PPRuNe Forums

PPRuNe Forums (https://www.pprune.org/)
-   Tech Log (https://www.pprune.org/tech-log-15/)
-   -   AF447 final crew conversation - Thread No. 1 (https://www.pprune.org/tech-log/466259-af447-final-crew-conversation-thread-no-1-a.html)

J.O. 30th Dec 2011 22:04

Dick Rutan? Really??? :rolleyes:

overthewing 30th Dec 2011 23:53

As humble SLF, I do know when a plane is descending, because my ears react with distressing sensitivity. I could probably tell you how high we are by how deaf I am, and how much swallowing / yawning I'm having to do. Even small changes in altitude affect me. I often wonder if professional flight crew are less bothered than I am? I can't imagine dropping 37,000 ft + without being well aware of it.

airtren 31st Dec 2011 00:51


Originally Posted by ChristiaanJ (Post 6928253)
I knew some of the people who designed the "Airbus computer system algorithm(s)" (I worked in the next office...). They were not 'computer nerds', I can assure you. They knew what they were doing.
Whether the original design specifications, thought up by 'system engineers' (not 'computer nerds' as such), with totally new 'control laws', were equally well thought out, is a question I'm still asking myself to this day.

Glad to see a post of yours I can resonate to.

The computer algorithm(s), while may be done by very thorough/experienced professionals, and while being full proof, and allowing some elegant way of doing things, ultimately they need to follow entirely, and strictly the system architecture/design.

And it's that system architecture/design in which some of the shortcomings/deficiencies pointed out on these threads show designer/architect weakness/immaturity.


There IS too much confusion here between system design and it's physical implementation, which can be mechanical or electro-hydraulic-mechanical (as in the very olden days), with analog computers (as in 'my' Concorde days... but there were many other similar systems, from Trident to early 747 to VC10 to the F-104... I'll leave it to the oldies among us to expand the list), or with digital computers (as in the A320 for a start, and most present-day A&B flight control systems).
True, but that's true in general, for those that are off their domain of expertise.


The fact that the switch from analog to digital happened at the same time as the switch from 'steam gauges' to 'glass' tends to confuse the issue even further.
Digital gauges can have a maximum effectiveness in conveying information. Unfortunately, from what can be concluded, and as a personal opinion, it is NOT the case with the A330 ones.


Given the challenge, the A320 flight control system could probably have been implemented as an analog system, but it would have been heavier, more expensive, and more difficult to flight test and optimise. "Been there, done that"....
While I agree 100%, with what I would call a lot less flexibility, which has drawn ultimately the much more successful path of the use of digital computers/electronics, I think that analog systems, don't have to be a lot more expensive, or a lot heavier, as long as the used components would be of same generation technology, and of same type of mass production, i.e. similarly to the special digital components used in aviation - analog computers allow modular designs, with modules usable on a wide range of frames too. Certainly it would mean a completely different ratio of hardware to/versus software work.

pattern_is_full 31st Dec 2011 01:41

Just to avoid confusion:

BURT Rutan is an innovative aircraft designer and engineer....

DICK Rutan is his brother. Primarily a pilot, not a designer (although he is a partner in his brother's design business - Scaled Composites LLC). Flew his brother's design (Voyager) for the record unrefueled non-stop global circumnavigation in 1986.

jcjeant 31st Dec 2011 02:11


And it's that system architecture/design in which some of the shortcomings/deficiencies pointed out on these threads show designer/architect weakness/immaturity.
If this is true .. how many of them will be recognized as responsible for the deaths of 228 people and how many of them will end their life in jail ?

iceman50 31st Dec 2011 03:02

It is funny how some are now trying to BLAME the designers and engineers. They were not and I use the word sadly, FLYING the aircraft. We as PILOTS bear the responsibility of flying our passengers or "packages" safely from A to B. This accident appears to stem from a basic "pilot" failure to control the A/C following the loss of airspeed indications resulting in a STALL. A 15 degree pilot induced pitch up at 35,000' close to the rec max altitude will give nothing else in what was now a "normal" A/C in ALT LAW, especially when the input is held and STALL warnings ignored!

Bubbers most of the previously stalled airliners, if not ALL, were from YOKE aircraft so give it a break on the sidesticks. The Yokes did not SAVE the others even though the other pilot "KNEW" what the other was doing.

Gretchenfrage 31st Dec 2011 06:24

I simply have to say it again:

It is a puerile reflex to defend any design shortcoming by stating that the "other" design had accidents too!

Grow up, please!

Sure enough you can stall a yoke as much as a stick. Once more: It is not about stick or yoke, not about conventional or fbw.

It is about enabling the pilot at commands still be the commander.

If you deprive him of primary sensory systems, then the design is worse than the one that still has those.
If you deprive the last resort (pilot) of FULL control by sometimes not processing his commands, then the design is worse than the one that still allows this.

Both designs may be fbw and stick-driven, that is NOT the issue. Both may work almost the same way, have the same protections, even programming commonalities.

But the difference I pointed out, sensory feedback and ultimate control, is the flaw in the Airbus design.

It would take very little to implement this:

1. A simple rumble-stick from the gameshop around the corner has primitive feedback, weighs almost nothing more, and is simple to install.

2. A simple escape-button, that functions analogue to the complicated procedure of switching off 2 prime and one sec, thus giving instant direct law, would be just as simple and cheap enough to install.

With these both features, the Airbus would be a very nice plane to operate.

iceman50 31st Dec 2011 07:01

Gretchenfrage

Can I assume your last post was directed at me as you have not addressed it?

If so, I think you are now being childish as my comments about sidesticks were directed at Bubbers who keeps insisting that if it had been a "real" airliner with a yoke instead of a sidestick it would not have happened.

As for the design being wrong that is your opinion and you are entitled to it, what would be interesting to know is your experience/knowledge in making this statement. Having flown A330 and A340 for the last 16 years, I do not see it as a design problem. We do not need stick shakers nor pushers in Normal law and in ALTN law the A/C was shouting at the AF crew that it was stalled. That is until it was held into a regime that no airliner was expected to be put in by a competent pilot. Perhaps you should speak to the designers of all the new military A/C as most cannot be flown without the use of technology and the "pilot" only makes requests through a computer.

As for the "simple" red button that gives control back by switching off some Prims and Sec's, way too complicated and dangerous. How would it work with the MEL and if the pilot was getting totally confused and thinks by removing the protections he will be safer, stand by for some major accidents / incidents. As pilots knowing our A/C, no matter what manufacturer, is the only option and continuing the learning process not just sitting back accepting the paycheck each month.

Dream Land 31st Dec 2011 07:28

2 cents
 

1. A simple rumble-stick from the gameshop around the corner has primitive feedback, weighs almost nothing more, and is simple to install.

2. A simple escape-button, that functions analogue to the complicated procedure of switching off 2 prime and one sec, thus giving instant direct law, would be just as simple and cheap enough to install
What about a gigantic parachute system when all your feed back devices fail to save the day? :ugh:

All the feed back in the world didn't seem to save several other accidents with similar circumstances, I know, I should grow up by pointing this out, sorry. :}

Many of the totally preventable accidents lately (and several old ones), had to do with bad pitot information displayed, why not design a system to deal with suspect airspeed information, maybe some automatic system that changes over to a synthesized indication, just an idea, because the feedback systems in place now seem to be overruling the obvious attitude information which pilots don't seem to pay attention to anymore in modern push button flight decks.

Warnings, shakers and alarms going off just complicated the situations.

cats_five 31st Dec 2011 07:32


Originally Posted by overthewing (Post 6928394)
As humble SLF, I do know when a plane is descending, because my ears react with distressing sensitivity. I could probably tell you how high we are by how deaf I am, and how much swallowing / yawning I'm having to do. Even small changes in altitude affect me. I often wonder if professional flight crew are less bothered than I am? I can't imagine dropping 37,000 ft + without being well aware of it.

No, you know it's cabin pressure is changing.

Typical pressure is equivalent to an altitude of 8,000' and for most of it's descent AF447 was above that level, so cabin pressure should have been constant down to that altitude.

Of course once it started increasing cabin pressure the occupants may well have felt the pressure changes. However whatever caused the problems happened way, way above 8,000' and pressure changes would not have been a clue.

The flight crew were well aware there was a problem but they failed to realise they were stalled. They should have realised they were dropping like a stone - the variometer would have told them that - but I suspect the vertical acceleration wasn't that obvious.

A33Zab 31st Dec 2011 08:14

@Gretchenfrage
 

But the difference I pointed out, sensory feedback and ultimate control, is the flaw in the Airbus design.
@dreamland: :ok:

Considering the several reports published this year,
IMO there's one common weak link (flaw) in ALL modern A/C design.......and it is called ADIRU!

f.i. the stick shaker design of 777 may look more robust(with WOW switching)
but if ADIRU quits sensing data (777 if CAS < 30 Kts) the shaker will 'silence' and flags (PFD) will show up too.

What remain is how this is - computed - and handeld by you Professional Pilots!

Clandestino 31st Dec 2011 14:08


Originally Posted by BOAC
I have only seen the AAIB report 6/2001, and would be grateful if you could elaborate on the 'massive overreaction' which according to that report was to close the throttles and disconnect the autothrust, which is pretty much what I would have expected? Perhaps a quote from your report copy would help?

Sure, its not an easy read for someone not acquainted with Airbus peculiarities.


Originally Posted by AAIB
Five seconds after the autopilot disengaged, the thrust levers were closed and then the autothrust was disconnected, probably by the handling pilot in an effort to prevent another overspeed condition.

As thrust levers on FBW Airbi set the thrust limit, closing them commands idle thrust, which is also what you get when you override autothrottle on B and D brands when pulling to idle. Nowadays putting TLs into idle detent also disconnects ATHR but I cannot comment whether this feature was installed at the time of the occurrence. Nevertheless, result was engines trying to spool down to flight idle. Only thing stopping them was activation of alpha floor - you can think of it as automatic low speed protection via autothrust.

Now, I would not expect 0.02 overmach in level flight to be dealt with by closing the TLs shut and keeping them there until high alpha protection kicks in and then watching and doing nothing for 17 seconds as aeroplane pitches up but that's just me.


Originally Posted by GretchenFrage
My quote was not attached to any race or region.

So I see. I am puzzled why in the Welt would you feel my comment to be written in response to your post. :confused: Think of it as pre-emptive non-agression and not reaction. :E


Originally Posted by Captainplaystation
Clandestino, I was thinking principally of the similar AF A340 "whoopsy" last Summer, where they ended up within (if I remember) about 5 -7(?) kts of the KIAS stalling speed

I am glad to read that, as it strongly indicates you have information regarding the aforementioned AF471 incident that confirm there was:


Originally Posted by captainplaystation
instrument failure/false overspeed indications & uncommanded climb by the "wonder-plane", to counter this

I would appreciate if you share your info with the rest of us. Thank you.


Originally Posted by captplaystation
when I read things like this I maintain an open mind

Very well sir, your open-mindedness will certainly help you appreciate that due to clogging of all three pitots, AF447 slipped into ALT2 law where there is only G protection left. No high sped prot. No alpha prot. No bank prot. Nothing to pitch you up except hitting the -1G, which was never approached.


Originally Posted by Mac the Knife
Would not the attitude indicator(s) have been showing an unusually NU attitude?

Well...yes. They would show attitude not normally met while flying at cruise altitudes, yet they were far from extreme, maximum being +17.9 ANU just before the aeroplane stalled. If the CM2 believed that aeroplane's behaviour and performance were the same at MSL and at FL330, then we are deeply and truly effed.


Originally Posted by GretchenFrage
I simply have to say it again:

It is a puerile reflex to defend any design shortcoming by stating that the "other" design had accidents too!

Grow up, please!

1. The same design has suffered at least 36 times from the failure so similar to the one affecting AF447 it can be considered the same. All other 36 flights continued and landed uneventfully. Therefore, design of the airbus flight controls does not imply accident is inevitable or even probable when all reliable airspeed data is lost.

2. Tu-154, B727 and Q400 were lost when their crews pulled into stall despite the ample warnings. All of them had interconnected yokes installed. Therefore, yokes do not provide absolute protection.

3. Wolfgang Langewiesche in 1944. wrote about pilots that did not understand that pitch flight control does not actually control pitch but rather AoA through elevator and kept pulling into stall until the ground put them out of their misery. Therefore, behaviour of CM2 is not unprecedented and similar cases can be traced as far back as there were three-axis fixed wings.

Care to revise that little statement of yours?


Originally Posted by bubber44
The Airbus computer system algorithm is designed by a bunch of computer nerds who have no understanding of aviation.

Fact that there are thousands of flights safely performed by FBW Airbi every day for last two decades would then to be ascribed to pure luck, eh? :hmm: Those who lack capacity, will or both to understand Airbus control system can do far better than ignorantly insult everyone who put their effort into creating pretty fine and functional system.


Originally Posted by A33Zab
IMO there's one common weak link (flaw) in ALL modern A/C design.......and it is called ADIRU!

Airspeed indicators were failing ever since there were first ASIs and ADIRU can not overcome that. Pilots unprepared to deal with their demise in safe manner have no business in aviation, from microlights to 380s.

OK465 31st Dec 2011 14:27


Nothing to pitch you up except hitting the -1G,
'Hitting' -1G will not 'pitch you up'.

The FBW flight control surfaces may move in an ANU direction to prevent exceedance of -1G, but the aircraft will continue to pitch down at -1G.

BOAC 31st Dec 2011 14:46

Clandestino - (TC/AC incident). It is still my opinion that your condemnation of a 'massive overreaction' is unwarranted. I agree subsequent actions were tardy, but faced with a significant overspeed in turbulence and a TCAS 'Descend' RA I think closing the T/Ls and disconnecting the A/T were spot on. Then with no indication whatsoever of the invocation of some bizarre AoA law that is going to cause the a/c to rear up on its tail and climb steeply, I am not really surprised that there were some 'what the ** is it doing now?' moments. I assume from your post that you would have been straight 'on the case' and sorted it out immediately. Bravo.

May the good Lord protect all of the average pilots from such software anomalies.

Clandestino 31st Dec 2011 15:37

OK465, thank you for correcting me, my pedantry selector is really set too low. :ok: Now we know there's absolutely nothing in FCS software that would pitch you up in ALTN2.

BOAC, the crew was faced with TCAS "Descend" and did not command pitch down. What do you think is expected pilot reaction when faced with TCAS RA? Perhaps:

Originally Posted by AAIB
For 18 seconds after the autopilot disengaged the aircraft remained within 200 feet altitude of FL 360 but once AoA law was invoked at 14:21:50 hrs, the aircraft’s attitude began to pitch nose-up. The pitchup
trend continued for 17 seconds reaching a peak of 15° nose-up shortly before the first nose-down sidestick command was applied.

..or not quite. "Descend, descend, descend" was issued at 14:21:41, first ND input was recorded at 14:22:07, mere second before TCAS gave up and degraded RA to TA as 340 was now above 330 and climbing.


Originally Posted by BOAC
the invocation of some bizarre AoA law that is going to cause the a/c to rear up on its tail and climb steeply,

Saying that alpha prot cause aeroplane to rear up on its tail is quite caricatural. Alpha prot's function must be known to every FBW Airbus pilot all the time, every time. As for labeling alpha prot "bizzare", this post refers.


Originally Posted by BOAC
no indication whatsoever

ASI full of low speed cues with speed eventually going below lowest slelectable, N1s rising uncommanded and aeroplane pitching up on her own accord. I'd say it is indication enough but that's just me.


Originally Posted by BOAC
I assume from your post that you would have been straight 'on the case' ad sorted it out immediately. Bravo.

Please don't assume that. Given the turbulence they were bouncing in, it is entirely possible that I would have made even bigger mess than the Turks did. It is ironic that if they left TLs in climb detent, autothrust would eventually have taken care of overspeed on its own. Yet if I made the same (or worse) mistake, eventually refused to accept it was my "to much, to soon, followed by too little, too late" that caused the incident and tried to blame the aeroplane for it, well, it would be criminally irresponsible of me. I'm not implying in any way that the incident crew or their airline did that, this statement is purely hypothetical.

Anyway, following StJohns and Bilbao mess ups, alpha prot has been desensitized. It would take a couple more seconds for it to activate nowadays in the same circumstances.

It's not about blame, it's about accurately analyzing what happened to prevent recurrence but then any aerospace professional knows that, I think. I guess. I hope.

HazelNuts39 31st Dec 2011 15:55


Originally Posted by clandestino
Now, I would not expect 0.02 overmach in level flight to be dealt with by closing the TLs shut and keeping them there until high alpha protection kicks in and then watching and doing nothing for 17 seconds as aeroplane pitches up but that's just me.

The FDR data show that the airspeed did not change significantly in the 5 seconds between closure of the throttles and capture of alpha-prot. (To put this into perspective: the engines take more than 10 seconds to spool down to idle, the A/C then takes over 2 minutes to decelerate to V-alphaprot in level flight). From AAIB Bulletin No. 6/2001:

Five seconds after the autopilot disengaged, the thrust levers were closed and then the autothrust was disconnected, probably by the handling pilot in an effort to prevent another overspeed condition. Ten seconds after the autopilot disengaged, the corrected or phase-advanced angle of attack (a computed parameter which is not recorded but can be calculated by Airbus Industrie from the DFDR data) reached the ‘alpha prot’ value. This angle of attack excursion beyond alpha prot caused a change in the pitch flight control law from normal law (NZ law) to angle of attack protection law (AoA law).
(...)
Changes to the A340’s flightpath caused by the aircraft’s flight control system response to the overspeed warning and autopilot disconnect were negligible until AoA law was triggered. The fact that this law was not triggered until 10 seconds after the autopilot disconnected was a random event driven by the severity of the turbulence.
In the 17 seconds after capturing alpha_prot several attempts were made to reengage the autopilot, before moving the sidestick.

It would take a couple more seconds for it to activate nowadays in the same circumstances.
Are you sure of that?

cessnapete 31st Dec 2011 16:05

AF447
 
Never flown the Airbus, but to put things in perspective.

At all times in this accident the two experienced F/Os had control of the flight surfaces (albeit degraded), two fully functioning engines, a standby horizon and compass, and a selectable groundspeed from the GPS.
The position of the Capt on/off the flightdeck should not be a factor even if incapacitated,thats what co-pilots are for.

Plenty enough information to keep in control of the aircraft if trained properly ????.

BOAC 31st Dec 2011 16:24

[QUOTE=Clandestino "Descend, descend, descend" was issued at 14:21:41[/QUOTE] - you are mixing up your RAs. This one (short term) was caused by the descent of the traffic above, not the later climb of the TC.

The 'desensitisation' of which you speak involves, I believe, an airspeed or mach input? Not much use in 447's case. I still feel uneasy about some underlying similarities between the TC and 447 - turbulence, a speed issue, A/P and A/T disconnect followed by a rapid climb. Yes, I know what BEA say. I still think it significant that without rearward stick movement in AoA law the system maintains alpha prot - the reasoning for this I cannot understand. Do we know whether 'AoA law' was triggered for 447?

thermostat 31st Dec 2011 16:27

Dream Land. Do you mean a "a simple GPS with ground speed read out"???
I mentioned in an earlier post that I started carrying with me a hand held GPS unit on my flights a year before I retired. Had to use an external antenna to get a signal. It was attached to a suction cup which was placed on the side window. It was interesting to see the ground speed within 1 knot of the aircraft computed one. This simple unit would have saved the lives of many pax and crew in the past had they been available and used. It has all you need to safely do an approach, GS, track and true (not pressure) altitude. Long live GPS.

Mr Optimistic 31st Dec 2011 16:44

Not being a pilot in any way shape or form I can only track these arguments with a general engineering interest. However, I think there is a significant issue in that the crew on this flight seemed to have despaired of thinking through the situation to understand and respond to their predicament. Something put a lock on their brains. If you are going to design ever more complex systems which nevertheless can still fail and require effective intervention, such a situation needs to be understood so that the warnings/protections/information presentation can quickly bring the humans up to speed.

Maybe believing that the a/c is a complex system allows the possibility of defeatism, ie that the machine has betrayed you in a way no-one had previously anticipated so we are the unlucky guinea pigs. The a/c not responding to increased thrust or NU commands may just have reinforced that and everyone effectively shrugged and gave up.


All times are GMT. The time now is 23:46.


Copyright © 2018 MH Sub I, LLC dba Internet Brands. All rights reserved. Use of this site indicates your consent to the Terms of Use.