Machinbird

At Night? Wasn't the CB pretty much behind him and off to starboard by the time the pitots froze up? (assuming the 12 degree turn to port was reasonable).

TTex600

It is virtually impossible to understand how an Airbus flies when all you have for training material is FCOM's - and when the first rule says an Airbus flies like any other airplane.

OK465

Re: A320

Quote:

The only way my instructor could force Dir Law left us with no roll control spoilers, slow ailerons, and resulted in an extremely unresponsive airplane.

FYI TTex & CONF:

1. A330 with only one SEC FCC available (other FCC pb's off):

One pair of roll control spoilers (depends on which SEC) and one set of ailerons

(During full SS roll input, down going aileron only deflects about ˝ way, the up going aileron goes full deflection.)


2. A330 in ALT2 roll direct as a result of triple ADR disagree (ALT2B?):

Two pair of roll control spoilers and one set of ailerons

(During full SS roll input, NEITHER aileron deflects as much as they do with only one SEC.)


It's no T-38, but 'responsiveness', it appears, is in the eye of the ‘respondee’.

Thank you guys for bringing this stuff up.

PJ2

Most here know that the idea behind any close examination of any accident, this one in particular, is learning and hopefully prevention. I think the conversation regarding important factors such as intention, willingness, capacity (of both airplane and pilot) are especially valuable contributions to an understanding of what happened and perhaps a bit about why.

Some comments regarding the sidestick, apropos franzl's and Machinbird's comments:

The THS is manually set prior to takeoff. On takeoff, the sidestick is held slightly ND until 100kts at which point it is neutralized. It is moved back to initiate rotation, usually about half-way or a little less, depending upon CG. Sometimes the rotation rate feels as though it requires checking-forward on the SS but the FCOM warns against this and one maintains sidestick position until the climb is established.

As described the takeoff is done in direct law with pitch normal law gradually blended in within 8 seconds of pitch > 8deg.

Once the initial climb pitch attitude is established, pressure on the stick is relaxed. In fact, speaking technically only, one can let go the stick and the airplane will remain in the last selected pitch and bank attitudes. Sometimes that is a reasonable way of avoiding stirring the pot, as the airplane doesn't need any input if it is at the intended attitudes. In other words, no back-pressure is required to sustain the takeoff pitch attitude.

One can fly a SID with small, "squeezed" stick inputs for roll and especially for pitch, (only because tiny changes are all that are required for speed control). Even in turbulence, that is all that is required, because in normal law, everybody out there on the wing is busy maintaining the last selected attitude and stirring the stick only adds to the busy-ness without much effect.

The airplane can be flown to cruise altitude and, again technically speaking if planned in RVSM airspace, in cruise in normal fashion. At high altitudes one is always gentle with any transport and not just "Airbus", due to reduced damping of the thinner air. (Many here know this already...I'm being thorough, not pedantic!)

The airplane can be flown in turbulence, even heavy turbulence, (moderate, not severe - which I have never experienced). One keeps inputs small to avoid stress but other than avoiding small (and I emphasize momentary), inputs due to turbulence the aircraft is as straightforward to fly as any I have flown.

Descent is still 1g flight of course and the airplane can be flown from ToD to landing, and I often did this for the A320, less so for the A340/A330 but only due to long-haul (fatigue) issues and STAR complexities. Again, no stick input is required in this phase.

On approach, the usual small movements to counter slight changes in pitch and bank are normal. To alleviate any impression that NU stick is somehow needed in the approach phase, there is no "back-stick" required until 50' when the AFS feeds in a slight ND bias to give the feel of a conventional flare. Pitch direct law is gradually fed in beginning at 100' while roll remains in flight (normal law) mode until the pitch is < 2.5deg.

In terms of a subtle bias towards "pull" vice "push" due to AFS laws in various flight phases, I have to observe that the required operation of the stick as described coupled with my own experience would not cause a bias either way. The control laws do not engender this sense that one must "pull" most of the time - the notion just doesn't apply.

Unless we know what was intended, I think the notion of "intention" has no "correct" application here in examining what occurred in the first 30 seconds after the UAS event. Clearly, we can intend something with all our being and still be wrong. "Intent" can be informed by knowledge, training and experience, or it may not be, yet one may fully "intend" an action based upon incorrect assumptions, mistaken understandings or shortcomings in knowledge or experience. To me, this is an important part of examining the crew interactions. Willingness to accept another's interventions in such circumstances will be based upon the strength of intent (one's interpretation of what is wrong), and not always (reliably) upon knowledge. CRM is designed to counter the natural unwillingness to let go of one's own assumptions to adopt, (perhaps initially on faith alone), another's interpretation of what is happening. We have all been in situations when we know damn well that we are right, until we aren't, and we may even surprise ourselves that it was possible to be wrong.

Applying a steady, not just a momentary NU control input to a transport aircraft at cruise altitude to a 10deg pitch attitude is counter to all intuition, training and knowledge yet there it is, applied to the point of stall. Why?

What was the basis of the individual intentions which initially overcame and subsequently replaced an understanding of a number of fundamental aeronautical principles such as energy management, high altitude flight, aircraft performance and operational principles of SOPs and CRM? What was more powerful than these?

Post-stall, there are contributions from the airplane which led, perhaps inevitably, to further confusion but in the first moments nothing from the airplane indicated that a sharp and then sustained pull-up was the correct solution to a particular flight problem.

The rapidity with which situational awareness was lost and not regained is also an important area of learning.

These are some of the thoughts that occurred as I read and re-read through HN39's, franzl's, gums', Machinbird's & CONF iture's contributions highlighting the challenges of understanding what really happened.

Lyman

As ever, I remain convinced this accident had its beginnings with loss of a/p. We think we know better, but we do not. We assume PF knew right away the conditions, I suggest neither did. A/P loss due turbulence happens, and the 330 will remain in Normal Law. This would explain his cavalier attitude with Pitch, and focus on a newly twitchy roll axis.

We need to be careful about assuming things that have no direct evidence to support, and cognizant of possibilities that do not rest on a fundamental understanding of the recorded events. They had nil access whatsoever to records, none. They had their senses, we think, and what is left can be explained by any combination of cues and data that can be imagined.... It is common here, for some to base a condemnation of the crew on a single fact, and then assume truth by asking "Why"? I submit that is absurd... It is possible PF assumed NL until PNF announced Alternate Law. That is sixteen seconds of NU. NU that may have resulted from an erroneous, though forgivable, assumption of his, which thoroughly would explain his "negligent" back stick...

RetiredF4

Thank you for your contribution, PJ2.

You explain the handling of the SS ion the different phases of the flight, and i have no reason to think otherwise. But do we know, what the procedure for hand flying the A330 on long haul flights is with AF and how much practice in the real aircraft that left outside the T/O and landing regime? When do they take over in the landing phase? When established on final no input is needed until flare mode activates, correct? As you describe and as i understand the NZ law, no stick input is required if no change of flight path is intended. Can we talk about flying manual in this case?

Imho no, it is changing flightpath manual instead of pushing buttons or turning knobs, but that´s it. Bank angle compensation acounts for no need to make any pitch adjustment when turning and autothrottle takes care of the energy management . The FBW concept as implemented here reduces the necessity for pilot input to one single axis operation concept. There is no necessity to manage all three axis of the aircraft and the energy at the same time with elevators, ailerons, rudder and throttles at the same time, meaning with manual inputs trained in years of expierience. It´s reduced to "point and let go, the system will take care of the rest ". Wether you do it with SS , a yoke or the knop on the dash board is no longer important. It´s a nobrainer.

When AF447 dumped AP, ATHR, normal law and protections together with the speed indication in the blink of a second, the PF was forced to use strategies he was no longer trained for. No system was taking care of roll when he concentrated on pitch, no autothrust was taking care of the energy management, and no bank angle protection available to stabilize the pitch during roll.

Let´s look at a different thing, which keeps me thinking about:

We know, that the aircraft maintains stick free 1g, and that SS commands a change of g blended with pitch rate starting below 210 knots. In a newspaper article from the early A320 flights i remeber, that the ratio is 50/50 at 150 Knots. The speed responsible used for this changeover was faulty, down to 60 knots which would equal nearly a pure pitch rate change. But the aircraft was traveling still at over 230 knots in the regime where only g command should be present. I could not find any reference from where we could draw a conclusion, wether this different SS command would have influenced the outcome of any SS order. There is reason, that this changeover takes place in the low speed regime, what influence does it create wehen this changeover is taking place when the airframe is still above the change over regime in degraded mode?

The assumption, that the PF initiated and sustained that climb intentionally disregarding all basic principles of flight, busting assigned FL, disregarding CRM and not anouncing his intentions neither to the PNF nor later to the captain explains all following events. It´s an easy solution (just change the pilot everything else is fine), but it´s based on thin or even no evidence.

PJ2

Hi franzl;

Thoughtful response, thank you. The points are eminently worth discussing, and for the time being concluding who's theory is right is less important than the discussion!

By all means, and what you say in your next paragraph, (. . . "It's a no-brainer"), is in my view largely though not exclusively true for the present standard of training and expectations, a clear risk which is now being widely discussed, examined in the many conference topics since AF447 and how has the attention of at least the US and European regulators, (in fact AW&ST provided articles in August, 1989 and January/February 1995 discussing these aspects in automation development).
I think your comment may be accurate in this case and this case may, among a couple of others, be a harbinger of such a developing trend. However, the Air Caraibe event and thirty-odd others which occurred prior to and also after AF447 do represent counter-examples to the "no longer trained for" point because to a greater or lesser degree, these were successfully completed flights. I don't believe such circumstances would leave the majority of crews in a situation they were no longer trained for, at least, quite frankly, I hope not because this wasn't in and of itself and all else being equal, (I recall, and take your interesting point regarding the pitch/flight path item), a serious emergency, (as in loss of pressurization, hydraulics, electrical power generation, engine thrust or engine disintegration, etc). If we consult JACDEC or Aviation Herald we can read about a number of incidents, events and near-accidents in which crews addressed them as trained and which did not result in loss of control or loss of the aircraft.
The evidence is in the absences of an expected standard initial response to an abnormality, which, I will add, is a point which thus far has yet to be discussed and countered.

To my knowledge and experience as an airline pilot there are no circumstances, save for perhaps extremely rare and dire events, in which SOPs, CRM and discipline take a back seat. These responses are proven, primary responses, heavily-emphasized and trained in airline operations and to deviate from them requires significant operational factors.

No such factors or events prior to the stall warning are in evidence. If the airplane pitches up due to some anomaly, one tries to get it back down to stable flight, period, yet the inputs are mostly NU; one does not permit the airplane to do what it will, not, at least, without vigourously trying to counter what it is doing. If one doesn't counter the anomaly, the evidence is that one agrees with what the airplane is doing and it should "do more". So no, this is not an "easy solution" which dismisses this crew out of hand. This is an extremely difficult solution to come to terms with because it is human factors-based and one must be very careful to examine such factors while avoiding the narrow focus of "blame". Finding out "why", despite some commentary to the contrary, is the way to prevent this kind of accident. If there are training and standards issues, that needs to be examined as do priorities in terms of autoflight and manual flying it is to be discovered first through this unfortunately-blunt process.

BOAC

- and I would dearly like to know why, but as I said months ago, I fear we are 'dancing on the head of a pin' here since the reaction of this crew was so out of the ordinary as to render, in my opinion, any logical approach to this accident of no value, except in re-focussing at least AF's attitude to basic flying skills in their crews and some vital changes to AB software and philosophy.

That not-with-standing, the dissection of the FBW system has been of great interest and my hearty thanks to all who have taken the trouble to expound.

HazelNuts39

Perhaps the final report answers your question, at least part of it? At any rate, the response of the FCS and the airplane to longitudinal sidestick commands is recorded in the traces of elevator/THS and pitch attitude.

bubbers44

Safely at FL350 maintaining attitude would have saved them. They did the low level terrain profile of 15 degrees nose up and toga power. You can't do that at FL350. They screwed up and I think everybody knows it. They were both autopilot monitors, not hands on pilots.

Lyman

HazelNuts39

"In alternate 2 law, the longitudinal control law remains a load factor law and the lateral control law is a direct law. In the specific case of alternate 2B law, some coefficients used in the longitudinal flight control law become speed-independent and are set for the maximum speed for the aeroplane configuration (330 kt in clean configuration). This hardly modifies the behaviour of the aeroplane in comparison to normal law, but can nevertheless induce an unusual response dynamic when the aeroplane has an abnormally low speed for the configuration."

You have said before you believed the "unusual response" might be sluggishness. As an aside, could it be the reverse? Could the Pitch rate be extremely high, if the coefficients are computed for Indicated speed, when it is erroneous? Would that not impart an emphatically increased authority to the elevators, relative to pilots inputs? Especially if the actual velocity was .80 Mach?

If the responses were sluggish, would that not possibly explain PF's dogged pursuit of some G? Might he have been seeking the twitchiness in Pitch that he found in Roll? If he thought the aircraft was unresponsive in Pitch, would he not command as much as he could get? If he sussed sluggish, and got no immediate "response" would he be tempted to trade rate for continuation of input?

He could not have seen AoA, and he would be using the horizon as his guide, which although high, did not jive with his sensation of load? Much of the g post 1.65 and the reduction to 1200 fpm from 7000fpm was less than one, was he flying knowing that his Pitch was high, but negating it to some extent because he did not truly believe the Pitch jived with the loading? Nose high and descending without a Stall Warning would communicate that one of them was wrong. Especially without Airspeed.

Did he choose to follow the wrong cue? The lack of gee over the horizon?

gums

Hey, Lyman! The coefficients are most likely the "gains", as we FBW pukes call them.

The basic gains are mostly for determining the amount of control surface movement and the rates they move. They are also modified according to configutation, so gear down they will be generally lower in movement or rate, but it depends on the jet. In the Viper, the movement was higher, but the rate was lower.

When our pitot-static system went FUBAR we went to "Standby Gains", which were about 140 -160 knots gear down and 300+ knots CAS gear up!! Surprisingly, almost the same as the 'bus on the high end. Hmmmmm.... Remember that we still had the gee and rate limits, but it still worked well, and we were used to high gee maneuvering.

Bottomline is that the 'bus gains actually provide a kinder, gentler handling when in the back-up mode. Not skittish or overly sensitive.

My personal beef is the lack of AoA inputs beyond the warnings. I could live with basic AoA limits ( "protections") throught the envelop unless in "direct", and then clear warning indications that I was gonna stall the sucker.

Alber Ratman

Thread... Please DIE!

Lyman

Hi gums....

You say.... "Bottomline is that the 'bus gains actually provide a kinder, gentler handling when in the back-up mode. Not skittish or overly sensitive."

I think it is possible our PF noticed the above in Pitch, and when he determined his inputs could not effect rate, as in Roll direct, he traded rate for "holding" aft stick, (persistence). It would be the first time he experienced different "Law" in two axes. In selecting Roll as his preferred (Direct), he may have decided his efforts would have more effect in Pitch if held. A mistake, surely, but he had no experience flying by seeking g, another first time only time. Til then, the aircraft had always done it.

Bear in mind, most everything that happened after a/p loss was a first time event in the aircraft, along with a deluge of alerts. He appeared terrified of G<1.

His training had been nothing like yours, not his fault. I don't think it is equitable to slam him for trying what he thought would work. Not that you are, I think you get what he was up against. The sobering thought is that he was one of thousands who might have the same responses to unexpected gremlins.

Respect,

john_tullamarine

The sobering thought is that he was one of thousands who might have the same responses to unexpected gremlins.

.. and that's why this series of threads is so important ... sorry, Alber Ratman

bubbers44

We must always remember how to fly by attitude and thrust, no matter how automatic the airplane is. If we don't we are not really pilots. We are programmers of autopilots. Don't let it happen to you.

gums

Nice concept, Bubs.

Although my IP had explained a lot about AoA and lift and such, on my first flights he taught me to use pitch attitude looking at parts of the canopy rails and such. Hold an attitude at a power setting and watch airspeed. Slowing down? Lower the nose a bit, and vice versa.

There are few planes in the commercial arena that will suffer catastrophic failure if you simply "hold what ya got" if you have any warnings except terrain collision. Sure, complete loss of power such as Sully had requires keeping airspeed for maneuvering and such by lowering the nose. But even my LEF emergency only required that I keep doing what I was already doing ( lots of sidestick pressure to keep roll under control and yaw trim would come later).

Another pilot here has already commented that the nifty flight director bars helped to get on course or intercept the ILS or...... But I can tell you that when I got my interceptor assignment outta pilot training that the Air Defense Command liked those of us that had flown the old T-33. No fancy ADI or steering or such when on instruments. "Primitive" would be a good description. So they let us fly target missions and such after only a single checkride. The T-38 troops had to wait until they were at their permanent assignments.

First jet I flew with the steering bars was the SLUF, and it was very "loose". Most of us did better using the raw data, heh heh. Years later I got to the Viper and no steering, just raw data( see my video of the LEF landing approach). Hmmmm.....

Somewhere in the aftermath of this tragedy, the powers that be must take a deep breath and implement better training. Ya think?

CONF iture

This seems to match with the FDR data where only INB Ailerons were solicited at a max deflection of 16 degrees out of 25, and spoilers 4 and 5.

After PRIM and SEC1 were turned OFF, only spoilers 4 remained active. I suppose the zipper shape for spoilers 5 and 6 show their inactivity related to PRIM and SEC1 status.

Another hit at the BEA :
Why no further comment on the PRIM and SEC1 episode, especially in regards to the human factor analysis ?
IMO, this one clearly indicates that the crew (or at least the PNF) lost faith in the aircraft, and probably thought they were facing a computer bogus, or a protection takeover.

Lyman

CONFiture,

Can you expand on that? Switching off the PRIM came up right away, did it have to do with trying to get the bird?

"We have tried everything"...

I have sensed the presence of workarounds in this cockpit, how are you meaning "lost faith in the a/c"? How does the suspicion of Prot take over fit in?

BEA have left all out but the rudiments, what are they doing? Do you consider the zipper to be generic, or specific to each instrument, mode or system?

nikplane

Hi all.


AF 447.

In this terrible disaster of flight, if you allow I would want to remember to you the following points:


1) the strong vertical movements of the air inside the cumulinembo CB.

2) with pitot-tube clogged (iced) the airspeed instrument then acts like an altimeter.


What do you think?

From 4 Aug 2011 I read and reread-reread all your messages about A330 disaster.