/ Had Air France AF477 been a Boeing....

This topic has been archived, and won't accept reply postings.
Bjartur i Sumarhus on 30 Apr 2012
a good chance it would not have crashed

Amazing conclusion and quite unnerving read

http://www.telegraph.co.uk/technology/9231855/Air-France-Flight-447-Damn-it-were-going-to-crash.html
GrahamD - on 30 Apr 2012
In reply to Game of Conkers:

I don't think that is the first conclusion I would have drawn from reading that.
butteredfrog - on 30 Apr 2012
In reply to Game of Conkers:

Why would a trained commercial pilot carry on pulling the stick back, and carry on trying to climb through a stall warning, especially if he has 30,000+ ft to play with?
Also why was he so confused that he thought he was in a landing abort situation when he took control? (asleep, just wokeup?)
I know they say there is a lack of feedback in the cockpit controls, but you would have thought the captain having to walk uphill to the cockpit would have given him a clue.

I haven't done a great amount of flying, but the couple of stalls I have had demonstrated to me (light aircraft) were quite violent(ish) with very obvious changes of aircraft attitude.

Cheers Adam
mypyrex - on 30 Apr 2012
In reply to Game of Conkers: I read that article and found it quite interesting. It does seem to me to be a bit stupid(if I've read it correctly) that the movements on one "stick" is not replicated the other. Also seems daft that the physical throttle levers remain stationary irrespective of the auto-throttle adjustments.

Although I've never flown (as P1) anything bigger than a twin engined Cessna I think that the actual physical feel of the controls should always be available to the pilot.
mypyrex - on 30 Apr 2012
In reply to butteredfrog:
> (In reply to Game of Conkers)
>
> Why would a trained commercial pilot carry on pulling the stick back, and carry on trying to climb through a stall warning, especially if he has 30,000+ ft to play with?

I did wonder that myself.

Some aircraft I have flown were equipped with "stick-shakers" to give an enhanced warning of an approaching stall.
Bjartur i Sumarhus on 30 Apr 2012
In reply to GrahamD: Apart from pilot error, I think the conclusion that different cockpit design features between the two manaufacturers could have made a difference is very interesting and is entirely plausable.

That's not to say that the Airbus joystick and engine thrust lever design is wrong, but in this instance it certainly didn't aid the other pilots assessment of what on earth the junior pilot had been doing.Yes, all the information would be available on various monitors, but in a split second decision, you would hope the captain would have seen immediately on returning to the cockpit that Bonin had the joystick pulled back between his legs whilst the "stall" alarm was sounding. But in the Airbus, that is not apparent at all.

It's good to know that pilots now rehearse this exercise in simulators.
Toby S - on 30 Apr 2012
In reply to butteredfrog:

Mypyrex will know more about this than me but don't planes fly with their noses slightly up anyway? So he may not have noticed a few degrees increase in pitch.

Utterly tragic regardless, reading that article I'm thinking Boeing have the right idea.
mypyrex - on 30 Apr 2012
In reply to butteredfrog:
> (In reply to Game of Conkers)
>

> I haven't done a great amount of flying, but the couple of stalls I have had demonstrated to me (light aircraft) were quite violent(ish) with very obvious changes of aircraft attitude.

I believe with a lot of bigger aircraft the stall can be a lot less noticeable, especially if there is no external visual reference(ie at night or in bad vizibilty) but, as I said before, I have no experience on large commercial aircraft so I am only going on what I've been told.
mypyrex - on 30 Apr 2012
In reply to Toby S:
> (In reply to butteredfrog)
>
> Mypyrex will know more about this than me but don't planes fly with their noses slightly up anyway?

I think you're right as far as airliners are concerned. Next time you're on a flight somewhere and get up to go to the loo you might notice it (about 4 - 5 degrees I think)
David Martin - on 30 Apr 2012
In reply to butteredfrog:

Its very hard to tell which way is up in cloud/darkness and roll, yaw and pitch will create artificial sensations of completely different attitudes than you are currently in. Quite understandable that they wouldn't know which way is up if not referring to or losing trust in their instruments (inconsistent readings). But their failure to follow the instrument scan and account for a blocked pitot is an extremely basic failure in airmanshap. First rule in all problems is "Fly the aircraft".

I'm certain their are control surface indicators on the displays, and while probably not foremost in the two pilots minds, were surely visible.

As for the Boeing/Airbus difference, the US aircraft are prone to their own design quirks. The interplay between trim and elevators springs to mind as having caused a crash at Schipol a few years back (737?) which probably wouldn't have occurred in an Airbus. No doubt there are many more incidences attributed to both design philosophies.
Markel on 30 Apr 2012
In reply to Game of Conkers:

There is a slightly more detailed write-up from last year that deals with a few of the psychological aspects. Fascinating read if slightly sickening. I've been told that confusion is a very common state for the pilots of crashing Airbuses for exactly these sorts of reasons.

http://www.popularmechanics.com/technology/aviation/crashes/what-really-happened-aboard-air-france-4...
mypyrex - on 30 Apr 2012
In reply to Game of Conkers:
> (In reply to GrahamD) that Bonin had the joystick pulled back between his legs

I think the "stick" in the Airbus is mounted to the pilot's side and would not necessarily be visible from anywhere else in the cockpit - which could have been part of the problem.
Bjartur i Sumarhus on 30 Apr 2012
In reply to mypyrex:

It does seem incredible that between the three of them they couldn't figure out what was happening.

I thought the "stall" alarm coming on when they managed to correct the climb into a dive very odd, and even odder that Bonin responded to this alarm by resuming the climb. AF will have difficulty avoiding liability for this accident I would imagine
Bjartur i Sumarhus on 30 Apr 2012
In reply to mypyrex: Exactly, my point being that in a Boeing it would have been a larger joystick pulled back between his legs....maybe I didnt make that clear
David Martin - on 30 Apr 2012
In reply to Toby S:

Its probable the AF flight remained wings-level throughout the stall, hence no spin or possible spiral dive. The fly-by-wire system more or less ensure that. Therefore the pilots had a very real chance of recovering the situation at any point prior to impact if they had followed the most basic of stall response rules.

If this had been a Boeing aircraft, and keeping in mind I have no idea of the stall/spin characteristics of either type, on entering a stall it is entirely likely this could have turned in to a spin or eventual spiral dive. If the later this could have led to an uncontrollable situation long before impact.

As much as a number of crashes do appear to have resulted from the Airbus design, it is equally certain that a great many have probably been averted. The interim report does make it look very much like pilot error.
lithos on 30 Apr 2012

interesting reading - why dpont they put the 'side' stick on the other side - ie in the middle between them, where they are visible ?
mypyrex - on 30 Apr 2012
In reply to lithos:
>
> interesting reading - why dpont they put the 'side' stick on the other side - ie in the middle between them, where they are visible ?

Because that's where the throttle quadrant is.

mypyrex - on 30 Apr 2012
Neil Williams - on 30 Apr 2012
In reply to Game of Conkers:

Given that air crashes are usually caused by pilot error, I think there's a good chance that the Airbus automation has prevented the odd few crashes as well.

With an Airbus in normal conditions, the aircraft will not allow pilot inputs that would take the aircraft out of normal parameters.

I'm perfectly happy to fly on either, personally. Both are safe enough to me. Just different approaches to automation, that's all.

Neil
pneame on 30 Apr 2012
In reply to Game of Conkers:
It seems that this a classic case of poor user interface - the Airbus, fine plane that it is, is not really designed with the needs of humans for some sort of visual feedback. It seems that there are no indications of what control inputs are being applied. This seems quite bizarre.

But even so, pulling the nose up in a stall seems to be such a beginners error. Especially at high altitude where planes are operating a few knots above stall speed and a few knots below supersonic (which they are not engineered for).

A good article and one can be thankful that the panic of the pilots was unlikely to be transmitted back to the passengers. A tragic accident.
jkarran - on 30 Apr 2012
In reply to lithos:

> interesting reading - why dpont they put the 'side' stick on the other side - ie in the middle between them, where they are visible ?

Even assuming they did it sounds like they'd still have the problem of there being no visible stick input once the 'desired' attitude is selected. I'm quite surprised to hear that is a feature, it must under normal circumstances be actively managed by the flight computers to remain within the safe operating limits of the aircraft (and the available space!). I suppose the problems arise when the sensors begin to deliver garbage to the computers.

Flying with no control position/force feedback and speed/attitude holding controls must take quite some getting used to when converting from something simpler.

There are two very odd things I see in that story: The apparently very strange behavior of the (experienced) pilot Bonin not attempting to lower the nose despite the repeated stall indications and the implication the auto-throttle remained engaged despite the unreliable airspeed information, presumably its primary source of feedback. Very strange and tragic for all involved. Perhaps the full report will shed a little more light when it's released.

jk
Bjartur i Sumarhus on 30 Apr 2012
In reply to Markel: This part from your linked website is also interesting

"As the plane approaches 10,000 feet, Robert tries to take back the controls, and pushes forward on the stick, but the plane is in "dual input" mode, and so the system averages his inputs with those of Bonin, who continues to pull back. The nose remains high.

Any pilots on here explain the benefit of a mode where one pilot can average out the movements of another?
digby - on 30 Apr 2012
In reply to jkarran:

> There are two very odd things I see in that story: The apparently very strange behavior of the (experienced) pilot Bonin not attempting to lower the nose despite the repeated stall indications and the implication the auto-throttle remained engaged despite the unreliable airspeed information, presumably its primary source of feedback.

I remember reading the speculation of experienced pilots on the flying version of UKC before the black boxes were found. At the altitude the plane was flying, in the so called 'coffin corner', the difference between stall and overspeed buffet is very very small. There was no speed information as the pitots had frozen. There were multiple sound warnings occurring. Speculation then was how a pilot would know whether to dive to increase speed to get out of a stall or climb to reduce overspeed, there being no easy way to distinguish between the two. There was also the possibility of the stall warning at one point indicating the opposite of reality, though I can't remember why.
Speculation is as unreliable as hindsight. Newspaper articles like this one become rather specific with hindsight. The thrashing around of speculation must have been an echo of the incomprehension in the cockpit.
Bjartur i Sumarhus on 30 Apr 2012
In reply to digby: Thanks, I just checked wiki for coffin corner and found this. Interesting additional information to the story
http://en.wikipedia.org/wiki/Coffin_corner_(aviation)
Bjartur i Sumarhus on 30 Apr 2012
In reply to Game of Conkers: ".....Some aircraft, such as the Lockheed U-2, routinely operate in the "coffin corner", which demands great skill from their pilots."

Blimey, avoiding detection as well as flying an unstable plane...impressive.
jkarran - on 30 Apr 2012
In reply to Game of Conkers:

> Any pilots on here explain the benefit of a mode where one pilot can average out the movements of another?

What else do you do with the information? If both pilots want to do something different there's always going to be some contention, either electrical or mechanical.

jk
Phase - on 30 Apr 2012
In reply to mypyrex: a further issue is secondary stall, which can happen after stall recovery has taken place.

My thoughts go out to all on board, it must have been a very scary and confusing time on the flight deck.
mypyrex - on 30 Apr 2012
In reply to Game of Conkers: I think those of us who have had some experience of aviation/flying(as pilot) - and most of us seem to - are forever going wonder about the actions of Bonin and his apparent failure to take stall recovery action or the apparent failure of the other two pilots to remedy the situation.

Have to say the nearest I got to flying a big jet was a VC10 simulator.
EeeByGum - on 30 Apr 2012
In reply to butteredfrog:

> I haven't done a great amount of flying, but the couple of stalls I have had demonstrated to me (light aircraft) were quite violent(ish) with very obvious changes of aircraft attitude.

There are lots of examples of aviation crashes that have happened at night as a result of the pilots becoming disorientated, often due to instrument malfunction or even worse - perceived malfunction. Also, large commercial jets fly through the air in a completely different way and the sensations you receive are entirely different to those of light aircraft.
Tim Chappell - on 30 Apr 2012
In reply to EeeByGum:


A slightly off-the-point comment: if the compensation ruins Air France, that's fine by me.

It will concentrate the minds of all the other carriers.
ads.ukclimbing.com
Simon4 - on 30 Apr 2012
In reply to David Martin:

> As for the Boeing/Airbus difference, the US aircraft are prone to their own design quirks ..... No doubt there are many more incidences attributed to both design philosophies.

This was the aspect of the article that appeared to me to be loaded, partisan and irresponsible.

I believe that the Airbus fleet actually has a better safety record the Boeing fleet, though this may well reflect the fact that the Boeing fleet is in general, much older than the Airbus fleet and has a significantly higher proportion of old-technology aircraft, with poorer fuel consumption and other less desirable characteristics (though this probably does not apply to newer Boeing models).

It does seem wrong to try to turn a tragedy that should be investigated as objectively as possible into a point-scoring exercise, as I have seen some American commentators doing.
David Martin - on 30 Apr 2012
In reply to digby:

Perhaps the example of a U2 isn't the best. Commercial flights between 360 and 400 have a sizable window between stall and overspeed. Setting a pre-determined fan speed/exhaust gas temperature and an appropriate pitch has long been an acceptable means for achieving a reliable and safe speed and attitude.

Freezing/wet pitots or static pressure ports can provide all manner of strange speed/altitude outputs though. The interplay between the pilot and co-pilot having airspeed/altitude data potentially sourced from different ports/pitots can make the situation all the more confusing. How either manufacturer sets up the logic for how its instruments interpret these situations can be mind-boggling. So, understandable that this may have caused strange behaviours, but it is exactly the reason why you have pilots in the cockpit and a low hours CPL holder could explain the necessary responses to various instruments in disagreement.

All aircraft are engineered and quirky in their own way. Some are no doubt a human factors nightmare but it is still essential of the pilot to fly the plane as it should be flown, and one pilot at least appears to have spectacularly failed to do so in this case. Total speculation here, but looking at simulator "profiles" for pilots undergoing checks, most seem to throw dramas at pilots on take-off and landing. Its possible that dealing with more mundane matters, such as what occurred at the beginning of this event, is seen as so straight forward that these pilots were simply out of practice.
David Martin - on 30 Apr 2012
In reply to Simon4:

Agreed. It is one thing to speculate on possible causes now, but to turn it in to an insinuation that an entire engineering ethos may be somehow flawed was short-sighted. At least a more balanced analysis of what Airbus has done for the aviation industry would have been god.

Besides, there are increasingly more manufacturers out there than just Boeing and Airbus, so many many schools of thought exist on what could be done better. The alternative isn't necessarily just Boeing.
butteredfrog - on 30 Apr 2012
In reply to Game of Conkers:

A root around the internet gives the following figures

Airbus 330 operating parameters at 30,000ft fully loaded: Min safe speed 230kts, max speed 475kts, stall speed aprox 190kts.
Thats about 300kts between stall and critical mach no, not quite coffin corner.

Cheers Adam
Neil Williams - on 01 May 2012
In reply to Tim Chappell:

While it's right that vast amounts of money are spent on avoiding air crashes, and will be spent on training changes to avoid a repeat of this one, no doubt, air travel is still massively safer than driving to the airport.

I don't think we need to make such comments about an incident ruining an airline, unless that airline has a serious lack of a safety culture, and I see no reason to cast that aspersion at Air France; that sort of criticism is probably validly levelled at some of the smaller African airlines, for instance, but not a major European carrier.

Neil
AdrianC - on 01 May 2012
In reply to Game of Conkers: If you're interested in this stuff the book "FLy by Wire" by William Langeweische is a good read. It is about the Airbus that was landed on the Hudson rive a few years ago after striking a flock of geese and losing all engine power. He argues that the Airbus system was part of the reason for the 100% survival rate and that had that plane been a Boeing the outcome might have been worse.
Neil Williams - on 01 May 2012
In reply to AdrianC:

Cheers, just went on Amazon and bought that, will have a read, no doubt on the plane on the way home!

Neil
mypyrex - on 01 May 2012
In reply to Neil Williams: I'm reminded that, when I was actively flying, there was a British and World aerobatic champion who I occasionally met. His name....



















Neil Williams.

Sadly he was killed in 1977 flying an ex-Condor Legion CASA(Heinkel) bomber from Spain back to the UK
Kemics - on 01 May 2012
In reply to Game of Conkers:

Must have been utterly terrifying and surreal being on the flight deck at the time.

I understand why the plane initially started to fall, but why when the plane reaches the much lower altitude is the full thrust and incline not enough to gain height once more? Is it because at this point the plane has now lost too much speed? What they were saying about the toga procedure, why would this have not worked once they were right down at 2000 ft?

Makes for an eerie read.
Simon4 - on 01 May 2012
In reply to Neil Williams:

> I don't think we need to make such comments about an incident ruining an airline, unless that airline has a serious lack of a safety culture, and I see no reason to cast that aspersion at Air France;

Indeed, I was very surprised to see that they actually had 3 experienced pilots on the aircraft, although this seems to have in practice confused matters. It is no good saying "it was the least experienced pilot flying for most of the incident", as he was still a qualified pilot, with quite a lot of hours (though much less than his colleagues), also how exactly do you get experienced other than ... all pilots have to start with 0 hours in control.

It is true that a single incident can sometimes bring an operator down (unfortunate turn of phrase), but it is not necessarily correct and proper that this should be the case. There are plenty of incidents with boeings and the airlines flying them where either the airline or the aircraft failure caused the crash. I recall a very serious crash of a Japanese 747 that was down to the manufacturer.
Neil Williams - on 01 May 2012
In reply to mypyrex:

Well there you go, I must be famous. However I didn't even exist until 2 years after he was killed, and the only flying I've done is as a commercial passenger (albeit rather frequently at present, twice a week), or indeed (to bring it back on topic) when falling off from above the clip on a climb...

Neil
Neil Williams - on 01 May 2012
In reply to Kemics:

Because essentially the nose was pointing so far up that the aircraft was still stalled despite the thicker air. The way to recover the stall would have been to push the nose down, AIUI, thus changing the angle of attack and gaining useful airspeed.

There's a further, slightly scarier side to this. If I recall, T-tailed aircraft (these are usually smaller ones with body rather than wing-mounted engines) can end up in a deep stall that is unrecoverable (because the wings are blocking the airflow over the elevators). Nasty.

Neil
mypyrex - on 01 May 2012
In reply to Neil Williams: Your namesake was involved in quite an incredible incident in 1970:

http://www.aaib.gov.uk/cms_resources.cfm?file=/15-1971%20G-AWAR.pdf

Now that IS airmanship.
Neil Williams - on 01 May 2012
In reply to mypyrex:

Wow, that's impressive.

Neil
Liam M - on 01 May 2012
In reply to jkarran:
> (In reply to Game of Conkers)
>
> [...]
>
> What else do you do with the information? If both pilots want to do something different there's always going to be some contention, either electrical or mechanical.
>
> jk

Most of the systems documents I used when working on aviation software went for the approach of prioritising a particular signal and following that until reason was found to believe that signal was unreliable, at which point another signal would be chosen to follow. There were very few times when signal averaging was used, and generally that was only when using values calculated from instruments at different locations.
Toby_W on 01 May 2012
In reply to Game of Conkers:

Scary stuff. Reading some of the comments is interesting. It seems from reading the black box dialogue that a mistake was made and then the junior pilot did the opposite of what he was told whilst confirming he was following instructions from the co-pilot. When the captain arrived the identified what was happening but it was too late and he had no options.

So sad.

Toby
jkarran - on 01 May 2012
jkarran - on 01 May 2012
In reply to Liam M:

> Most of the systems documents I used when working on aviation software went for the approach of prioritising a particular signal and following that until reason was found to believe that signal was unreliable, at which point another signal would be chosen to follow. There were very few times when signal averaging was used, and generally that was only when using values calculated from instruments at different locations.

That's fine if you're picking between an array of sensor outputs or calculation results but it's a slightly different situation with human input, both pilots have equal access to the primary flight controls or at least in traditional mechanical system they do.

jk
Liam M - on 01 May 2012
In reply to jkarran: Mechanical systems will obviously differ, but I would have expected a software system would have some mechanism to select between signals if inputs are simultaneously received from both control levers, rather than average them.

The details of how it chooses would probably be complex, but I can't see the value of averaging the signals if you don't need to - maybe a control engineer could explain if there were reasons you may want to.
Neil Williams - on 01 May 2012
In reply to Liam M:

Just an electronic version of the Boeing system, in which the pilots would be able to "fight" each other by trying to counter each other with the controls? The only difference would be that on the Airbus both pilots would be assumed to have precisely the same strength.

Neil
Markel on 01 May 2012
In reply to Neil Williams:

> Just an electronic version of the Boeing system, in which the pilots would be able to "fight" each other by trying to counter each other with the controls? The only difference would be that on the Airbus both pilots would be assumed to have precisely the same strength.


Not quite the same because with the Airbus system the pilots wouldn't know that they were 'fighting' each other. The really sad thing is that the inputs from both pilots were being monitored and recorded; just not displayed on the flight deck. I guess the designers never even considered that the pilots wouldn't be communicating properly.

jkarran - on 01 May 2012
In reply to Liam M:

> Mechanical systems will obviously differ, but I would have expected a software system would have some mechanism to select between signals if inputs are simultaneously received from both control levers, rather than average them.

Why? What is the merit in having a computer choose which pilot to listen to and how would decide? If you're talking about a user input select switch then that is effectively the system you already have, the pilot making control inputs flys the plane. No need for a switch.

> The details of how it chooses would probably be complex, but I can't see the value of averaging the signals if you don't need to - maybe a control engineer could explain if there were reasons you may want to.

You do it because it gives you a system where either pilot can fly the aircraft without having to flip any switches or change any settings. There's nothing to get wrong, no additional steps to take control, if you're making control inputs and your co-pilot isn't you have control. Same as for the mechanically linked systems, control is shared by co-operation, it's simple simple safe and well understood. Your co-pilot can of course 'fight' against your control inputs but that's just the same as for the mechanical system, the only difference is with force feedback, mechanical or electro-mechanical the strongest* wins.

*or the first one to the trim controls

jk
Liam M - on 01 May 2012
In reply to Markel: A big part of design of such systems is trying to avoid information overload. One of the ecu/epmu systems I was working on last year had an extensive procedure for establishing what signals would be recorded for post-flight analysis and what was critical to send to the cockpit given the status of other signals.

The inputs may have been displayed somewhere, or may have been relegated due to other information for the displays. I can imagine being a systems engineer working on cockpit displays is a nightmare role in some ways, trying to establish what information is likely to be needed when.
Markel on 01 May 2012
In reply to jkarran:

> You do it because it gives you a system where either pilot can fly the aircraft without having to flip any switches or change any settings.

But it isn't the only system that let's you do that...

Based on the assumption that there should never be a situation in an electronically controlled aircraft where both pilots need to use the controls at the same time, you could have dual controls with one set dominant to the other. That would allow either pilot to fly with no settings change.

Markel on 01 May 2012
In reply to Liam M:
> (In reply to Markel) A big part of design of such systems is trying to avoid information overload. One of the ecu/epmu systems I was working on last year had an extensive procedure for establishing what signals would be recorded for post-flight analysis and what was critical to send to the cockpit given the status of other signals.


I can see that, but there are other ways that information could be presented (e.g. force-feedback to the controls) that would avoid that problem, or is it more complicated than that?
jkarran - on 01 May 2012
In reply to Markel:

> Based on the assumption that there should never be a situation in an electronically controlled aircraft where both pilots need to use the controls at the same time, you could have dual controls with one set dominant to the other. That would allow either pilot to fly with no settings change.

How would it decide which is dominant and why would you want it to?
jk
Markel on 01 May 2012
In reply to jkarran:

> How would it decide which is dominant and why would you want it to?
> jk

It could be preset to the captain's controls, or set via a central switch with visible indicators on the controls. As to why you would want it: read the article above!

That said, I imagine that a far better system would be for the two sets of controls to be locked together so that there was no ambiguity.
elliptic on 01 May 2012
In reply to Markel:

> you could have dual controls with one set dominant to the other.

...and if the dominant stick is in its neutral, centred position how does the system decide if it's actually being held there..?

Markel on 01 May 2012
In reply to elliptic:

> ...and if the dominant stick is in its neutral, centred position how does the system decide if it's actually being held there..?

By setting input thresholds.

EeeByGum - on 01 May 2012
In reply to Markel:
> I guess the designers never even considered that the pilots wouldn't be communicating properly.

This sadly brings to mind this famous image detailing the requirements and software development process that happens in the real world and not that of software engineering text books.

http://cfc.kizzx2.com/wp-content/uploads/2009/06/68408_f520.jpg
elliptic on 01 May 2012
In reply to Markel:
> (In reply to elliptic)
>
> [...]
>
> By setting input thresholds.

Input thresholds on what?

Not on deflection, because the pilot holding the dominant stick might *want* it in its neutral position...

jkarran - on 01 May 2012
In reply to Markel:

> It could be preset to the captain's controls, or set via a central switch with visible indicators on the controls. As to why you would want it: read the article above!

You've lost me. Why would a physical switch be needed? It already exists, if one pilot or the other wants to take control all they have to do is say so... "I have control".

jk
Markel on 01 May 2012
In reply to elliptic:

> Not on deflection, because the pilot holding the dominant stick might *want* it in its neutral position...


He might, and I would agree that it would only be a marginal improvement on the current situation. However, it would still allow the captain to see whether the aircraft responded to positive inputs which would have helped in the current case.

As I said, it would be far better if the control positions were linked. Perhaps this could still be done even if there were a dominant set of controls.
Markel on 01 May 2012
In reply to jkarran:

> You've lost me. Why would a physical switch be needed? It already exists, if one pilot or the other wants to take control all they have to do is say so... "I have control".

If, for example, the captain was unconscious and jamming the controls.

Yes, they should have been able to decide between them who was flying. However this case shows that under extreme stress, this may not happen!


jkarran - on 01 May 2012
In reply to Markel:

> If, for example, the captain was unconscious and jamming the controls.

That's a problem with pretty much any shared control system. It does not appear to have been the cause of the AF accident nor any other airline accident that I can recall.

> Yes, they should have been able to decide between them who was flying. However this case shows that under extreme stress, this may not happen!

If you haven't got the conviction to say "I have control" then are you really going to flip a switch to turn off your co-pilot's controls? I don't see it.
jk
Markel on 01 May 2012
In reply to jkarran:

> If you haven't got the conviction to say "I have control" then are you really going to flip a switch to turn off your co-pilot's controls? I don't see it.

It may have helped, it may not, in this case it was the actions of the last experienced pilot that caused the disaster so if the left hand controls had over-ridden his, then it probably would not have happened. It certainly couldn't have hindered...

Anyway, that is not the main point. If the controls had been set up so that the inputs of one pilot could not be hidden from the other pilot then this would not have happened. I would be interested to know if there is a real design barrier to getting round this, if there are other considerations to be made, or if it was simply an oversight.

Simon4 - on 01 May 2012
In reply to Game of Conkers: An interesting discussion, both informed and measured.

It is clear that nobody buys the simplistic implication in the thread title, but it is still chilling as an account. For me the most distubing part of the more detailed description listed above is where it says (after the plane had descended somewhat, due to the erroneously induced stall :

"Another of the pitot tubes begins to function once more. The cockpit's avionics are now all functioning normally. The flight crew has all the information that they need to fly safely, and all the systems are fully functional. The problems that occur from this point forward are entirely due to human error."

The reference to going around as from an aborted landing, at 37,000 feet, is quite weird.
Markel on 01 May 2012
In reply to jkarran:

> You've lost me. Why would a physical switch be needed? It already exists, if one pilot or the other wants to take control all they have to do is say so... "I have control".


Apparently (from wiki) airbus already incorporate a 'priority switch' into the controls so that one pilot can do exactly this. Obviously not quite the same as it requires an active intervention before either set of controls are dominant.

Would still be interested to know if there is any argument for having unsynchronised controls though.

David Martin - on 01 May 2012
In reply to Markel:

> Anyway, that is not the main point. If the controls had been set up so that the inputs of one pilot could not be hidden from the other pilot then this would not have happened.

If the pilot had correctly responded to the situation this would not have happened. No matter how many safeties you build in, there is little you can do if the wrong action is chosen - all the stick shakers, force feedback, or control isolation in the world would probably not have helped. Plenty, probably most, accidents occur where it is abundantly clear the wrong path of action is being undertaken but panic, tunnel vision, fatigue, or misunderstanding doom crews all the same. One pilot clearly believed pulling back on the stick while clearly in a stall was the solution.

Besides, no doubt there could be a circumstance where input (even opposite inputs) may be required from both pilots and where the locking out of one crew member could result in a disaster. What would the Telegraph article then say about Airbus design?

I suspect the Airbus system is by design and probably considered a sensible trade-off. Sidesticks are apparently comfortable, allow clearer viewing of instruments and make eating your meal a lot easier.

Bottom line is, if one pilot is pulling and the other pushing then your problems are much bigger than the design of the stick; someone has fundamentally misunderstood which direction and speed the aircraft is headed.
John Rushby - on 01 May 2012
In reply to Game of Conkers:

some interesting comments and experience on the CiF section, and a rather Francophobic troll.......
Markel on 01 May 2012
In reply to David Martin:

> Besides, no doubt there could be a circumstance where input (even opposite inputs) may be required from both pilots....

I'm struggling to imagine. Given that the combined input from two pilots (in fly-by-wire) cannot exceed what is achievable by one pilot. Indeed, it seems the controls already allow one pilot to be 'locked out'. However, even accepting that, I can't see any advantage to controls that move independently. Not that this means there are none, just that I would be interested to hear them.


Obviously the controls didn't cause the accident, but surely it is reasonable for designs to mitigate the effects of pilot error.

Markel on 01 May 2012
In reply to Markel:

It seems that Boeing fly by wire aircraft (e.g. 777) have controls that move together. That means that it was either a design decision to have them move independently in the airbus or an oversight.
David Martin - on 01 May 2012
In reply to Markel:

It is a widespread design ethos. Airbus, and most other maufacturers, use throttles which sit in detents while actual delivered thrust is set by the FADEC (hence the mention of TOGA thrust also). Not quite the same as separate control columns/sticks but the disconnect between aircraft control positions and actual control inputs have been present in aircraft for decades. Pilots do not expect tactile feedback in the same was as previous generations and in instrument flight conditions this is actually beneficial; for 99% of the time pilots heads should be in the cockpit referring to instruments and commanding heading, speed and altitude changes through autopilot inputs.

I see what you are saying by having as many cues as possible providing information, but I'm skeptical as to whether a mechanical control stick would have made any difference in this case given how the pilot(s) had so clearly lost awareness of their speed, attitude and altitude.
Markel on 01 May 2012
In reply to David Martin:

> I see what you are saying by having as many cues as possible providing information, but I'm skeptical as to whether a mechanical control stick would have made any difference in this case given how the pilot(s) had so clearly lost awareness of their speed, attitude and altitude.

Perhaps, I guess it's impossible to say. It seemed to me that it would have made a difference because the other two pilots realised instantly what was wrong when the third pilot told them what he had been doing. Given that we know that one of the other pilots had tried manipulating the controls it's hard to see how it wouldn't have affected his understanding of the problem.

It isn't so much getting feedback from the aircraft; just knowing instantly what the control inputs to the flight surfaces are as soon as you touch the stick. Reading through the airline forums, there are many commercial pilots who think its a stupid system too (and, admittedly, some who think it would have made no difference).

Obviously its a bit of a side show to the real causes of the crash, but I do think that it an interesting design/ interface aspect that could maybe be done better.

Daysleeper - on 01 May 2012
In reply to butteredfrog:
> (In reply to Game of Conkers)
> Thats about 300kts between stall and critical mach no, not quite coffin corner.

Unfortunately it doesn't work like that. Those figures have not been converted for altitude.

I don't fly the A330 but a quick look at the airbus operating manual says... The Max permitted Calibrated Air Speed is 330kts at 30,000 ft. This reduces linearly above that to be 255 KCAS at 41,000 ft. So assuming you are correct about the 230 kts min safe bit (and it sound about the right ballpark) you can see that at 41,000 ft (max cruise altitude) there is a 25 kt difference between min safe and max safe speeds. That's the corner of the flight envelope (referred to as coffin corner as, on older types, the handling qualities outside of that 25 kt window would be pretty horrid. At 30,000 ft it's a bit wider.


jkarran - on 01 May 2012
In reply to Markel:

> I'm struggling to imagine. Given that the combined input from two pilots (in fly-by-wire) cannot exceed what is achievable by one pilot. Indeed, it seems the controls already allow one pilot to be 'locked out'. However, even accepting that, I can't see any advantage to controls that move independently. Not that this means there are none, just that I would be interested to hear them.

Any selectable control isolation will be intended not to isolate a faulty pilot, rather to isolate a malfunctioning control.

You link them not because you need to sum the inputs to overcome a control surface reaction force (as you might with an older mechanically linked, unassisted system under extreme conditions) but because both pilots need to be able to make control inputs at all times. No matter how they are linked there is always the possibility of contention.

I suppose the physical/mechanical isolation of the controls exists for a number of reasons, installation simplicity being one of them, total redundancy being other big one. A mechanical link would require space and could bind, if it adds nothing the pilots and engineers deem necessary then it's probably reducing safety, not adding it. This crash may of course shift that balance of that judgement, it may not. Electromechanical linking of the sticks adds a lot of complexity and a little extra potential for trouble. I suspect artificial force feedback (surfaces-stick) has be omitted as it would make little sense in the context of these controls. They appear to be (I could be wrong here) rate of change of attitude demand controls rather than 1:1 mapped position:position to any flying surface. The control force required to maintain a given control surface position is is not changed once the stick is centred in fact it will vary as the plane interacts with its environment and other parameters (airspeed for example) change. Since this artificial and somewhat meaningless feedback is apparently deemed unnecessary the sticks are built without actuators to couple in the other pilot's inputs, something that was again presumably deemed unnecessary.

I could be way off the mark, I don't know much about modern jets, I only have a couple of years undergrad flight control and recreational gliding to fall back on but the lack of physical link and attitude-change-rate controls while totally unfamiliar don't seem unreasonable.

jk
Markel on 01 May 2012
In reply to jkarran:
OK, thanks for that. I guess there are all sorts of complexities that you miss if you aren't considering the bigger picture.
birdie num num - on 01 May 2012
In reply to Game of Conkers:
The throttle lever on the Airbus A330 is exactly the same as the automatic gear change lever on the Nissan Micra 1.0Ltr Auto. Folks in the civil aviation authority investigating this crash may not have spotted this similarity and the confusion that may have resulted if a sleepy pilot had thought he was driving his car.
Gordon Stainforth - on 01 May 2012
In reply to Game of Conkers:

I'm hoping that soon someone brilliant (with the technical know-how) is going to come along and summarise this interesting thread for the benefit of mere mortals/ordinary airline passengers.
birdie num num - on 01 May 2012
In reply to Gordon Stainforth:
I just did
Neil Williams - on 02 May 2012
In reply to birdie num num:

You jest, but it is true that people die in plane crashes because they try to release their seat belt in the manner of a car one and thus can't get out of their seat quickly enough. I can believe this because I have done it (not in a crash), feeling a bit of a muppet when I realised what I was doing.

Neil
Martin W on 02 May 2012
In reply to Markel:

> The really sad thing is that the inputs from both pilots were being monitored and recorded

I thought, from the description of what happened, that the junior pilot had control most of the time. From the Popular Mechanics transcript, look at 02:11:37: Left seat taking control! ...Bonin soon after takes back the controls, and 02:13:43: Robert takes control and finally lowers the nose

I don't read that as saying that the pilots both trying to make control inputs at the same time.

> just not displayed on the flight deck.

Don't they have an attitude indicator on these things? That would show pretty clearly if the plane was strongly nose-up - and isn't it more important to know what the plane is actually doing, rather than what the guy at the controls thinks he's trying to do? Mind you, they seem to have managed to ignore the plane telling them that it was stalling, so who knows what other helpful information they chose to disregard.

I'm no pilot but even I know that continuing to pull back on the stick when the plane itself is telling you that it's stalling is not likely to be a good way to give your passengers a comfortable flight.
butteredfrog - on 02 May 2012
In reply to Daysleeper:

Will check on that, a friend is a 767 pilot. FWW i can't see a commercial airliner operating within the margins you are suggesting. The only time I have read about coffin corner, was in reference to the U2 spyplane - a different beast altogether.

Link to the A330 manual is quite interesting:

http://www.smartcockpit.com/data/pdfs/plane/airbus/A340/misc/A330-A340_Flight_Crew_Training_Manual.p...

TOGA is a reference to the maximum thrust setting for the engines as opposed to the pilot thinking he was in a take off situation.

Cheers Adam
Bjartur i Sumarhus on 02 May 2012
In reply to Martin W: "I'm no pilot but even I know that continuing to pull back on the stick when the plane itself is telling you that it's stalling is not likely to be a good way to give your passengers a comfortable flight. "

I believe the general consensus is that the passengers would have had very little knowledge of their imminent deaths. If the plane has very little forward velocity then pulling the stick back will not make it go vertical and stay there (although I could be wrong but is my understanding)
Markel on 02 May 2012
In reply to Martin W:

> I don't read that as saying that the pilots both trying to make control inputs at the same time.


The BEA interim report has an appendix with graphs of the inputs from each pilot. You can clearly see that at one or two points there are overlapping and opposite inputs from each pilot.
Martin W on 02 May 2012
In reply to Game of Conkers:
> (In reply to Martin W) "I'm no pilot but even I know that continuing to pull back on the stick when the plane itself is telling you that it's stalling is not likely to be a good way to give your passengers a comfortable flight. "
>
> I believe the general consensus is that the passengers would have had very little knowledge of their imminent deaths.

You believe the actual dying bit would have been comfortable? Jeez, talk about pedantic.
Bjartur i Sumarhus on 02 May 2012
In reply to Martin W: ?? sorry, you have lost me there
Neil Williams - on 02 May 2012
In reply to Game of Conkers:

I guess he means it'd have been over quickly. It would be far worse to know for 10 minutes you're going to die painfully, then die painfully, than for it to just creep up on you and be over quickly.

Neil
Bjartur i Sumarhus on 02 May 2012
In reply to Neil Williams: of course. I was just responding to the comfortable flight part. By all accounts, the passengers would have had no knowledge that the plane was out of control, which is a blessing.

Making that point seems to have annoyed Martin ,
Daysleeper - on 03 May 2012
In reply to butteredfrog: Coffin corner is a bit of a misnomer which has survived from the early jets. An A330 or 767 or any "modern" (i.e. anything you can buy a ticket on these days) aircraft has the worst of the handling qualities designed out of it. But somewhere you have to say "this is the limit" and then you put a (wide) safety margin on it.

For example minimum safe speed will be about 1.3 times the stalling speed. Maximum safe speed will be set at a level which has a good margin till the point where the aircraft sustains damage. You draw big red lines at these speeds and put alerting systems on them to wake up the pilots and get them to do something. In addition there is quite a long way beyond the red line before anything bad actually happens (apart from a considerable amount of paperwork and a tea no biscuits interview with the chief pilot).

On paper (and on the air speed indicator) the margin can be quite tight but in day to day normal operations it generally is not. For instance you wouldn't operate a max weight A330 at 41,000 ft it would be horribly inefficient.
David Martin - on 03 May 2012
In reply to Daysleeper:

Weights are an important consideration when talking "coffin corner" too. Up at 41,000 at high weights the gap between high and low speed buffet boundaries is not great - this appearing near the top of the A330 operating envelope. This incident seems to have occurred about a 1/3 of the way in to the flight though so maybe the aircraft was a fair way from the limits.

Working from the right of page 8 of http://www.smartcockpit.com/data/pdfs/plane/airbus/A330/misc/A_0_Limitations.pdf in a zero G situation, at 200,00kg and the ref line for CG, at Fl 410, the buffet boundaries go from about .685M to Mmo. But if you factor in G for an upset, the buffet margins reduce quite quickly.

Without a cruise graph, itís hard to tell what weights corresponded with what altitudes, so hard to get the split. If that minimum safe speed is correct, there would also be a buffer between it and buffet onset, so 25kts is probably a bit small.

Probably safe to say major upsets don't generally occur at this sort of altitude (a few recent A380 incidents excluded) so while the margin sounds thin, in reality it isn't. Even with a few hundred kts of margin, drawing the wrong conclusions about attitude and airspeed will still result in dramas.
Neil Williams - on 03 May 2012
In reply to AdrianC:

Been reading that book the past few days. A seriously interesting read that has made me more impressed with the Airbus control concept than I was before.

I will feel all the safer on the EZY A319 I'll be plonking myself on in about 8 hours' time.

Neil
Rigid Raider - on 05 May 2012
I've been following this story with especial interest as I use AF most of the time and I will be in an AF 'plane on my way to Lagos in just over 24 hours.

The fact is that the crew were in what's colloquially known as a "blue funk". They were so overwhelmed with conflicting information, which the more experienced captain could have sorted out had he realised what the inexperienced co-pilot was doing, quite unaccountably and illogically. As others have pointed out, in a state of panic the human mind shuts down its higher intellectual functions and reverts to the most basic learned behaviour, which is when you have to hope that the training has been sufficient and the right actions are taken. If you understand French and you read the transcript you'll see that the crew had reverted to schoolboy levels of language, actually slang, in those last couple of minutes.

If any good can be said to have come from it; at least we can hope the same thing won't happen again. Fingers crossed for tomorrow.....
MG - on 05 May 2012
Curious about the assumption that had the captain been there sooner things might have been different. Is it likely that his greater experience would have been relevant? Presumably most pilots would never in an entire career need to deal with this sort of situation. So would more flying hours help, or was simply a calm assessment of the situation needed, with the ability to provide one down to personality rather than flying experience?

This topic has been archived, and won't accept reply postings.