THS
 

OK, I stand corrected that the THS wasn't "stuck up". I recall a number of posts trying to figure out why it was up, and whether it was stuck, broken, etc. But it seems to have been up (for whatever reason) and stayed up. Has the BEA explained why?

RWA-

Amongst all the hot air generated on this topic, I think your analysis may pretty well hit the nail on the head. Unfortunately, a lot of people may not be seeing things in quite the same light. Quite a few posts ago, I made the comment that a very small pull on the stick at high altitude, if you were heavy and close to max altitude, (should'nt be there in the first place, especially if turbulence is present, as the FCTM clearly states), could put you in stick shaker very quickly, together with a zoom up into "coffin corner" territory, (don' t hear much talk about that phrase these days, do we!)

Combine that startle factor (for the unprepared/untrained) with a full up THS and you can see why a certain element of confusion may have arisen. I think we will eventually find there is more to this than simple pilot error. Not bashing Airbus, (they build great aircraft, and I speak from experience of operating the type as well as Boeing), but nevertheless Toulouse will have some explaining to do here.

Regrettably, over the years, Airbus have usually blamed the pilots in most accidents ("they did not understand the systems"), but then they quietly go ahead and change some bit of software/hardware.

The reason is simple. The pilot's nose-down inputs only reduced the elevator deflection a little - the elevator still stayed "up".

He didn't keep the stick forward far enough or long enough to move the elevator back to neutral, and then into the "down" position. The THS would only have started moving "down" again once the elevators were pushed beyond neutral.

The THS behaved exactly as designed.

Not wishing to sound facetious, but would you care to provide examples of where this has been the case any later than, say, 1994 (when the death of senior test pilot Nick Warner on a demonstration flight caused them to re-examine their priorites)?

Neither, importantly, did he exercise the option to re-trim the aircraft manually using the trim wheels, which would have taken significantly less time - presumably because he was not taught how.

Hi,

One from head is Mount St Odile (some changes after .. despite pilots made some mistake about descent rate settings if remember well) about interface

That might be how the system is designed to work but is it how a human "naturally" expects a system to work? Hasn't it been proven over the years that in an emergency we revert to how we expect things to work regardless of how they actually work or how we're trained to use them?

If I turn the thermostat up on my heating system I wouldn't expect the temperature to get hottor and hotter indefinitely.

If I turn the steering wheel of my car I wouldn't expect the turn to get tighter and tighter until the steering rack is on full lock.

Even the accelerator pedal on my car mostly behaves like a speed control rather than an accelerator. Once I have reached the desired speed I don't lift off the pedal to maintain that speed. I have to keep it pressed.

What other familiar systems have a "runaway" behavior like the auto trim?

puzzling...
 

One thing that is puzzling me is the speed of PF's hand on his movements aft and forward on the sidestick.
From 02:10:13 to 02:10:16 (only three seconds) he moved his sidestick 9 times. That's three times a second...
And with significant amplitude also.

What was he chasing?

The answer can be speculated from the BEA report.
When AP disconnected the a/c was pointed down slightly (Pitch=0 deg) and it was descending. Also altimeter might have jumped down a few hundred feet because of impared pitots and air temperature sensors due to a possible ice build up from the strom they were in.
Pilots are trained to maintain altitude in cruise so he correctly pulls the stick back to keep at FL350. It appears he was in a down draft because it took a large stick deflection to maintain altitude for a few seconds. At high altitude and speed his strong stick input has dangerous effects and a stall warning is triggered. He then notices airspeed is broken. We can speculate that he was distracted, took his attention away from altitude and tried to process what was happening. More errors occur further distracting him. Finally the a/c has risen to FL375 before he figures out he has climbed, he had exited the downdraft and had held the nose up attitude.

Climbing the aircraft without adding power results in a drop in speed. Same as a pendulum swing up. Now a/c was in a dangerous high altitude and low speed condition with impared flight controls (ALT LAW).

It didn't runaway. The pilot made such high nose-up demands via the SS for such a long time that to satisfy his demands the system gradually trimmed to full nose-up. If he'd relaxed the back pressure for long enough (ie. taken "standard" stall recovery actions) the trim would've moved back towards the nose down position. It just trimmed to try to satisfy what the pilot what was asking of it - a high nose up attitude.

No, but if you set the thermostat to a temperature that was higher than the heating system could achieve the temperature would rise to that limit. Similarly, the pilot was requesting a nose-up attitude that the aircraft, with it's low speed, couldn't achieve and the trim ended up moving to the nose-up limit.

Agreed. Look at the accelerations too.
Looks to me like he was over sticking then having to correct for each input.
Instead i imagine he should have used smooth gentle deflections.
Ever seen a nervous person drive a car like this? Constantly twitching the wheel having to correct over steer in each direction. A smooth small input is all that is required. I have also seen people use accelerator and brake in the same way, constantly pressing one peddle then the other. Instead of just calmly pressing accelerator to desired speed. You are pulled back and forwards in your seat. Can't say i've seen that behaviour much in pilots.
Totally speculative - is this consistent with anxious/un-trained behaviour from the onset.

In alternate law we are still in load factor demand correct?

Could be a passenger was flying it while waiting for the toilet? Or could be an instrumentation or recording error ;)

Flickering Flight Display
 

xcitation, it looks to me that he is "chasing" a flickering PFD.

It was an attitude indicator, because he is correcting roll also.

But anyway, at least to my knowledge, the normal is to be "quick" on ailerons and "easy" on elevators.

The graphics show the opposite.

According to the docs ALT1 / ALT2 have Load Factor.

That was in 1992 (2 years before the cutoff date I specified), and interestingly was the occasion on which the BEA invited the NTSB over to perform a co-operative investigation in order to prove they weren't trying t oprotect Air Inter or Airbus.

(The fact that Air Inter specifically ordered A320s with GPWS deleted didn't help in that case either)

One thing that is puzzling me is the speed of PF's hand on his movements aft and forward on the sidestick.
From 02:10:13 to 02:10:16 (only three seconds) he moved his sidestick 9 times. That's three times a second...
And with significant amplitude also.

What was he chasing?

__

I've flown with a few FO's that did that all the time hand flying. They have plenty of experience but don't seem to trust themselves hand flying so keep doing PIO's to touchdown. You would think some hands on flying would stop this but a lot of airlines discourage hand flying these days. The PF in the AF447 crash didn't seem to have a clue when the automatics failed. They need to look at the Bob Hoover you tube flick where he pours an iced tea pitcher into a glass while inverted in a roll in a Shrike on page 83 of the tech forum. I know Bob and he knows how to fly an airplane. He will be at Reno at the air races in mid Sept.But he has retired from demo flying.

Dozy
So what are you implying?That there was conspiracy to deceive and obfuscate prior 1994 but everything 's been above board since?

Would you say there are any latent dangers in this autotrim design in abnormal ops or that the pilots always instantly and intuitively recognize the transition from auto to manual?

Would you say that the stall inhibit below 60 is excusable in design terms as its outside the realms of probability?

Would you say that the SS is ideal in a multi crew environment especially when considering scenarios where stick input must be timely and CORRECT(stall/x-wind/sev turb etc)?They have the stick in fighters but thats one man.Why did AIrbus employ it in commercial airliners?To be NEW?You seem to be an avid Airbus apologist so I'd like to know.

Do you have an opinion on why BEA withdrew the stall logic recommendation(factors affecting safety can and must be included in initial/interim reports but maybe theres a valid reason)?

Do you have an opinion on why no mention was made of the Captain's decision to leave the cockpit even though he now appears to have known that weather was up ahead?

Hi Rananim, long time no see... :)

Not at all. They were overconfident in the abilites of their systems prior to 1994, that's for certain - and as such predisposed to put things down to pilot error when interface issues should have been looked at (the dual-mode Vertical Speed and Flight Path Angle in the Honeywell FMC being a major case in point, but the same computers were used in other manufacturer's aircraft). Airbus weren't alone in trying to point the finger at pilots in the '90s - Aircraft manufacturers are billion-dollar corporations open to massive liabilities claims and as such their legal departments will tend to demand it - look at how Boeing tried to pass the 737 rudder PCU issue off as pilot error until the NTSB had the "eureka" moment and tried a thermal shock test.

My honest answer is I don't know, but the fact is that the inputs made to the system were way in excess of any input that should have been applied at cruise level. As (I think) PJ2 says, manually controlling the Airbus via sidestick at altitude requires light thumb pressure in pitch, and gentle lateral movements. This guy wasn't taught how to handle the thing, so he's bashing it on the stops.

I wouldn't say anything other than that design tends to happen at the limits of what is known. I definitely think the logic needs looking at, but I find it difficult to be too hard on the aircraft and systems designers for not contemplating what would happen if the aircraft lost all speed data and ended up in a mushy stall due to overcontrolling. For all we know the stall warning on other modern airliners has similar inhibit modes, and I'm sure Airbus will share whatever information they have in that regard.

For a start, I am an apologist for no-one, see my earlier post for further details there (should only be a page or two back [EDIT - http://www.pprune.org/rumours-news/4...ml#post6622053]). Secondly, if you're going to start an airliner design from a fresh sheet, why not? Yokes can be a pain in the ar*e, and have noticeably been used as a crutch in certain situations (some of which happened after the A320 was well off the drawing board) when they had been effectively rendered completely useless. Sidesticks are smaller, lighter, de-clutter the flight-deck, allow more freedom of movement and the lack of backdrive makes the system less complex (technophobes don't even want to *think* about the numbers of lines of code required to make the B777 yoke behave like the old-school ones). It would also appear that the majority of people on here who have an issue with the sidestick have never used one. Neil Armstrong certainly didn't have a problem with them...

Presumably to clarify what the situation regarding stall warning actually is. This is an interim report after all, so let's wait and see what the final report says. I have a strong suspicion that stall warning logic is a complex beast that's going to need to be looked into by all manufacturers. Stall warnings and pilot's responses to them have been a prickly issue since at least 1972 and the BEA548 Trident accident.

Not at all, other than from a layman's perspective (in airline pilot terms) it appears that AF has a few CRM issues it needs to address if it hasn't already done so, but primarily, I believe that placing two F/Os in the flight deck - in the ITCZ at night notwithstanding - when neither of them has had any high-altitude manual flight training is frankly inexcusable.

From the BEA site (french)
communiqué 3 août 2011

Google site translation:
Google Vertaling

Transcrypt of press meeting is also now available (english)
Transcript de la conférence de presse du 29 juillet 2011

If the PNF had a yoke in front of him flailing all over the place mostly in his lap don't you think he would have responded and put it where it was required to be for the no airspeed situation? I know the yoke costs money and weight but isn't it a wonderful way to see what your fellow pilot is doing? Maybe Airbus should put yokes back in their aircraft.

What the lawyers lawsuits are going to cost them negates all the years of SS controls to reduce weight and cost.

This could happen again.

Recovering a Cessna 152 from a students stall is easy with dual yokes. It would be more problematic with a side stick of Airbus design.

That's why training aircraft don't have non-interconnected sidesticks. :)

One would hope that by the time someone has ended up in the flight deck of an airliner that they would have the requisite training for controlling the aircraft at that level (in terms of ability and altitude!). Airbus certainly haven't condoned pilots being certified without that training, but it would appear Air France let it happen anyway. Why the desire to crucify Airbus, when Air France seem to be the problem here?

Airbus's ace-up-the-sleeve has always been the ease of type conversion between their narrowbody short-haul types and their widebody long-haul types - weight saving was simply an added bonus. The advantages of starting from a clean sheet in that manner seemed at the time to have provoked a severe case of sour grapes from the other side of the Atlantic, which at one point considered the monopoly of western airliner design it's god-given right. The early years of competition between the two were bitter, but it seems to have resolved into a more friendly rivalry these days - unfortunately some seem to not to want to let it go.

The comment about no high altitude training hand flying being inexcusable is naive. It was the norm when we came to the airlines. We were expected to know it. We were dispatched at high altitude with no autopilot with FAA blessings. The problem with inexperienced pilots going through airline training programs is they don't have the money or time to spoon feed new hires. They start their job barely qualified in normal conditions. You can't expect much performance in the AF 447 crash and you didn't. That is the new culture.

I have to enter at least 10 characters... But this was all I wanted to say :ok:

Then would you not agree that the culture needs changing?

Look, I may for various reasons have abandoned any attempt to fly professionally early in life - but I'd have thought that if it's in your blood you want to do it to the best of your ability. To reduce it to "just another job" is ridiculous. The airlines expect you to handle these things manually at cruise altitude without training when things go pear shaped? I suspect there'd be an uproar of sizeable proportions if it ever got out that that was the norm...

Sidestick
 

Hi Bubbers44,
After the plastic vertical stabilizer snapping off, the crazy computer automatic dive, the flat spin caused by too far aft CG, the total loss of elevator control due to THS autotrimming, the complete loss of artificial horizons (meaning a quadruple failure of inertial systems), the supposed lack of alpha indicators (which exist but that no company really care to add on pannels) or control surface feedback, etc. theories... we are back, now to the so flawed sidesticks that can not allow any cross control check of imputs...

Actually, this feature exists on all Airbus aircraft, but it's not displayed after take-off. Then, one should ask Why it is like that?
In fact, because nobody using those aircraft seems to be very concerned about such an "obvious" need or it would have been made possible from two decades. Nonetheless, it could be very easy (and cheap) to add it, without changing anything to the actual stick system, via a stick pushbutton commanding to display this function (see below) during whatever phase of the flight.

Going on lawsuits against Airbus on such a case would be a non-winning game. Nothing is showing that it was a factor at all, neither that doing otherwise is better. It is obvious that the PNF complained about PF initial imputs almost immediately: he saw them and was able to take over at any point (including by blocking PF further actions) but he refrained from doing so.

Function displaying sidestick imputs:
http://takata1940.free.fr/sidestick.jpg

SS vs Yoke
 

When they started to design the 777 (and that's a long time ago) Boeing consulted pilots from around the world. Their preference was clearly in favour of the sidestick.

Not wishing to be seen to follow Airbus, Boeing disregarded this opinion.

Quoting Sidestick:-

Sidestick, I don't know everything about the Airbus displays, but as I understand it the PNF would normally have been able to monitor the control movements of the PF (in the captain's seat) from the Flight Director; but that this depends on the availability of some 'crosshairs'? If you refer to the FDR/CVR transcript (Page 86 onwards) you'll find that the FDs were cutting in and out at intervals throughout the descent; and in the body of the report the BEA says specifically that the crosshairs were cutting in and out even when the FDs were working in other respects? So I for one am none too sure that the PNF had access to the information he needed?

Oddly enough, the BEA has 'covered itself' (in what is only yet another 'interim' report) on this aspect; saying that finding out what information was actually available to the pilots at any particular time will need 'ongoing analyses':-

They also put in some stuff that just about sums up what I for one currently suspect turned an upset into a crash. We all know that such accidents almost never have a single cause: and also that the thing that started off the whole event was the sub-standard pitot tubes (which could and should, IMO, have been replaced months previously). My own view, on present evidence, was that the next most important cause was the inadequate stall warnings; that they tend to communicate 'approach to a stall' and don't 'change their tune' if the aeroplane HAS in fact stalled. And also 'shut up' if the ASI drops below a given speed. Not getting at Airbus specifically on this, I wouldn't be at all surprised if Boeing's stall warnings work the same way.

RWA - I think we are pretty well convinced by the AB 'knowers' that there is no way PNF would know what PF was doing with the SS when airborne except by a/c manoeuvre. I also believe that the failure of the stall warning system to perform adequately, while a hugely significant factor, is overshadowed by the initial and so far unexplained and 'un-noticed' pull-up to above safe altitude. Once there the dice was indeed heavily loaded.

If what Bubbers says about 3 pitch/roll SS movements a SECOND..................!?? - that is either a fault in the system or one of the most bizarre episodes I have seen of 'stick thrashing' and AF need urgently to review the 'pilots' in their fleets. Was he simply 'chasing' a wild FD?

No.

takata, I would be curious to know if you are actually one or more of the following :

1. FBW Airbus pilot
2. Multi crew pilot
3. pilot

My reading is that the stall warning system performed exactly as designed and accepted by FAA, EASA and others.

Performed Exactly as Designed
 

I've seen that so many times on these pages in defense of AB, in effort to end a line of thinking. How about considering the design is just plain defective?

How much of the PNF attention was on the ECAM and not on the PF joystick actions? In a yoke airplane, PNF would have been aware of PF actions at all times. Same goes for stationary throttles.

FAA data cited in Mitchell and Revans (2004) AvSpace&EnvMedicine

Mitchell cites the risk of pilot error inducing fatal accidents commercial RTP flight by age is lowest age 50-54 years old.
Taken as an index of one, the relative risk of a 40 year old commercial pilot is 1.25, and a multiplication of the risk by 2.5 for pilots under 30. This explains the intuitively correct CRM call by Sullenberger for "my aircraft".
In a situation loaded with risk for aircraft upset, the most experienced pilot would be , according to the accident figures, by far the superior. Sorry all you young atpls...you'll be closer to magic in your late 40' s.Accident rates for pilots at retirement age are also significantly lower than the 30's and under.Interestingly,if one includes back into the stats all those older commercial pilots who have lost their medical clearance due to risk of sudden cardiac death, the statistical overall rate of pilot error accident risks improves dramatically

Froom et al (1988) describe an ever greater jump in risk (by a factor of ten) between 30 year old pilots and their 10 year younger colleagues in the miitary aviation setting.
The situation also reminds me of the now famous "Rumsfeldts", where there is alot of unknown knowns, and unknown unknowns ...the worst place to be in any aircraft.

The AF 447 situation strikes me as one in which the incipient risks required the most experienced hand flyer right from the word go...no time for liberte,egalite,fraternite.....

if you get my drift.

@ CONF iture
Well who cares what I'm doing for a leaving. I know that the purpose of such function is not aimed at controling PF imputs. But, as both stick imputs are already linked to PFDs, it would be no big deal, if such was really asked and needed by customers, to modify it for cross checking pilot imputs (and I seem to remember that we already had such a discussion about two years ago).
Even adding some feedback thru electrical commands would not add tons of hardware (weight) nor tremendeous development costs to the manufacturer. The very same could be say about "non-moving" lever thrust, silent autotrim, etc.

On the other hand, you just hate this brand, whatever their reasons being to do it like that. I can't remember a single feature you actually liked about an aircraft you are supposed to fly. On my side, I'd like to know why is your grief so deep? Personnal vendetta?

Hi Greybeard,
It seems that there was not a single ECAM TB sequence popping up before 2 minutes after UAS, hence 1 minute+ into the full stall developement (0212+); the only one showing up, quite late, was ADR DISAGREE; there is barely no call about a procedure performed by him, either...

re #2697
 

Hi takata,
no big deal but just to make it clear:

The post you are referring to was not issued by myself.

Thank you for your valuable infos on the VSI!! I checked 34-14-00ff but somehow missed this.

AMEN to ALL of that, Brother. But, you'll NEVER get the AB supporters to agree, unfortunately.

Quoting bubbers44:-

I think this may well be a matter of 'super-sensitive' recording, rather than anything more sinister. If you ignore the 'snaking lines,' and just count the basic 'ups and downs,' rather than the 'zigzags,' there appear to have been only about four of each in the period referred to. And we have to remember that the PF had just taken control after the autos signed off - according to the Report, his main concern at that time was to correct a tendency to roll. He'd have been mainly concerned with moving the stick sideways, not forward or back; if he'd moved it even slightly forward or back during the sideways movements (and assuming, as one has to, that the systems were probably recording in fractions of a second) it was probably 'no big deal.'

Looking again at Page 111, though, I noticed another thing. Assuming that I've interpreted the graphs correctly, while the THS was still moving towards 'full-up,' the PF applied full nosedown for quite a long period; probably because, as the report says, he was (correctly) countering the zoom climb. But the THS just continued on it's 'merry' way to full up.

What's more, if you look lower down, the Angle of Attack graph starts increasing at exactly the same time that the THS starts moving, at what looks like an exactly-equal angle (though opposite on the graph). And reaches its highest angle just as the THS arrives at 'full up'?

So (assuming I've read the graphs right) why didn't the THS react to the PF's apparently 'solid' nosedown input and 'get sensible'?

Hello Mimpe
Your numbers are very interesting.
But is it really a question of experience or is there more about this?
If its is only experience than this should be a constant about the history of aviation.

If its not a constant than there might be other factors like for example changed traning involved.

The "Airplane with HAL" :

Number of A330s built : 802
Number of A330 hull losses : 6

Percentage lost : 0.75%

The "Old-fashioned pilot's aircraft" :

Number of DC-10s built : 386
Number of DC-10 hull losses : 31

Percentage lost : 8%

You were saying?

How about a boeing?
 

Do most of you agree that a boeing aircraft faced with the same conditions and failure would have clearly handled the problem and not ended up in the drink?