I have never fllown with a side stick but no Boeing pilot at FL350 would pull back to the stops if he lost his airspeed and autopilot, why would an Airbus guy do it? No monitoring pilot would let the pilot flying do it either. Is this really how Airbus pilots normally fly?

@ bubbers, et al

Yep, beats the hell outta me. Hold what ya got and analyze the situation.

maybe the "boldface" or whatever the commercial folks call it was not a good idea - like pull to 5 degrees of pitch and so forth.

For others, re: elevator versus stick inputs

Good grief, we've been thru this. The position of the control surfaces on a FBW jet do not reflect the pilot inputs via sidestick, wheel or whatever. They move to achieve the commanded gee, AoA or roll rate. The rudder will move to help roll coordination, with no input from the pilot.

In the 'bus, the THS will gradually move one direction or the other to reduce the amount of control input to achieve the "commanded" gee.

Waiting for the final and the immense amount of comments from we in the peanut gallery shortly thereafter.

The BEA's English translation of the recommendations issued following their report into the Tarom incident is as follows:-As far as I can ascertain, those basic recommendations have been fulfilled, e.g.

• The A/P and A/THR disconnected when a couple of ADR's had a disagreement over speeds.
• The THS trim continued the way it was designed to, though due to a change in the control law, it would appear that the crew had no idea that the Alpha protections had been lost.
  • Except when the Abnormal Attitude Law has been triggered, auto trim is still functioning and a constant stick NU/ND command will cause the THS to follow in an attempt to maintain a load factor of +1g - be aware!
• The Stall Warning operated correctly while the aircraft was within the Normal Flight Envelope, but wasn't heeded.

In short, the visual and audio clues/warnings provided, possibly added to the cognitive overload. Would a pseudo artificial feedback to the SS helped?? Perhaps the centrifuge simulators of tomorrow may help to reinforce what it feels like when the aircraft is maneouvered in abnormal conditions.

Bubbers, that's *precisely* what the Birgenair 757 captain did.

Not at FL350 he didn't. which maybe means bubbers wins on a technicality.

The point, however, remains that LOC accidents, and specifically mishandled stalls, are happening across range of types/mfrs/airlines.

If it was just 'buses falling out of the sky I'd be clamouring for investigation of the bus specific sidestick and fbw etc. - but that isn't what is happening.

If the effort were focused on the bus stuff only, then we could only ever fix part of the problem and risk fixing nothing if in fact the causes are elsewhere and common across types. Far better to try and identify common causes, surely ?

Role of A/C
 

Pariès cited 16 events similar to AF447, all of which showed poor understanding, rare implementation of unreliable airspeed procedures and stall warnings which were "perceived but mostly not believed".

Granted, but in that case the Birgenair jet never made it to cruise level. While it doesn't satisfy all of the criteria Bubbers set, it satisfies the important ones - namely doing precisely the wrong thing minus speed indications and with autopilot behaving erratically. It also puts a dent in the "other pilot would see the inappropriate control deflection and take over", because in that case the F/O did not. The Stony Point NWA B727 was at FL 248 when exactly the same thing happened.

Not THIS again!!!
 

Dozy:
You keep on saying that, and I keep on countering that your premise is based on false logic. One or two cases proves nothing at all. One could (I do) also speculate that, if one had a data base of, say, 100 such incidents, equally distributed between SS and yoke aircraft, it might be more appropriate to draw the conclusion that you do IF there was no difference in the associated statistics. I.E., if we're going to make guesses on this issue, I would guess that in 100 cases of LOC-to-stall, MORE pilots would notice the inappropriate input when it was a yoke. But....I have no more proof for that than you do for your opposite conclusion. :ugh:

True, I can't fly, but I can think! :\

But this guy was an Airbus pilot, and under Normal Law, isn't pulling the side stick to the stops quite permitted because Alpha Protect would limit any pitch change to prevent LOC? (Indeed, I thought such such flight protections were one of Airbus' avowed selling points.)

So what about the possibility that the PF simply did not process/comprehend that when the A/P went into Alternate Law, that Alpha Protect was no longer active -- and that he would need to use the side stick more like a Boeing yoke than an Airbus joystick? Thus, it wasn't that he was dumb, just that he had a brain freeze in not fully comprehending that his protections were gone (despite this warning being given on the ECAM).

Just catching up on the discussions.
I found the Jean Pariès presentation very interesting. In the conclusion he recommends an "Overall paradigm shift" in design and training. Well, how about giving the crew a synthetic view of the aircraft generated from the sensors, something like the MS Flt Sim outide spot view, say 3/4 rear view. We see this kind of display on visualisations of FDR readouts. It would give the crew (when at their most 'startled') an immediate understanding of the aircraft's pitch/roll attitude relative to the horizon, and also (if it were possible to generate) their FPV in 3D.

Dozy,

The same thing did not happen except for the end result. There was a different reason. The NWA B727 crew forgot to turn on the pitot heaters. So when they iced over, all indications were that they were all of a sudden speeding up, therefore they reduced speed. There was nothing to tell them they were at the right speed in reality. The speeds they were seeing on the instruments were erroneous, they didn't go away and they appeared real to them sadly.

I think caution needs to be applied when comparing accidents of this type, one 3 years ago and one ~ 40 years ago.

"Overall paradigm shift" in design and training.
 

Hi,

SRMman:

Long time ago some of us raised the issue. A very serious one.

Unfortunately i am too busy to do more in this Thread:

http://www.pprune.org/tech-log/48135...anomalies.html

Initial cause of the pitot failure was not the same but after that it's all depressingly similar.

You could write those statements about either accident.

Yes, there is a big difference in aviation between 40yrs ago and 3yrs ago but that doesn't mean history provides nothing to learn from. The problem of crews failing to recognize stall, pulling up into stalls, pulling back through stall warnings, and failing to correct the other guy doing it is not new and didn't start with fbw and sidesticks.

IF789 said:
You're sure right about that, but some of us are suspicious that sidesticks could make this problem harder to suss out under pressure, since the other guy can't see them.

=this has been a pre-recorded announcement=

:ugh: :ugh: :ugh: :ugh: :ugh:

Admittedly that's a good point taken in isolation, but as you so eloquently put it, the amount of evidence available is not sufficient to make a statistically valid conclusion. "Suspicions" based on an assumption that something stands to reason are just that - opinion and conjecture.

The fact is that whether you can see the sidestick or not (and FWIW in the sim I could get a good idea of what the guy in the opposite seat was doing with the stick based on his posture), a combination of monitoring the aircraft's response plus timely and effective communication with your colleague in the opposite seat should nullify that potential problem.

@ DW
 

Good! Then I'll stop maintaining that yokes would have helped, if you'll stop saying that they wouldn't. :p The only thing for sure is that we don't know.

Indeed it should, but back to the real world-- that just didn't happen this time. Even if some pilots shouldn't be fools, the fact remains that some are, so "foolproof is better." If one outta 100 pilots need a situational icon showing attitude, let's just give 'em one! Hang the expense.

=I HAVE SPOKEN= :8

Camparing pears with apples, do you?
 

@DW

As far as I remember the Birgen Air accident was mainly caused by failure of shutting off the A/T system, which by design took thrust to idle because of the false overspeed sensed by blocked pitot. The FO very well was aware of the situation, but it seemed to be also a cultural issue that he did not take control.

So IMHO this has very little resemblance with a yoke versus SS issue.

Why do I sense some reluctance throughout all your postings to realize that humans not always react the way that engineers plan and sometimes design the human - machine interface the way that the human has to configure himself to the machine instead the other way around?

Re Overall paradigm shift" in design and training.
Pariès expands paradigm shift elsewhere as “the change from safety through conformity, to safety through variation management”. Variation is described by Hollnagel:- http://ipac.ca/documents/The_resilie...ion%5B1%5D.pdf


"… other pilot would see the inappropriate control deflection and take over"
I doubt that there would be a fair range of data to support either view, but anecdotal evidence suggests that pilots don’t ‘see’ a control input as inappropriate, particularly where both have similar but erroneous situation awareness. Aspect of this are discussed in http://www.pprune.org/safety-crm-qa-...ervention.html.

In addition, it would be poor practice to train pilots to use an input as a measure of aircraft control, where the really important aspect is the output - what the aircraft is doing – what has the control input achieved.

Not what I read. The blocked pitot tube caused all kinds of problems for the autoflight systems. The autopilot's manouvering limits kept the aircraft from stalling while autoflight (A/P and A/T) were engaged. The automation made no change to the thrust setting - that was set by the Captain when he pulled thrust back in response to the warnings. By pulling it back the autopilot could no longer prevent stall, but regardless, the captain pulled back on the yoke repeatedly throughout the follwoing sequence.

Well now, let's break that down a little. Engineers do not design aircraft in a vacuum and expect the pilots to cope with decisions made in isolation. In any development process there is a constant dialogue going on between all interested parties.

I've said this before, but there seems to be a persistent rumour that the Airbus FBW design was the work of engineers and management alone with no pilot input - which is categorically not true.

The difference between the Airbus and Boeing FBW designs is solely down to the dialogue having different conclusions (because the pool of engineers and pool of pilots was different). As Organfreak says, there is insufficient evidence to prove that one design is any safer or more intituitve than another, so all we have to go on are our own conclusions based on the information to hand.

Remember that the yoke design grew from the requirement to have cables connected to all flight surfaces, and until the late '80s was a de facto standard that reached acceptance over time - it was not designed to be the ultimate piloting interface, nor was it ever so.

I never said it wouldn't - I said that there's a roughly equal probability that a connected yoke or stick would have helped versus the probability that it wouldn't. I've only ever taken exception to posts which state that it categorically would have made a difference when the evidence is not there. :)

A german translation of the accident report is available here.

Now watch it, this may backfire.
The crew of AF447 after the loss of control wanted to arrest the descent and climb back to their assigned flight level, therefore they pulled on the stick. The aircraft continued to descent, therefore they pulled more.........

From a pilots view it is most important to evaluate the output in relation to the input. As mentioned before, thats the control strategy of automatic systems as well. Take the input value or the output value away from the logic, and your automatics will go banana.

You need to take both into account, the input and the resulting output.

Spot on, franzl - that was logic gone mad. How doe sit go? "To go up, pull back on the stick. To Go down, pull harder".

Aviation has always been about "is the result what I wanted". Output, as you say, matched to input.

DW - why, in order to link the controls to the surfaces electronically, was the control column removed - so we can pull out a tray and eat our food?

So here we need to dig deeper. The aircraft is descending and pulling back does not help. At some point you'd hope that the problem-solving mindset would say : "We're pulling back and still descending - why is this? We need to try a new approach".

At which point the most common reason for descending while pulling up would have to be "we're stalled". Sadly they never got that far. One of the big questions is *why* they never got that far.

Hi,

DW
You will have the answer in two days
If no answer .. you will ask the same question the next time ... .........

The side stick reached the aft stop at FL363.

I think that in fact we'll have a plethora of possible answers, all of which must be addressed - there simply isn't enough information to determine the HF aspects for certain.

Among other things... ;)

In all seriousness, I've gone into this so many times now that to repeat myself would only be hogging posting space. Here are a few links to more recent posts on the subject:

http://www.pprune.org/tech-log/48235...ml#post7184647

http://www.pprune.org/tech-log/48235...ml#post7172036

http://www.pprune.org/tech-log/48235...ml#post7168364

http://www.pprune.org/tech-log/46625...ml#post6779587

And, possibly to mind mind most importantly, the role of Capt. Gordon Corps:

http://www.pprune.org/tech-log/45465...ml#post6538934

One more time : The fact that a "timely and effective communication with your colleague in the opposite seat" is necessary, makes the point that something is missing.

By their presence, the yokes make such verbal communication unnecessary.
Unambiguous direct and free information to all in the flight deck : Priceless !

Some months ago somebody asked whether there were any training aircraft fitted with S/S.

I read that the specification for training aircraft states, IIRC, that dual controls were required for what I think, was professional instruction. It specifically stated that swing-over controls were not acceptable
At least on a DH86b it meant saying " You have control... " was done without any doubts by either pilot !

CONF iture, yes, the yoke requires no communications of what the other pilot is doing. Us Boeing pilots know when to take over and not talk. Just say I have it.

I don't see this crash happening in a Boeing because of the fact at FL350 you can not let another pilot pull a yoke into his lap with out knowing you will go into a deep stall. Those guys did let it happen. Their inexperience shown by the most experienced guy in the left seat calling the captain instead of taking control. He let the young pilot pull into a full stall and didn't stop it because his experience was slightly more than the PF. They should have not been flying together. The captain would have prevented this tragedy but he was required to trust his co pilots to let him have his required rest. Cruise pilots are not real pilots or they wouldn't be called cruise pilots.

CM2 2936 TT, 807 on 330/340
CM1 6547 TT, 4479 on 330/340
Capt 10988 TT, 1747 on 330/340

When is the pilot experienced enough to be safe?
Why do the certifying authorities allow pilots consider pilots with 200 hrs to be safe riding a RHS in airliner and consider 1500 hrs enough for command - provided pilots are properly trained and checked out?
What do you see that no competent aviation authority can't?
Is there any study showing less experienced pilots to be more incident/accident prone or that high-timers are immune to certain types of mistakes?

What does? Capability to fly aeroplane manually? Would perchance ability to comprehend what is going on from the cues on the instrument panel be more useful to AF447 crew?

Yes. In normal and alternate laws, stick free, FBW Airbi are flightpath stable so they'll try to keep the flightpath no matter whether the disturbance comes from turbulence or from power coupling. Also they move elevators to satisfy G command, if you slam the thrust levers and get classic pitch-power coupling while pulling on the stick, elevators will deflect less to produce commanded G.

Sounds complicated but feels very natural. You set the nose where you want it to be without thinking about normal acceleration and roll rates.

Now there is a very, very good example why flight safety should never, ever be left to commercial enterprises. Some good and valid ideas are mixed with radical departures from what is currently accepted to be true - which wouldn't be bad in itself if it were supported by some verifiable research, instead of I-say-so's and unnecessarily complicated graphs.

If Dedale SAS France has it its way - yes. Hopefully not. IMHO improvement is needed, complete rebuild, not yet.

Monsieur Paries cites 16 events without specifying when they happened, what were the circumstances and whether they pertain to A330/340 at all. So much for the "role of the aircraft"

BEA interim2 lists 36 occurrences of unreliable airspeed on 330/40, 13 of which were available for detailed analysis. It also makes clear there were five cases of descend following the stall warning - in controlled manner, obviously.

So are we to believe BEA in this dilemma because it is official investigating body?

Nope, not just because of that. BEA has provided pretty comprehensive reference about the occurences. Aeroplanes are listed by s/n, which can be used in conjunction with DoF to find out which were the flights affected, making its claims very verifiable. Monsieurs Paries references have, for the time being, been zilch.

You are right, because this Boeing MD-80 was stalled at FL330 and didn't lose a thing, but is 20 FLs such a big difference? Is the type relevant at all? This Tu-154 was lost near FL400, this CRJ 200 was lost as the crew tried to cling at FL410 while being blissfully unaware what "backside of the power curve" means. At Stony Point B727 was lost that never got above FL248.

Monitoring who or what; Dynasty 006 was a Boeing, too.

If you read interim2, you would have known there were 36 recorded cases of Airbuses 330/340 suffering from UAS and going into alternate laws, with crews handling the situation without any damage to airframe or injury to passengers, so your question would be completely superfluous.

Wrong!

People who keep maligning FBW & FADEC when discussing the Habsheim disaster with blatant disregard to facts, don't have anything to do with Michael Asseline's defense and definitively have no his support. Capt Asseline has accepted the BEA findings. He was no stupid yet he has flown aeroplane full of passengers, at an airshow overhead the grass airfield where he has never flown the display before, without checking for obstacles, executing flyby low and slow. In retrospect it was very stupid thing to do. Analysis of the process that led him to belief it is acceptable to perform such a seemingly idiotic thing and warning every pilot about it was supposed to be the main safety lesson of the accident. Diverting the focus from it by blaming the computers, either through ignorance or agenda, is irresponsible at the very least.

Heavier than air, fixed wing aircraft specifically dislike being flown with low energy, colloquially known as: low and slow. It hurts them a lot.

At 2:10:22 CM1 called out "Alternate law, protections (law/low/lo). Anyway, what's the use of alpha prot in cruise? Other crews lost it too and lived to tell the story.

Yes, but alpha prot gives you maximum possible climb performance, there's absolutely no reason to use it when terrain is not a factor. Pilots should be constantly able to tell whether they are flying high or low.

What is so remarkable about AF447 but is mostly glossed over, is that loss of control looked extremely benign on the inside. There were no extreme attitudes, no rapid attitude changes, no high or low G, no spin, nothing. Just unwinding altimeters. How would centrifuge training help pilot feel what AF447 crew felt is way beyond my comprehension, unless we are talking about stopped one.

Actually, with connected controls, who made which input is always matter of conjecture. Of course capt was designated PF, so chances he made the final pull are extremely high. On FBW Airbus, there is no mistaking who did what with the stick.

Here: quick refresher on Birgenair, by the time it departed the envelope, autopilot and autothrottle were off.

Correct. Panic pull has been with us since there were first aeroplanes, it has nothing to do with type, sidesticks did not eliminate it or made it worse, also triggering condition can differ but result is exactly as you have described.

It cannot be cured by flying more raw data manual approaches. To prevent it it is essential to have a pilot that doesn't panic when startled. Is it achievable through training? Damned if I know.

F-16D
Because 60 years ago it was legal to train pilots on aeroplanes with single throw-over yoke, such as early Bonanzas had. Took a while to realize it wasn't the smartest thing to do.

Read more closely sir. I wasn't referring to those who in modern times malign FBW and FADEC as a result of Habsheim. I was referring simply and purely to the idea (contemporary to the late 1980s) advanced by Asseline's defence team that the BEA were obliged to defend Airbus due to the involvement of the French state apparatus. A position which is easily disproved by the facts on the ground.

That some of the posters on here have not let this go and continue in the hope of humiliating Airbus in the manner that Asseline (and later, Norbert Jacquet) were brought down is not merely a matter of conjecture - it is a matter of fact.

Clandestino, I wouold like to meet you some day and see what planet you come from. No Boeing pilot would pull a yoke into his lap like the PF did with the side stick. Good luck with your long story but I don't think you will have any followers.

Bubbers, it has already been proven beyond a shadow of a doubt that at least two of them did.

There is no such thing as "no XXXX pilot would or wouldn't YYYY", only that which various pilots, regardless of type, have done.

It's time to step outside our comfort zone and realise that there's something significantly wrong here.

A-340/330 UAS Events
 

For those who might be wondering about the 36 A-340/330 UAS incidents, here is the listing and details Clandestino refers to...


http://i1166.photobucket.com/albums/.../UASEvents.png

:=So what about the possibility that the PF simply did not process/comprehend that when the A/P went into Alternate Law, that Alpha Protect was no longer active -- and that he would need to use the side stick more like a Boeing yoke than an Airbus joystick? Thus, it wasn't that he was dumb, just that he had a brain freeze in not fully comprehending that his protections were gone (despite this warning being given on the ECAM)....from SRMman

Yep. Absolutely. The first we know the crew recognized Alternate Law is at 2:10:22, seventeen seconds after loss of autopilot! full stop.

Nothing is provided us to counter this conclusion in evidence. A lot of harrumphing, posturing, etc., but no evidence. It is the salient take away from the initial (post startle) CVR data. We cannot be sure the displays were operating, nor can we conclude either pilot sussed the fact until the PNF callout.

By this time, the upset was under way. Clandestino remarks about the benign nature of the upset in his recent rather lengthy post...

The upset began the moment autopilot dropped out, arguably, and for many reasons, not all damning, many exculpatory.....

There is no going back in some chains, many cannot be unwound once begun.

Whether you believe it or don't, the case can be made that PF's stick work was based on him assuming Normal Law. It makes more sense than an endless onslaught of "No pilot in his right mind ....." ad absurdum....

To be clear, know that a loss of autopilot can occur without degrade.....frequently. Loss of autopilot is not some trained in "oh, naturally, we degrade then..."

Clandestino; That's really not what I was getting at. Put another way, if your butt sensors are only used to the usual about 1g, and you rarely pay any attention to the small inflight 'g' changes, then it seems rather obvious to me that neither of the crew noted the lack of 'g' change when applying NU/ND commands. This can only happen when you are on the ground, or as they were - stalled, and the aircraft response coming from just bigger or smaller 'barn door' effects.

The other clue was possibly the 'soft' rolling; certainly something you wouldn't expect with normal airspeed and AoA. :hmm:

I know its easy to sit back and claim they were just a "clueless" crew and the chances of a similar result happening again is probably something approaching infinitesimal odds, but the "why" surrounding this event has to be addressed and corrected. So if that means crews need to experience the full range of acceptable 'g' forces in normal flight and those that could be expected in recovering from LOC, then so be it. ;)

Let's be clear - what accidents like Flash Air have proven is that the physical response from the human body is more often than not wrong. Getting too deep into it is going too far. It doesn't matter what control law the machine was in - the input sequence was inappropriate.

No-one's saying the crew was "clueless", but let's be honest. If you're at full thrust, full back-stick and the aircraft is still falling - what else can it be other than stall?

Dozy;

Well, you just said it.:ok:

Hi,

Machinbird
Interesting to know that Jean Pariès is a ex DGA ex BEA
The loop is closed .. lol
Jean Paries - Lucina

.....a failure in the barostatic system causing not only erroneous speed indications, but also a downwinding altimeter at a never before seen rate (."...that can´t be.....") and a pegged VS?

Not saying that it was like that, just spinning your question a bit