How safe is (airbus) fly by wire? Airbus A330/340 and A320 family emergency AD

[RetiredF4]

Flying the A320, Flight International 6. July 1985

In training, the "follow me through on the controls" instruction will no longer be possible. There may also be some room for doubt when a captain, dissatisfied with a copilot's rate of flare, edges in a bit more "wrist back," just as the copilot also makes a correction. A lit caption to warn of simultaneous input has been considered. Airbus has argued that the captain can assess the progress of the flare visually, but not all pilots are convinced that this is the complete answer.

[Lonewolf_50]

@Lyman - To what benefit? In airliner operation the PNF/PM is supposed to be monitoring the aircraft, not the other pilot.

Dozy:

Sorry, old chum. You are wrong about that.

In a crew of two, you do indeed have to keep aware of what the other pilot is doing, particularly when he/she is flying and you are not. A great way to save your own life, his/hers, complete your mission, and retain the confidence of anyone in the back.

Standard CRM.

[DozyWannabe]

Originally Posted by Chris Scott

Fair enough, and I'll do my best! But please also read my post of January 02 @ 1959Z.

I did then, and I've just re-skimmed it now to remind myself. Thank you very much for taking the time to inform me. Please understand that any of my responses are intended to be in the spirit of discussion only and are not intended to disrespect your view.

1) Tactile feedback to the PNF for monitoring, but that only applies if he/she is touching the stick.
...
The advantage of the conventional yoke (and this also applies to old-fashioned joysticks in tandem cockpits), is that (peripheral) vision of its movements gives the PNF some idea of what’s going on.

Granted - this is a real benefit, but this kind of monitoring in the sense of correcting a less-experienced pilot tends to happen less than it used to. When cadets were being trained by doing go-arounds and circuits in Tridents and the like - essentially doing it for real - this was arguably necessary. But in this modern era, when simulation has progressed to the point that such is the fidelity of the experience in terms of response and handling that the only major difference from the real thing is the psychological knowledge that it is indeed a simulation, could it not be argued that this becomes less of a necessity?

Additionally, the way the A320 should be handled according to Airbus procedure (small input-observe-correct if necessary) is slightly different than for more conventional types to counter any risk of over-controlling, as well as the minima and numbers being determined in a much more scientific and strict fashion than had been done before, with a significant margin of error embedded within the system. I suspect the intent of this was to provide an extra margin of safety compared to the methodologies employed before.

2) The ability of the PNF to takeover control without pushing a button. Pilots are very reluctant to do that, because it would be a clear expression of no-confidence in the PF‘s handling.

Is it really any more of a clear expression than the "I have control" (or "My aircraft" in the US) call-out? Begging your pardon, but to me it simply seems like a technical codification of that procedure, which has been around for as long as aircraft have had dual controls.

The reluctance of the PNF to interfere applies on traditional types also (see above).

Believe me this is something of which I'm well aware, but I'm hoping that coming from you it will have more clout than it seems to coming from me!

The frailty of the human condition is not unique to Airbus cockpits! But momentary intervention (like correcting an undesired wing-drop on a gusty, crosswind landing) is easier on traditional types.

Agreed, however - as with your first point - the procedures for flying the A320 are slightly different from those in conventional types in order to balance that out. The gusts were bordering on minima in the Hamburg incident and the Captain should not have permitted the F/O to continue the approach if he was not confident in her ability to handle it.

Having said that, the Airbus in Normal Law is, IMHO, a more forgiving aircraft to fly “manually” than most of those. Can’t (and won’t) second-guess the Boeing version.

From my understanding it's not a great deal different from the 767. What I have been told is that the software system providing the servo feedback alone is more complex than the A320's flight logic in its entirety. This is why I get a little chagrined when I read posts that berate the Airbus FBW system for being too complex and insist that the T7's system is superior in the same breath.

In 14 years of line flying the A320, starting from type-certification, I never found myself in other than Normal Law.

Quite - while she may not have been an easy bird to love at first, it seems that a growing number of pilots have found that she's not as bad as they feared - and in fact she's a tough, reliable and friendly companion.

@LW_50 - Please permit me to re-phrase. Would qualifying the statement with "the *primary* job of the PM is to monitor the aircraft, not the other pilot", be more acceptable? Otherwise every ride is, in effect, a check ride.

EDIT : This article:

Old Facts, New Insights – Lessons from A-320 » Aviation Medicine :: Aerospace Medicine | Aviation Medicine :: Aerospace Medicine

posted by jcjeant back in the AF447 thread some time ago contains some interesting information and is worth a read, but it was pointed out to me that on this page:

Old Facts, New Insights

of a survey of 167 pilots in a US airline that had adopted the A320 (and I think I'm right in saying US pilots tend to be more conservative than most), after some time with the aircraft only ten of them still had misgivings about the lack of interconnection. That's only 6%!

[Clandestino]

How safe is (Airbus) fly by wire?

Fly-by-wire is just generic technical term like "aileron", "flange", "wheel" , "empennage" or "cable". It is flight control systems, it has nothing to do with automatic flight systems. In real world (the stuff outside internet) any aviation system has to be deemed airworthy. There are entities given the task of airworthiness certification and their approval must be confirmed by the events occurring in everyday use or it gets revoked. That's the way things work in the real life.

Airbus fly-by wire is safe enough that:

1. No civil aviation authority has put its airworthiness as a system into question.

2. No statistics, (and this is statistics as a branch of mathematics, not some numerology liberally applied around PPRuNe posing as statistics) or accident/incident report prove that Airbus FBW is inherently unsafe.

3. Folks trying to sell their gut-feeling that FBW is unsafe as a fact have to use huge amounts of distortion and aeronautical ignorance to have their theories sound plausible. Whether their theories are ignored by aeronautical-powers-that-be because of severe lack of resemblance to anything meaningful and true or due to worldwide conspiracy of Illuminati taking over the all CAAs , well anyone can make his own decision. More or less paranoid.

Anyway, just when I hoped that threads with abundance of nonsense about Airbus here on PPRuNe somehow went the way of the dinosaurs, someone triggered new outburst of same old, hundred times refuted Airbus fairytales.

If you are really interested in AB airworthiness following the EVA incident and not in another A vs B idiocylimpics, investigation is underway. AD is issued as a completely justified overkill measure that should help all the crews cope if that one-in-a-couple-of-million-hours-event reoccurs before we exactly know what happened. It might turn out to be another case of "Perpignan pressure washing". It might be that new probes are worse by design. It might turn out they are not but they were both from defective batch. It might turn out to be freak icing. Or any of other couple of thousand possibilities. Wait and see.

Originally Posted by TyroPicard

Good heavens, what a fruitless discussion.

I beg to differ. Testimonials along the lines"I have flown such-and-such aeroplane for so-and-so hours" followed by hilarious display of ignorance of basic aeronautical facts are some kind of fruit. Not particularly palatable if one is more interested in meaningful aeronautical discussion than Beckettian satire, yet fruit anyway.

Originally Posted by Chris Scott

The similarity of the two readings is thought to have misled the crew but, as Clandestino also points out, that assumes the crew also failed to ensure that the relevant FMA was indicating FPA (if the theory is correct, it would have been indicating VS).

Yes, but it was not just the unread FMA that doomed the flight. David Learmount has nicely summed it up:

As if failure to spot unwanted 3200 fpm descent is somehow on the par with the getting the sum wrong in the first grade arithmetics.

So, yes, FCU selector had the design flaw but its redesign was just one small contribution towards prevention of future CFITs. Major points of Strasbourg investigation are: know where you are, know where you need to go and don't trust the autopilot blindly or basically: Flight order says you are a pilot, so act like one for 's sake!

Originally Posted by Chris Scott

Some FMA changes are worthy of mandatory callouts

Lufthansa and Air Berlin have the right idea; not many of them are. It is important to know what autopilot is doing but it is more important to know what the aeroplane is doing, and despite the doomsayers spilling the bile over alleged lack, there are abundant clues in Airbi to tell the pilot what is going on.

Originally Posted by Chris Scott

the practice of dialling zero feet into the altitude selector and selecting OP DES ("open descent") with A/THR was always a hostage to fortune

Err.... perhaps that's the reason I was never, ever allowed to dial anything lower than FAF or initial G/A altitude on any aeroplane, not just Airbus. Allegedly, practice was developed soon after altitude selectors made inroads in my country's aerospace and that was quite some time before I was born.

Originally Posted by Chris Scott

which is clearly indicated as IDLE on the thrust FMA

Yes, but despite the name, Airbus is aeroplane, not bus. It is even more obvious on the thrust display, be it EPR or N1. Woe to the any powered aeroplane pilot who doesn't pay attention to his power gauges.

Originally Posted by CONF iture

.. as should be a protection !

If Airbus could once for all acknowledge this, they would nicely integrate somewhere in the instruments panel that single guarded switch the pilot could sensibly push to allow him to regain control in the shortest time when protections go wild.

Find me a case of protection gone wild that is not QF72 or MH124.

There are recorded cases of stickpushers trying to prevent the stall but confused and panicked pilot overriding them with fatal results. So both soft and hard protections have their good and bad sides, as (in the real life, outside of PPRuNe's Gedankenexprimenten) everything technical has. I wonder whether the locked groove track of "we want the override" would change if someone kills himself while overriding the functioning protection.

Triple sevens are very expensive widebodies. They are not abundant with less than scrupulous operators.

Originally Posted by The Robe

Mark my words, it will come out that you can't ground an Airbus enough to with stand a direct 100,000 volt lighting strike and not fry something.

Airbi get lightningstruck regularly for last 20-odd year without dire consequences. Your words are marked. Not for the reasons you'd find agreeable, though.

Originally Posted by The Robe

Air France went down because those guys lost situational awareness at night in the soup. They didn't know if up was up, down was down, if the plane had air going over the wings or not because the screen weren't given them the information. The CVR conversation confirms this. Three guys looking at the panels and none of them knew what to make of it.

Correct.... nothing to do with the machine.

Originally Posted by bubbers44

Toga is for go around. Everybody knows that.

Well then airline pilots just don't fit your definition of "everybody". There is more to TOGA then just go-around, however it is indeed not particularly applicable at FL330 over the ocean.

Originally Posted by Lyman

Let's ask the pilots who fly them...

They just don't care about some loud, arrogant and largely ignorant group pretending to be the voice of the silent majority.

Originally Posted by Bengerman

Others here have chuntered on about training, procedures blah blah....yes, they are vital but it has to start with the actual piece of kit in the first place!

No, it absoly must not! Training has to start with basics, type rating is no place to be learning them and do-whatever-you-need-just-control-your-attitude is very basic stuff! RTF Airbus manual! Disclaimer is on page two or thereabout!

Originally Posted by A33Zab

1000 T7s is not a minuscule fleet but only 12,6% of total boeing FWB and airbii FBW combined.

What about Embraers, 7x, Tu-204 and SSJ? Are we discussing just civilian FBWs or perhaps it is indeed not about FBW but rather another A vs B treadmill?

Originally Posted by Dozy Wannabe

Interconnection is not objectively better no matter how much you and others claim it to be so.

Exactly. It has good sides and bad sides with no controls configuration decisively proven to be better.

Originally Posted by Lyman

I think isolating sidesticks from view and from feedback, PLUS NO Connectivity, is a terrible idea.

Did you tell FAA about this?

Originally Posted by Lyman

No conclusions. Not yet.

FAA's current conclusion is that Airbus FBW is fine as it is.

Originally Posted by Dozy Wannabe

you're supposed to trust your colleague, no?

It's not just your significant-cockpit-other, it is ATCO, mechanic, weatherman, dispatcher and many more folks dealing with your flight that you have to trust enough to put your life in their hands but not enough to let anything they do and say go unchecked.

Originally Posted by hoggsnortrupert

Does Air Bus acknowledge that its machinery is flown by PILOTS, or are they SYSTEM MONITORS:

I have no idea what Air Bus is but Airbus does. Preface of their manuals says it explicitly so.

Originally Posted by hoggsnortrupert

In my opinion A/B is ignorant of HUMAN FACTORS:

EASA should be informed of your opinion immediately. If you are right, then safety is seriously compromised. We must not allow this to continue, must we?

Originally Posted by hogsnortrupert

it is quite disturbing:

Entertaining, too.

Originally Posted by Retired F4

Flying the A320, Flight International 6. July 1985

Wow. Some things never change.

Did you know that about-face in some cultures is viewed as sign of weakness, even if it would make one's opinion more aligned with the freshly available facts?

Originally Posted by Lonewolf 50

In a crew of two, you do indeed have to keep aware of what the other pilot is doing

In a crew of trainer and trainee during basic training it is indeed important to know what the trainee does with controls. Airliner pilots usually get formally trained when and how to take over the controls and learn to watch the aeroplane's behaviour, not what yoke or stick does. That some of them are going against their training (or they weren't given it at all) could be a leftover from their instructor days or lack of confidence in their copilots - mind you, it needs not be paranoid but pretty justified, depending on the operator.

Makes me wonder how anyone ever got rated on Airbus if "tactile feedback" from other side is indeed so important.

[RetiredF4]

Clandestino
Did you know that about-face in some cultures is viewed as sign of weakness, even if it would make one's opinion more aligned with the freshly available facts?

Now it´s time to leave this thread, as the BS hits the fan.

[DozyWannabe]

A few additions...

Originally Posted by Clandestino

Major points of Strasbourg investigation are: know where you are, know where you need to go and don't trust the autopilot blindly

Or the ATCO - he gave an incorrect radar vector that put them significantly left of centreline.

Find me a case of protection gone wild that is not QF72 or MH124.

Both of which were recovered and landed safely, the latter of which was in fact a T7, not an Airbus.

There are recorded cases of stickpushers trying to prevent the stall but confused and panicked pilot overriding them with fatal results.

D.P. Davies explicitly lamented the knee-jerk reaction of a minority of pilots that condemned the introduction of the stick pushers in HTBJ.

[@RF4 - Clandestino may be too pugnacious for your taste, but if you fact-check him, he's rarely misinformed!]

[Lonewolf_50]

Reading comprehension time, Clandestino.

In a crew of two, you do indeed have to keep aware of what the other pilot is doing

I didn't say you have to ride the controls, did I? No.

"Trust Him!" is a good way to die, which is one of the reasons that CRM was developed. What the other pilot is doing when you aren't flying is usually shown to you on the flight instruments, by aircraft performance, etcetera. Of course, there are times when what the other pilot is doing is just what you are doing: monitoring what the robot is doing for you both.

Nice to see you back in the scrum.

[Clandestino]

Originally Posted by Lonewolf 50

Reading comprehension time, Clandestino.

(...)

I didn't say you have to ride the controls, did I? No.

Did you just admit that non-linked controls are indeed fine for airline flying?

PNF checking what PF is doing is standard CRM. Performing it by looking at the yoke (or feeling it) instead of at the instruments is someone's very wrong idea of how it is done in airlines.

[Lonewolf_50]

Did you just admit that non-linked controls are indeed fine for airline flying?

When did I ever say it wasn't? Make sure you know what conversation you are having, and with whom, before you toss out wisecracks, OK? There is more to this discussion than "must link yokes" "sidestick fine" and you may note that I was pointing out -- in the post of mine that you excerpted - to a non pilot (Dozy) that he was wrong about how that flying thing works in multi place aircraft.

PNF checking what PF is doing is standard CRM.

No kidding.

Thank you for returning my own comment back to me. We are in violent agreement.

Happy New Year, and glad to "see" you again.

[DozyWannabe]

@LW_50 - With respect, Clandestino seems to understand what I was driving at even if I phrased it badly (note I attempted to re-phrase above) - namely that you're not generally supposed to constantly monitor the PF through linked controls, rather more usually through monitoring the aircraft and by communicating with them.

By "trust them" I certainly wouldn't expect that to mean trusting them blindly - if you think they're doing something wrong then of course doing something about it is paramount.

[Lyman]

In an emergent abnormal, it would be nice for PNF to check the controls, both of them, since the aircraft is climbing, the PF swears he is commanding ND, the aircraft keeps climbing, at a higher rate, the airraft keeps climbing, it STALLS, and ten seconds before ending up as millions of pieces on the sea floor, he admits: "I have been holding aft stick for some time..."

Most of the time connected or visible SS are not needed.

"why bother? It hardly ever fouls up...." Or, "even with connected yokes, planes crash..."

Stop talking standard, and start addressing abnormal, please....

[DozyWannabe]

Originally Posted by Lyman

the PF swears he is commanding ND

Where did he do that? I don't think he did, at least given the CVR and FDR evidence. No discrepancy between input and the aircraft's behaviour in that case ever existed, and there's no evidence to suggest any instruments other than ASI were malfunctioning. There were plenty of ways to see that the PF wasn't doing as he should have been.

[Lyman]

Yes, except the most obvious one, look at his hands.

You constantly try to distract from every question. They knew they had instrument issues; in an abnormal, everything is on the table, or should be.

"Yes, we go down..." how prophetic.... PNF had no choice but believe in PF. There is not necessarily something wrong with that, but after Stalling, the other two had no chance to cross check attitude with inputs. it is obvious that if the a/c had not been subject to invisible inputs, there is a strong chance the flight would not have been lost.

Cue "But other aircraft have crashed with visible controls..."

Over to you

But you rebut BEA as well... "The lack of Sidestick visibility may have been a factor..."

[DozyWannabe]

Originally Posted by Lyman

Yes, except the most obvious one, look at his hands.

It wouldn't be the "most obvious" to pilots (such as those we're discussing) whose main familiarity was with Airbus FBW types because they know that option is difficult and are trained to monitor using other methods.

An anomalous footnote (and I'm not rebutting BEA - "may have been a factor" does not mean they're recommending significant changes to the PFC implementation) to an accident which raised far more pressing technical and HF concerns is not going to be sufficient to compel a manufacturer to alter their entire control philosophy - one which has proven successful and as safe as any other (remember, yoke position has been ignored by PNFs several times in other accidents where the crew pulled up into a stall). I'm guessing this is why the "rumble" compromise has been put forward for the A350 (in response to the BEA's phrasing on the subject).

In this case it seems that the PNF allowed himself to become distracted from properly monitoring the aircraft and, by extension, the PF, by focusing on waiting for the Captain to return - that should be a far bigger worry.

[CONF iture]

Originally Posted by Cough

Technically, the speed will get to Vfe (flap 3/full) and then the aircraft will pitch up...

No - High speed protection applies to VMO not Vfe.

[CONF iture]

Originally Posted by Clandestino

Find me a case of protection gone wild that is not QF72 or MH124

Eva air seems to be the most recent one.
And the MH 777 does not qualify for the 'wild' definition as the crew could always have control over the protection.

I wonder whether the locked groove track of "we want the override" would change if someone kills himself while overriding the functioning protection.

I don't know about that 'someone' I know about my airplane which is able to give full priority to a wild protection and blind fully ignoring my more reasonable command.

[Lyman]

Why has no one mentioned the uncommanded excursions of Airbii that prompted OEB: "do not reactivate AP, "uncommanded climb can result..." "?

Air Caraibes?

[DozyWannabe]

Originally Posted by CONF iture

And the MH 777 does not qualify for the 'wild' definition as the crew could always have control over the protection.

Not in that instance. Check the article I posted in the other (T7 AoA) thread. According to the article (which quotes the ATSB report) the 777's systems re-engaged autothrottle uncommanded at least once after the crew had disengaged it.

[CONF iture]

Originally Posted by DOZY

According to the article (which quotes the ATSB report) the 777's systems re-engaged autothrottle uncommanded at least once after the crew had disengaged it.

Where does it say the crew could not have control over the protection ... ?
Quote from the ATSB report will do just fine. Leave the article to Clandestino.

[Lyman]

Praise God that 'articles' are different from "reports". Can you imagine if, like the uncommanded climb that can happen on A330, the AF:447 succumbed to same?

Now that inspires confidence....