Dozy, I don't think that is correct. Letting go would cause the aircraft to hold an attitude as long as it was above stall speed. Below stall speed it will bobble its nose up and down. It needs nose down input on the controls for a significant period of time or actually running the trim down manually to have a hope to recover.
The elevators have some influence, but the THS and the pitch up moment from the engines' thrust hold the trump cards.

The PF didn't actively hold the nose on the level flight pitch attitude when he took control and all bad things flowed from that.

When the nose tried to fall through at the stall, the PF actively fought it and that is the reason for most of his nose up input after the stall. Do you agree?

Nothing of the sort - Capt. Burkill himself says that all he did was raise the flaps while his F/O tried to keep the aircraft as on track as he could - no "oversteering", no disabling of the protections.

The Air Transat Azores incident proved that you can deadstick an A330 quite handily too.

@Machinbird, agreed - to be clear the nose coming down "to some degree" means exactly what you're describing - not enough to change the situation in itself, but the FCU would not be trying to hold the elevators in place if the PF let go.

Maintain altitude, full (or 'enough') power, indicated at approach to STALL.
Use back stick to maintain level.

TOGA, PITCH 15 degrees NOSE UP, indicated at Wind Shear.

Full back stick, full power, CFIT recovery.


All published, all trained, all wrong. Everything is in the "Stick". And the stick is 'invisible'. Without full disclosure, this gets stuck on myth, and bias. Or on a guess at the Pilot's thinking. Excellent way to improve safety, eh?

Let's have a look at the DFDR traces:

Dozy,
The PF released (sort of) back-pressure on the sidestick between 02:12:15 and 02:12:30, without any change in the elevator (full nose-up). When he then moved the SS forward, the elevator moved away from the up stop, pitch attude decreased, and lift (normal acceleration) increased. Similar but slightly lesser responses between 02:13:45 and 02:14:00.

Machinbird,
The traces show similar responses in pitch when thrust is reduced between 02:11:45 and 02:12:15.

Lyman,

I am constantly amazed at what I perceive to be total fabrication and mis-statement. Their is absolutely no indication that the PF acted in any logical or correct manner. He was totally incorrect in his actions and the PM knew it but for some reason did not take control of the plane.

As an ancient engineer, and one not fully familiar with all the Airbus FCS subtleties, I hesitate to bring this up once again....

But, wasn't there a suggestion, early in one of the earliest threads, that AB pilots are 'not encouraged' to use manual trim (except after full reversion to "mechanical")?

Not to mention the absence of a "bicycle bell" while autotrim was winding the THS to full nose-up, contributing to a lack of "configuration awareness"?

Is "trim" still being taught, or is it now left to George entirely?

Maybe I'm missing something here, but I fail to understand how none of the pilots recognised (or mentioned anyway), that the plane was stalling. As far as I know the vertical speed/altimeters and attitude indicators were functioning correctly. Once the plane was in a rapid descent despite the fact the nose was pitched up, I'm sure most trainee pilots or even aviation enthusiasts could tell you the aircraft had stalled.

Basic training used to emphasise the importance of not only recovering from a stall but also to recognise it. As MU99 says, slow speed, high nose att, RoD all give it away. I feel that these guys were disorientated, they had forgotten their basic flying training and were engulfed by the wrong mental model. As a general principle, if an aeroplane does not respond as you expect, reduce the AoA and try again. That's good for anything that you put in the air, fast or slow.

Emergency Action
 

A close reading of the thousands of posts shows that, apparently, there would have been little or no sensation of falling. Having said that, had I been on board, I would have "stormed the cockpit." (Yeah, right, buddy!)

It's clear to me, after having read everyone's input here, that there is a fundamental difference between 'A and B,' and that difference amounts to: to fly B, you need to be a thoroughly-trained pro. To fly A, you don't need to be unless the poop hits the fan. And so, they all died. :sad: So far, without complete information, there will be plenty of blame to go all around.

He blips the controls forward from full back-stick to half-forward for a second, if that - there's a corresponding blip in elevator movement (which will take a second or two to respond in any case), before he returns to ordering half-nose-up pitch, which pushes the elevator back to max nose-up position

Yes, and if you follow the overall trend during those periods, the stick spends more time around neutral than it does during the 2:12:15 to 2:12:30 period (of which approximately 1s is spent with the stick forward of neutral, the other 14s is spent with it back). I don't need to tell you that hydraulically-assisted controls need a few seconds to respond to demand now, do I? :)

@CJ - As I recall, it was only 1 poster, Svarin IIRC, who stated that his airline were very strict about their Airbus crews not touching the manual trim wheel. This is worrying to me because even if autotrim is there more than 99% of the time, you still need to know how to recover the aircraft if it isn't - so it should be trained for regardless. The trim "whooler" was deleted from the A320 onwards, probably because it would be an annoyance with autotrim active during turbulence - it would be sounding almost constantly. Maybe a compromise with the whooler sounding if the trim goes beyond a certain limit could be looked into?

I was referring to aviation accident investigation terminology, in which "structural failure" and "adverse meteorological conditions" are considered distinct and separate.

Knowing what we now know after the Hudson NTSB report, there is a good chance a 330 would have hit the road, fence, loc antenna for refusing to the pilot the possibility to reach Alpha Max.

• It is not trained or 'encouraged' to use manual trim in Normal or Alternate Law.
• Manual trim is used in Direct Law, and trained for.
• Manual trim and thrust are the only way to 'control' pitch in Mechanical Backup, and 'trained' for.

That's your opinion, but BA038 was never anywhere near alpha max after the FMS was taken offline and manual control established - in fact it should not have been anywhere near Alpha Max even under FMS control (although the FMS did order a relatively alarming pitch-up initially) because of the FMS's own limitations.

I'm still not convinced by your arguments on the Hudson incident either - but I'm not going to argue the point.

Self Loading Freight in Arizona.

You have some ironclad conclusions. You are without doubt basing these on what you read here. What the Pilot did, and did not do, was undoubtedly based on what he made of his situation. It turned out badly.

For you to extrapolate beyond that makes you one who parrots what is readily available here, in great quantity.

Oddly, someone as yourself might easily take the opposite tack. I don't presume to know how what happened, happened. I think you do. Is there an explanation for what the PF did? Without doubt. I am baffled. As you are not, I defer to your report of the accident.

What you see as fabrication I present as "what ifs". If you do not suss the difference, no harm, no foul.

Don't choose fish.

What some here do not get, is that this a/c is a known quantity, and has a diary that shows what she did and did not do. The pilots are human, and have to pay for their individuality in knee jerk condemnation because the BUS has a pedigree, a huge fan base, and has her own recorders to cement her reputation. The interpreters of said recorders are her friends.
The pilots are "unquantifiable", to some extent, and that looms large when one's jury is engineers who reject a challenge to their sense of the World as it "is". What bonus, advantage, or future consideration is worth the open mind? None, as they see it.

Human factors are not the same, and instantly put dead men on the defensive.

I would not expect the delay to be as much as 'a few seconds'. In fact it isn't - if you look at pitch response to SS at 02:10:05. The point is that I suspect that after having held the SS fully back for a considerable time, you probably have to push it forward of neutral to affect the elevator. In consideration of the control law in pitch, just releasing the stick does not demand a reduction in pitch or load factor.

@HN39 - It's going to take some time to respond from a 100% pitch-up to 50% pitch-down, as was the case during that trace.

Lyman
 

Help me out here, guys. Is it me, or are Lyman's posts increasingly odd?

It doesn't matter a damn what the aeroplane is. How crude or sophisticated. You do NOT pull the stick hard back when the autopilot disconnects at altitude. You fly attitude, attitude, and attitude. Then, you fly attitude.

@Dozy,
I was talking about the elevator moving from 30 deg ANU to 20 deg ANU.

Aileron Drag
AD, no, you are quite on target. His approach is often skewed from physical reality.
But sometimes he has his points. Just pick and choose what you want to actually consider and ignore the rest.

@HN39 - I must confess I'm a little confused. At approx 2:10:07, the pitch attitude trace registers nose-up in response to the initial pitch-up command from the sidestick. The aircraft is at normal cruising speed at this point, so the elevator deflection only needs to move very slightly to effect a significant result.

I thought we were taking about the later traces at 02:12:15 to 02:12:30, when the aircraft is stalled and the PF makes a 1-2s "blip" forward on the sidestick, whch triggers a very small deviation in the elevator trace, followed by the 2:12:31 to 2:12:45 traces where the stick spends most of it's time either neutral or forward, with the corresponding elevator movement, and the 02:13:45 to 02:14:00, where the PNF's inputs achieve the same thing.

Lyman . . .
 

Are the FDR recorded nose-up inputs by the P/F's stick in doubt?

...FDR at 2h 10min 51 "the P/F continued to make nose-up inputs." What difference would it make if he was commanding the climb or trimming into it ....when he should have been commanding nose-down and if trimming at all, then nose-down.

Tell us: With A/P switched off, could the A330-200 climb and trim itself into a 3000' climb without manual inputs?

Quote:
A true aviator flying a normal aircraft uses touch and sight.
Oh, I didn't know. So in your opinion all fighter pilots nowadays are no true aviators anymore... - gives me comfort when I share my cockpit with them (I can promise you, they still are). I never understood the "tactile feedback concept" on modern airliners since you are not getting feedback at all but working against springs and hydraulic units. Let alone these clumsy scratching autothrottles that are never completly aligned and where you spend most of your brain capacity to figure out how to put them in the position you want them (and they move again away from there).

No let me tell you, a true aviator is the one that understands that every aircraft is a true aircraft, that you have to handle her gently, moderatly, sensibly, intelligently and foresightedly. You can bring down every aircraft if you want to, be it an A300 American Airlines over NY 2001 with weired rudder input, or forgetting moving throttles like Turkish in Amsterdam. It's not easy, but you can do it. If you are no true aviator. They are spread over A and B about evenly...
Last edited by Dani; 19th Oct 2011 at 06:40.


I can't comment on fighter pilots.

I learned to fly without reference to flight instruments, to judge my attitude by outside references, to judge speed by wind noise and control feedback, and I trust that I am not the only Airbus pilot with that background.

You seem to be focused on proving that the pilots performed poorly and proving that the airplane is innocent, while I'm trying to understand WHY they failed at their task. I'm not anti Airbus. I guess you'll just have to disagree.

isn't it ironic that the I, the American, find myself defending the French pilots against their European brothers.

To: everyone else, Please remember that we do NOT know what the PF saw in his PFD. All of the indignation you read regarding the pilot flyings failure to properly react is coming from people with 20/20 hindsight. People who have had two years to study, in minute detail, every parameter of the initial onset and subsequent stall. Had the pilot had the same verifiable , reliable data we have the outcome would likely have been quite different.

TTex - you are a welcome arrival in this melee.

Pitch and power. (There were some intriguing comments some months back on how pilots are trained/incentivized to never approach "overspeed" but as we are not mind readers, we aren't sure just what the PF was thinking)

If you go back to the first two weeks of comment on this mishap, as the ACARS data became available, a lot of Professional pilots opined the following:

You fly pitch and power. That would have prevented this mishap, flying pitch and power for the altitude/temp/condition one was in just before the airspeeds went AWOL.

Now that the plane has been found, and some of the info of what went on is coming to light, that simple point remains true.

I asked over a year ago what TTex is asking now:

What was the PF seeing? We don't know, due to how FDR works.

Many have inferred that his artificial horizon/flight instruments, were, absent the airspeed tapes, working as they should. The BEA does not seem to have uncovered any info that leads them to believe otherwise. IF they did, I think that would have been released.

What is disturbing is that what appears to have been a scan breakdown by one pilot was not caught and corrected by the other pilot on hand in the cockpit.

But flying pitch and power are hard to argue against as the ways to avoid such mishaps as AF 447, in an A330, in the future.

Indeed, in most aircraft. :cool: That this approach was not taken is cause for serious concern among all SLF.

How common is the potential for this mistake?

Unknown.

Not at all, unless you're choosing to read it that way. You can understand a certain defensiveness on the part of some Airbus pilots however, given the amount of unadulterated horlicks that has been thrown at the type over the years. Also bearing in mind that some of those you are talking to have English as a second language, saying things like "A true aviator flying a normal aircraft uses touch and sight" could be read as accusing them of not being "true" aviators, whatever that means.

As are we all, but the fact is that there are many on the myriad threads on the subject who will not be satisfied until the aircraft is found 100% to blame for various reasons related to their personal psychology.

Nationality does not come into it - more so than ever these days. The facts on the ground thus far indicate that the PF made inputs that make no sense coming from a trained aviator - finding out why this was is absolutely paramount, but throwing pre-conceived notions like the Airbus being overly complex to a pilot, rogue computers going haywire and the like around do not assist understanding, they simply turn things into a slanging match which benefits nobody.

You're new to the subject and threads, but I'd like to point out that one of the very first things I said on the subject was to the effect of the situation those pilots faced was the dictionary definition of every pilot's worst nightmare - unreliable instruments in night IMC in turbulent conditions, and that no matter what happened on that flight deck, this must be taken into account. There's no evidence to suggest that the PF's PFD was displaying anything divergent from his colleague's in the LHS - and indeed, if the PFD was not making sense (over and above the unreliable airspeed and V/S unreadability) then there is no apparent indication of that, nor an attempt to cross check with the PNF and hand control over if the latter's was making more sense.

This isn't about two sides "defending" their turf, this is about trying to get to the bottom of the accident and doing our best to not fall into the old traps which infect every Airbus incident thread there has ever been on this site.

Yes, I thought so too, and the next 15 secs, basically.

So I don't see where we're diverging - stick goes forward and hovers around neutral, elevators go down, stick goes back (trending around 50% pitch up again), elevators come back up - and then it repeats (which in itself is interesting because it shows elevator response to the dual inputs at that time). Looks sensible to me, and I don't know why 2:10:05 comes into it.

Sir,

The Airbus is flown using purely on the basis of flight path. The pilot sets the path and it is coordinated in the short term by the elevators and in the long term by the THS. So "trim" is not used/mentioned when the flight path is selected (normal and alternate law). Trim is used in Direct law, but normally in the sim you deliberately configure for landing in alternate law, drop the gear and end up in direct law, and you remain in trim for landing.

Naturally the pilot is more reliant on his ASI than in a conventional aircraft, since he has lost the "out of trim" feedback. However, you have all the protections.... Young first officers I fly with are astonished when I bore them with stories of how you can fly from clean to fully configured quite accurately with all your instruments covered (in an old fashioned jet).

Now...

In the usual sim scenarios something goes wrong and you end up in alternate law. So, as Han said to Chewy, you "fly casual". No high speed, nothing dramatic, and when you drop the gear you often end up in direct law. No drama (and really, not much learnt).

In alternate law, you have no protections, but you still have a stall warning. It is not often drilled into pilots that this stall warning is really important when in alternate law. We rarely, no maybe never, do scenarios where we are in alternate law with UAS and we might easily stall. All UAS has been done low-level, where the recovery is simple and brutal, using escape-mode pitch and power. High altitude flight is hardly ever trained - and most new zero-to-Airbus guys have never hand flown a conventional aircraft at high altitude. The Airbus hides the delicacy and sensitivity you need, and masks the true lack of grunt available from the engines.

With only sidestick and seneca in your experience bag, you are not well positioned to start your manual high altitude conventional flight learning with UAS in a thunderstorm. Especially when all the training for UAS has been at low altitude. I can't reproduce the words of my certain trainer exactly here, but it was along the lines of "if you hear the stall warning in an airbus, then something has gone wrong and you have stalled, so you need to unstall and then work out how you :mad:ed up twice in such a short period".

Furthermore, when something goes wrong there's normally a drill or "recovery" to be done. Especially if you are an FO. Captains are required to sit and think, but FOs are required to recall and perform the drill. Of course here the drill was to do as much nothing as possible, but that doesn't feature in training either.

It is easy to end the thought process with the fact that the 2 FOs got it very wrong. It is clear that they did, but I would suggest that there is a whole generation of new pilots capable of making the same error. Through lack of experience of less able aircraft and through lack of conventional high altitude flight and failures.

The Airbus is an extremely capable and safe aircraft. By far the safest type I have flown. But it will still kill you if you get it badly wrong, just like any other type. If airlines insist of shoving low hour pilots onto the Airbus, then they need to seriously up the ante with training and standards or else this type of accident will happen again.

TTex600

Welcome indeed. Hope you stick around.

When a thread is paralyzed with rhetorical questions that parrot all the prior rhetorical questions, and there is so much arrogant condemnation, a little bit of leadership is well appreciated. Here be Hyenas on a kill.

Phew.


HundredPercentPlease, likewise.

100%Please, very good post, just a small hint: A330/A340 don't go into direct law when gear is down. That's a feature of single aisle Airbus.

discussion about elevator position, SS input
 

@DW, HN, and MB,

we had been through this discussions many moons ago.
In Alt2 we have the NZ Law without protections.

The pull on the stick sends a loadfactor demand to the FCPC, which in turn position the elevators and in the long term the THS. There is no direct conection SS to elevator.

When the PF pulled back on the stick for some time, he commanded a loadfactor increase, which the aircraft could not achieve neither by elevators nor by THS trim after 02:11:30. Therefore position of elevators are full NU and the THS was NU as well. Relaxing the SS to neutral would reduce the loadfactor to 1 G, which the aircraft could not achieve either in the stalled state, therefore no change in elevator position or THS trim, the FCPC´s still try to achieve the demanded loadfactor of 1 G (stick neutral) which would be a solid climb with the 10° pitch at release of the SS.

Only prolonged SS ND command would have started to move the elevators from Full NU to ND position.

Especially DW´s inputs are wrong and misleading concerning the behaviour of NZ law and giving the impression, that the elevator position can be correlated with the SS position, like releasing SS to neutral would have neutralized the elevators. That would be the case in an unstalled flight condition, where you point the nose with SS, then neutralize and the aircraft continous with one g flight on the new ordered trajectory.

But in the stalled condition AF 447 was in, the actual flight trajectory was 45° down and the ordered one was some 10° climb, which neither elevators nor THS could give, but the FCPC´s tried to achieve by elevator and THS full NU.

Obviously the crew did not understand the situation and the law and the consequences of that situation like some posters here. SS input from full NU to some short ND didn´t produce any noticeable response of the aircraft. Those pitch responses present had been mainly caused by thrust changes and not by SS inputs. Way back there was a graph with SS pitch input, elevator position and thrust setting, that one told it all.

The elevator never came in the ND region after the initial stall, the best it ever got was 15° NU command at 02:12:45 and 02:13:50 (page 108 of BEA interim report No. 3) when just prior the thrust was reduced and the SS was more ND then NU.

franzl

Franzl - it looks to me (quickly doing the maths in my head) like the elevator trace looks like it takes THS position into account and is relative to the THS value.

The FCPC will command what it is told to, this is true - but letting go of the stick should at least allow the aircraft to follow natural trajectory - it wouldn't be enough to get the nose down, and I'm not arguing that it would, but the elevators were being held in position by the inputs and came back down when the stick was held forward for a few seconds and then held around neutral. The question of whether the underlying systems were in G-loading mode or pitch command mode (which IIRC becomes active below a certain speed) is important when discussing this, because if it was holding G then the FCPC might have held the elevators there and if it was in pitch command mode the elevators should have relaxed.

Either way this doesn't help understanding to a great degree, because at the point the elevators returned from full-up, the aircraft was beyond saving. Why did it take so long?

It may look like that, but it is not relative to it. If the loadfactor demand would have been achievable, let´s say with a THS trim of 10° NU, then the elevators would be back in neutral after SS release. The elevators are used by the FCPC to change the loadfactor demand (short term), when achieved and stick is back in neutral the THS trim (long term) will trim out the load of the elevators , bringing those back to neutral.

That is wrong again, the natural trajectory of the stalled AF447 was 45° down, but the intended trajectory commanded to the FCPC´s was 10° climb, and thats the only order the FCPC´s will try to follow. Releasing the stick to neutral does reduce the load factor demand to 1 g, which in turn does not lead to follow natural trajectory but to an order maintaining a ordered trajectory with one g.
The trajectory you describe would be less than one g.


see above

Elevators were being held by the FCPC´s in order to follow the loadfactor demand, which could be changed only by SS ND.
Elevators came out of the full NU position to 15° NU, but never ND position, because the FCPC+s where not ordered to do so.


This pitch commanding mode was mentioned several times, however no reference to it surfaced until now. Would be interesting to know, when does it self-employ and how would it work.

That is the one million dollar question, some answer with adressing poor airmanship and some try to look deeper to prevent it from happening in the future.

First, I am NOT a FBW pilot, and my experience is limited to DC9's and more lately to non FBW corporate types, (mostly glass). Please feel free to shoot me down as one of the non informed on this thread, but I have often wondered about the following;

Is it possible in the A340, or any FBW type, for the crew to revert, at any time of their choosing, to conventional control of the aircraft? By that I mean, can they take over manual control (albeit by wires) of pitch, roll, yaw and thrust whenever they wish, and without any interference from computers?

On another more simplistic note, why would the computer logic of the A340 CANCEL a stall warning when the AoA had exceeded a certain value above that of the trigger point? The type I fly commences stick shake at a given value, and is then followed by stick push. NEITHER are cancelled until the AoA is reduced to less than a level that triggered the appropriate warning. Pray, what is the logic in this, when on the other hand a similar (albeit opposing) warning, such as overspeed, is permitted to continue forever?

I'm not so sure, not least because the traces indicate that the sidestick was *never* released - from the moment FMC disengaged until impact. The last two forward stick movements are held for several seconds, but the trace never goes above -20 degrees - it can't take tha long to respond to a full pitch-down command, surely?


Personally I reckon we should both hold our horses until this is clarified. :)

What bothers me the most (and this is not directed at the crew, it is directed at the industry that trained and passed them, as are all my airmanship questions on this matter) is that I can't understand how a rapidly unwinding altimeter despite considerable nose-up pitch and TOGA thrust could be perceived as anything *other* than a stall- and that's before we get to the 50-odd seconds of stall warning that sounded!

@deefer dog - That's precisely what Alternate and Direct Law are - full authority handed to the pilots, transmitted through the digital control system, no hard limits or protections - Alt 1 and 2 have "soft" (i.e. overridable through pilot input) protections except in the case of ADR failure. This is, however, determined by the computers rather than manual input (however, the computers know when they can no longer cope and make it very clear to the crew via ECAM message that the control law is degraded).

In the beginning, the STALL WARN was WARN, not a STALL. There were no airframe cues for STALL, Instruments came later on the screen. There is no Alarm for STALL. ........(Differentiated from WARN) AND Airstream noise is atypical at STALL. (It signifies "crazy speed", at least to the pilot). And the variety of crazy speed he had was not taught, he had never experienced it, nor entertained it in a bad dream.

One needs those cues, AND instruments. Especially when key instruments have gone South.

RetiredF4:
How do you translate "10° climb" in Nz demand? After all, once stalled, the g were around 1...
The law is not a "pure" g-demand one, IIRC. "Corrected for the path", or something like that. Cannot do the maths, but would be interested to know more.

deefer dog:
1/ No, unless switching off many computers (=not a feature/approved procedure). Against the Airbus philosophy.
2/ There is no such computer logic. Perhaps you've read Otelli litterally? Well, he's wrong here, there is no logic AoA>41°=Cancel SW.

The logic is :
- if current AoA > value calculated for SW then
-- SW is triggered

- if speed < 60kts then
-- AoA measurement is deemed not valid, whatever value the probe sends is "cancelled" (& AoA is flaged "NCD", no computed data); reason for doing that is [my guess] that will less than 60kt of relative wind, there is not enough airflow to get a correct/precise AoA measure from the AoA probes

The "problem" is that if AoA is "NCD", then AoA value (or "non-value") does not validate the condition for the stall warning: "NCD" is not superior to X degrees AoA.

Why do you call Autopilot "FMS" ?

Alpha Max equivalence was way behind.
Wasn't stickshaker already activated when AP was disconnected ?

Because FMS/FMC is the correct terminology for the autopilot in almost all modern airliners, and both Boeing and Airbus source theirs from Honeywell (which, incidentally, is one of the lesser known aspects of the early mode-confusion accidents - the "confusing" display of VS vs. FPA was identical in all Honeywell FMS displays, not just the Airbus units).

Autopilot is the correct term for the autopilot in modern airliners.

The FMS provides much more than just auto-flight capability, and provides it independent of A/P engagement.

And not all Boeing FMS's, nor Airbus FMGEC's/FMS's are Honeywell.

Dozy

The following describes (from the A330 Instructors Manual) the Pitch characteristics in Normal Law - same in Alternate Law.
http://oi56.tinypic.com/1z142nb.jpg

It can be deduced that the elevator response is related to pilot input and modified as function of Pitch Rate / Flight Path rate or G at High Speed.

As KCAS was low at the time you have been discussing, the elevator was acting more in accordance with pilot input, rather than G.