single failure is no big deal by itself.

Pitot heater
 

Hi,

Machinbird:

I would design like that. it is correct.

Thank you

Mac

PS

A33Zab could help us in more important questions :)

i am scanning fast previous posts after a "difficult mission" (trip) and now preparing to post a continuation of the "PF in the loop, transient" issue. Part B will cover from 350 to apogee. When like you precisely said, they had the beginning of the end. :{ by completely inadequate handling at this FL. And Why? Very interesting issue.

Crazy speedS
 

Hi,

Machinbird:

My rationale was based in the speeds (in plural). And also because "crystalline ice shower" IIRC started before.

And,

Occurred IIRC later when A/C was stalled. "turbulent flow around the aircraft (due to extremely high AOA"

The timing must be reviewed to verify. My feeling is of crazy speeds saw in his side and ISIS.


Plz, not consider:


Argument for my rationale:

Quote:

---

Actual airspeed indications were largely Invalid due to the extremely high AOA, or extremely low-for the same reason.

---

So erratic, "viewed" as crazy.

Error!

Crazy speedS
 

Hi,

Bear, :)

Quote:

---

Was it not the sound of the crystalline ice shower?

---

Based in a "pattern" memorized when read pertinent data (essentially BEA) my feeling was PF phrase was due what he was seeing.

This feeling could be verified looking the timing since start of ice crystals noise to the "crazy speedS" phrase.

Mac

PS

I posted before this analysis. Decision making, risk taking.

RR NDB

I think it is very hard to be accurate with this conversation. Firstly, it was hardly "discourse". I think also it was not CRM. If the CVR releases are complete, I would be astonished. I believe much is witheld, and, again, the translation is questionable.

If memory serves, Bonin mentions only "speed". Crazy or otherwise, crazy is not cockpitspeak for any condition I am aware of. PNF is closer to the ideal, but it is not fair to make any finding: conversation.

Entirely lacking context and continuity, until the final is in, this banter is confusing at best!

May your beacon find a direction, bear.

Glitch or intermittent failure
 

Hi,

Bear,

Probably facing a glitch, :)

Certainly.

I was looking to "garbage in" fed to PF. So, "crazy speeds" could be a confirmation for that.

Again very limited factual information (necessary) and others factors like you mentioned.

I will investigate closer this point.

Sometimes our beacon could present "crazy headings". :):}:E:{

Intermittent failures were one of my nightmares. I had one in RR (375 KHz) NDB many years ago. And in the return flight to my base it failed again, :{

Nothing compared to intermittent failures in "feedback Systems" where also the "failure" is feedback to the inputs of the System. :}

Failures in Feedback Systems
 

Hi,

I learned from CJ, on the Concorde YD was slightly disturbed when HF radio was transmitting. Yaw Damper is a good example of a "closed loop" feedback System.

EMI/EMC aspects caused this problems later solved, i guess through shielding, etc.

PF was inside the loop after A/P and A/THR dropped.

As i understand, his "inputs" came from the displays ahead of him and from other "inputs" like noise, G, etc.

His actions with SS were clearly inadequate specially near the "corner". For example, roll.

Question:

His initial "pitch up", the ones that put the A/C in the zoom climb could be justified by a set of erratic "inputs" he saw?

The change of speed to alt could be a "based" decision (acting thinking to be under imminent risk) and putting the plane above REC MAX?

Hi,

About "Crazy speed"
This is the exact transcript of the BEA in report N°3 in french language

J’ai l’impression qu’on
a une vitesse de fou
non qu’est-ce que vous
en pensez ?

I translate this as:
I have the "feeling" (impression) we have a crazy speed ... what do you think ?
So it's an evidence that this remark is not the consequence of instruments reading .. as he use the word "feeling"
what do you think ?
Add to the fact of "feeling" as he ask the other for infirm or confirm his "feeling"
This feeling may be the result of what he hears .. or an internal message of his body

QUOTE JCJEANT
J’ai l’impression qu’on
a une vitesse de fou
non qu’est-ce que vous
en pensez ?

I translate this as:
I have the "feeling" (impression) we have a crazy speed ... what do you think ?

You are perfectly accurate in your translation. I would add that in French," une vitesse de fou" applies to very high, excessive speed. A speed very much above what you would expect, and completely abnormal.
We cannot be sure we have the whole exchange from the CVR...The sentence translated is not technical, factual language normally used in cockpit conversations. It is certainly expressing high stress, and the anxiety of not understanding the situation as far as speed is concerned.