I came across this a discussion of this incident on another forum; I've been trying to find an accident/investigation report on it, but without success; perhaps someone can assist - and maybe answer a few questions?

Most of you will have heard of it; here's a Youtube video, which shows the incident at 2:20:

http://www.youtube.com/watch?v=0wp-Dbb2CO4&feature=player_embedded

Here's a text version:

Following an autopilot-coupled go-around, the pilot attempted to counteract the autopilot's programmed pitch-up by pushing forward on the control column. (In most circumstances pushing on the control column disengages the autopilot but automatic disconnect is inhibited in go-around mode. The autopilot should be disconnected or a mode other than go-around should be engaged through the FCU-Flight Control Unit).
As a result of the control inputs, the autopilot trimmed the stabilizer to 12° (nose up) to maintain the go-around profile, but the elevator was deflected 14° (nose down). After climbing about 600ft (to around 2100ft) the autopilot captured its preselected missed approach altitude and disconnected, as the go around mode was no longer engaged. In the next 30 seconds, the grossly mistrimmed A310 pitched up to 88° and airspeed dropped to less than 30kt. (The stall warning activated then canceled itself as the airspeed fell below usable computed values and the autothrottle system dropped off.) At 4,300ft, the A310 stalled, pitching down to -42° while pilot-applied control inputs showed full up elevator. Airspeed increased to 245kt then the aircraft bottomed out at 1,500ft, pulled + 1.7g, then climbed rapidly.
The second pitch-up reached 70° followed by a stall 50 seconds after the first. The nose dropped to -32° and airspeed rose to 290kt and the aircraft bottomed out at 1,800ft. On the third pitch-up (to 74°), the A310 climbed to 7,000ft then stalled again, about 60 seconds after the second stall. This time airspeed reached 300kt in a -32° nose down attitude before the aircraft leveled off at 3,600ft.
The fourth pitch-up reached 9,000ft but this time the crew's use of thrust and elevator control (and very likely retrimming the stabilizer) prevented a stall and the A310 leveled off at 130kt. Speed then increased accompanied by another milder pitch- up to 11,500ft where control was eventually regained.
All aircraft systems operated in accordance with design specifications. The reaction of ATC (the incident happened at Moscow) or the passengers is not recorded.

The decision by the crew to override the AP on g/a was (as I understand it) also behind the two CAL AB6 losses at NGO and TPE; these were 16 and 14 years ago, respectively and there have been no similar accidents or (reported) incidents since. So, what happened to stop these?
1) Better training on go arounds? (i.e. "whatever you do, DON'T try and interfere with the acft while it's carrying out an automatic go around)
2) Changes to the Airbus flight software on the newer "FBW" Airbus family?

Any comments?

Quoting Kaitak (Thread starter): 2) Changes to the Airbus flight software on the newer "FBW" Airbus family?

Airbus FBW uses a variant of the C* pitch control law originally developed by NASA. That type of control is basically immune to this type of event. The only limiter would be how fast the system can trim out the stabilizer, but that would happen automatically.

Tom.

Holy &*$%. Never heard of that incident. That must be one hell of a situation for everyone involved in there...

Quoting Kaitak (Thread starter): The decision by the crew to override the AP on g/a was (as I understand it) also behind the two CAL AB6 losses at NGO and TPE; these were 16 and 14 years ago, respectively and there have been no similar accidents or (reported) incidents since. So, what happened to stop these?

The series of accidents is described exhaustively in Macarthur Job's "Air Disaster" (vol 3). An activation of TOGA mode and a manual intervention with pitch control at the same time could lead to an out-of-trim situation without the crew realizing what was happening. All A300-310 models with a two-man crew configuration were affected.

A similar case also happened to OH-LAB of KarAir/Finnair in Jan 1989. During approach to HEL at 865 ft the captain inadvertedly activated TOGA at the yoke, and disconnected the autopilot 10 seconds later. During this time the autopilot had already trimmed the tailplane at 8 degrees noseup as the captain had manually resisted the TOGA actions for the convenience of the passengers. The a/c went off glidescope and the captain desided to go around "for real". After the configuration changes, just like in the fatal accidents, full power and full nosedown command were issued, with the nose 24 degrees up. The nose went further up to 35.5 degrees with the speed decaying down to 94 knots. What saved the a/c was that 8 seconds after full power was applied, at the height of 2,254 ft the captain began trimming the a/c nosedown and commanded the f/o continue doing so using the trim wheels, and a recovery took place at 1,539 ft.

Macarthur Job writes: (p 155) "One of the first dash 600s in March 1985, as well as [the case above] ... were the first to record the conflicting inputs and pre-landing upsets ... Airbus Industrie's resulting communications to its customers between 1985 and 1994 included Technical Notes, Service Bulletins, FCOM procedure revisions and cautionary warnings and Operations Engineering Bulletins. The second incident [above] in fact prompted Airbus Industrie to the unusual step of conducting an "Operators' Conference" at Toulouse in May 1990."

The most complete analysis of this failing man-computer interaction is probably by Dr Peter Ladkin: http://www.rvs.uni-bielefeld.de/publ...OCS/ComAndRep/Nagoya/NagoyaWB.html

OH-LAB still going strong 20 years later in 2009:


View Large View Medium

Photo © Reza - Iranian Spotters


[Edited 2010-02-04 05:09:18]

Quoting Kimon (Reply 4): Habsheim Crash:

I don't think we need to discuss any conspiracy theories, the facts themselves are obvious: the crew's preparation for the show flypast was very deficient, they disabled the dynamic envelope protection for a more spectacular show effect, flew too low, too slow with the engines idled too far back, and added GA power much too late for the engines to spool up before the fatal meeting with forest line.

Something similar happened to a TAROM A310 flight on approach to Paris Orly in September 1994:

http://www.skybrary.aero/index.php/A...TAROM,_Paris_France,_1994_(HF_LOC)

Faro

A short PP presentation on the China AIr Nagoya accident, and the final SB from Airbus.
www.myoops.org/cocw/mit/NR/rdonlyres/Aeronautics.../china.ppt

Quoting Kaitak (Thread starter): In most circumstances pushing on the control column disengages the autopilot but automatic disconnect is inhibited in go-around mode

I was under the impression that disconnecting the autopilot by moving the controls was a nice-to-have safety feature but that the recommended method was to use the disconnect switch on the yoke/stick.

Quoting Kaitak (Thread starter): 2) Changes to the Airbus flight software on the newer "FBW" Airbus family?

The article isn't about FBW. It's about automated flight systems on all modern airliners. Some of the incidents mentioned involved non-FBW aircraft.

The Air Inter 320 accident was caused by ambiguity of the displayed information, i.e. FPA vs V/S, not by the actions of the auto-flight systems. I seem to recall that the display was changed to better differentiate between the two.

I suggest a vote be taken amongst us forumistas to settle the safety debate between Airbus vs Boeing.

Quoting kimon (Reply 8): I suggest a vote be taken amongst us forumistas to settle the safety debate between Airbus vs Boeing.

Both can be equally dangerous when operated wrong.

Quoting kimon (Reply 8): I suggest a vote be taken amongst us forumistas to settle the safety debate between Airbus vs Boeing.

The pros and cons of each are discussed here quite regularly. The statistics don't show one being safer then the other and I don't think a vote would settle anything, especially if the reason for a vote doesn't need to be justified. In any case, this article isn't specifically comparing Airbus FBW with the rest - it's considering issues with modern systems from all manufacturers.

I'm pretty sure the problems in fighting the autopilot during a go-around aren't peculiar to Airbus FBW. Looking at the Airbus incidents that were mentioned, the A310 and A300 aren't FBW and the Air Inter 320 accident wasn't caused by the FBW systems.

Quoting AverageUser (Reply 9): Both can be equally dangerous when operated wrong.

I think that's more the point.

Quoting Kaitak (Thread starter): 1) Better training on go arounds? (i.e. "whatever you do, DON'T try and interfere with the acft while it's carrying out an automatic go around) 2) Changes to the Airbus flight software on the newer "FBW" Airbus family?

The difference between the A310 and the Airbus FBW family, is that the A310 is not FBW on the pitch axis. If I understood the FCOMs right, On the Airbus FBW... you can always disengage the A/P if you move the sidestick beyond a certain threshold... even in A/P G/A mode (but why?)... now if large sidestick movement disengagement is inhibited, U're on it for the ride, and let the envelope protection do its job.   

Simple rule of airmanship... when you're not flying... when you want to take control... disconnect the damn autopilot! Moving the yoke or yanking the stick may disconnect it, but isn't it much better to press that disconnect button on the yoke/stick?

Quoting David L (Reply 7): I was under the impression that disconnecting the autopilot by moving the controls was a nice-to-have safety feature but that the recommended method was to use the disconnect switch on the yoke/stick.

Ahhh... yes... nice feature... but as always, using the disconnect button is better, because it means "intentional" disengagement.

Quoting David L (Reply 7): The article isn't about FBW. It's about automated flight systems on all modern airliners. Some of the incidents mentioned involved non-FBW aircraft.

Need me to repeat stating the obvious for the others again there mate?  

Quoting AverageUser (Reply 9): Both can be equally dangerous when operated wrong.

Yes... you can kill yourself in either if you really wanna do it... so what's the point in the vote?   

Quoting mandala499 (Reply 11): On the Airbus FBW... you can always disengage the A/P if you move the sidestick beyond a certain threshold... even in A/P G/A mode (but why?).

I guess the reason why was just what the very accidents of the earlier models demonstrated: it's better to have the a/p always disengage, because the pilots' brains seem to prefer it that way.

Quoting AverageUser (Reply 12): it's better to have the a/p always disengage

It does disengage on Airbus FBW with sustained control deflection (but apparently not in Go-around Mode?). Is that different in other types? I'm also pretty sure I've heard pros saying that things can happen very quickly during a go-around (high power levelling off and climb, possibly followed by a rapid levelling-off and thrust reduction, all while dealing with the flaps and landing gear) so it's sometimes better to leave it to the autopilot.

However, as Mandala499 says, pressing the disconnect button implies intention while yanking on the controls might imply a certain amount of "panic" but I'm perfectly prepared to accept that I may well be wrong about hauling on the controls not being a recommended method of disengaging the autopilot.

Quoting mandala499 (Reply 11): Need me to repeat stating the obvious for the others again there mate?

It all helps.

Quoting David L (Reply 13): However, as Mandala499 says, pressing the disconnect button implies intention while yanking on the controls might imply a certain amount of "panic" but I'm perfectly prepared to accept that I may well be wrong about hauling on the controls not being a recommended method of disengaging the autopilot.

Horses for courses my friend... there are those who prefer nudging the controls to disconnect in certain situations... there were some on the 73 who'd disconnect the A/P by jiggling the trim switches (I was like, what the?)

Mind you, I've shared a sim session with someone who kept disconnecting the A/P by either nudging (unintentional) or accidentally pressing the yoke A/P disconnect... *aaarrrghh*...

The thing is, the Interflug case, it may not have been an intention to disconnect, but merely trying to "smooth it out"... something you can do at other times as long as you don't go beyond a certain force on the yoke... Hence, sometimes, someone forgot when that was not supposed to happen... and end up with a "what the? why is it doing that? OK, let me correct it some more" and end up oscillating due to trying to counter the A/P...

Had a case of an A320 A/P disconnect the other day (that thing doesn't disconnect easily unless you yank the stick or do something weird...)... plus an ECAM warning... something like SEC/ELAC disagree... and aircraft departed normal mode and entered one of the alternate modes I think... why? Someone accidentally pressed the rudder continuously (thinking his foot was on the footrest...

for me the rules are simple, if you need to take over from the A/P, kill it... until you hear that disconnect horn, it's still on. (Caught out in a sim once on that! ended up out of trim at low altitude (it was NOSEDOWN  &nbsp 

Quoting David L (Reply 13): I'm also pretty sure I've heard pros saying that things can happen very quickly during a go-around (high power levelling off and climb, possibly followed by a rapid levelling-off and thrust reduction, all while dealing with the flaps and landing gear)

Oh yes, fast ! When it's a dog, something's wrong and the brain starts to think why...

Another thing is, if it's auto G/A, leave it... some of these planes have some kick to it that you might end up overshooting the G/A level off altitude (Record at one of the local airlines here, 2000ft over! the pilot swore never to switch from A/P to manual or vice versa halfway through a G/A again)...

Mandala499

Quoting Kaitak (Thread starter): The reaction of ATC (the incident happened at Moscow) or the passengers is not recorded.

I'm sure many sets of shorts were soiled on that day.

88 degree pitch up followed by a -42 pitch down followed by 70 degree pitch up then a -32 pitch down? Yeah, I'm sure there were more than a few people screaming.

Quoting David L (Reply 7): The Air Inter 320 accident was caused by ambiguity of the displayed information, i.e. FPA vs V/S, not by the actions of the auto-flight systems. I seem to recall that the display was changed to better differentiate between the two.

Tragically, the issue had already been identified and a change to the display was being retrofitted on the fleet. The change had not been performed on this particular aircraft yet.

Quoting kimon (Reply 8): I suggest a vote be taken amongst us forumistas to settle the safety debate between Airbus vs Boeing.

And this vote would resolve what exactly? I guess we would find out who things what.  

Quoting AverageUser (Reply 9): Both can be equally dangerous when operated wrong.

Quite.

Quoting Matt Macmaster:
I have ten years experience flying airliners both conventional and fly by wire. I am currently an Airbus A320 pilot. Flying is a technically complex issue so I understand a layman may find it difficult to get it all correct in such a report. There is no difference between flying a conventional and fly by wire jet. Period. The pilot just flies it like a conventional aircraft. The brilliance of the French in Toulouse was to design an electronic jet that responded EXACTLY like a conventional pilot would expect (except in some very minor instances like X-controlled landing). It is a wee bit more complex when something goes wrong and those built in 'protections' alluded to above so cleverly thought up by our friends in France disappear in the Hudson case. The link between pilot hand and control surface is electronic in this case but it flew like a cessna 152 as far as Sullenberger was concerned. Nothing more nothing less. He just used a side stick instead of a control yoke. While speaking recently to a senior trainer, he relayed that he had heard that at no point in the 3-4 minutes post bird strike had the thrust levers on either engine been advanced (like pressing the accelerator in a car). Not sure if this is true or that if it was, the less damaged engine would have responded anyway. Worth trying to find out though.

Modern aircraft like the Airbus rely almost exclusively on fly by wire (FBW) control systems to keep them in the air, and to take off and land safely. FBW uses motors and actuators to control the flaps, rudder, engine thrust and so on via wires or fibre optics, a process managed on the Airbus by multiple redundant computer systems.

Software controls the aircraft’s flight according to sets of rules or ‘laws’, which on the Airbus are called ‘Normal’, ‘Alternate’ and ‘Direct’. Airbus pilots are essentially systems managers, and if successive systems fail, then the plane will eventually reach Direct mode, leaving the crew just about able to fly the aircraft using only the rudder, elevator trim and differential engine thrust.

While redundancy normally makes the odds of getting to Direct mode remote, some pilots complain that FBW does not give them full control in an emergency, comparing Direct mode to trying to drive a car without power steering. In the kind of turbulence Flight AF 447 is believed to have experienced just before the disaster it may have been impossible to keep the aircraft upright, overstressing the airframe and causing the aircraft to break apart.

Quoting kimon (Reply 18): FBW uses motors and actuators to control the flaps, rudder, engine thrust and so on via wires or fibre optics

Essentially all modern flight control systems use motors and actuators to control the surfaces. What makes if FBW is the lack of direct mechanical connection between the pilot inceptors (yoke, pedals, throttles) and the actuators...the essential characteristic of FBW is the wires and/or fiber optics (in which case it's FlyByLight, FBL).

Quoting kimon (Reply 18): Software controls the aircraft’s flight according to sets of rules or ‘laws’, which on the Airbus are called ‘Normal’, ‘Alternate’ and ‘Direct’

This is absolutely true, but it's not a necessary part of FBW. Analog FBW has no software.

Quoting kimon (Reply 18): Airbus pilots are essentially systems managers

I don't agree with this charachterization; Boeing and Airbus FBW systems are designed to mimic the flying charachteristics of their non-FBW breathern. Done properly, the pilot shouldn't realize it's a FBW airliner.

Quoting kimon (Reply 18): if successive systems fail, then the plane will eventually reach Direct mode, leaving the crew just about able to fly the aircraft using only the rudder, elevator trim and differential engine thrust.

Direct mode still provides use of all flight controls, as far as I know. It certainly does on a Boeing.

Quoting kimon (Reply 18): While redundancy normally makes the odds of getting to Direct mode remote, some pilots complain that FBW does not give them full control in an emergency, comparing Direct mode to trying to drive a car without power steering.

As the FBW system degrades, you tend to *lose* the protections. As a result, you've usually got more control authority in Direct mode. But, having flown in direct mode, it is much more "sloppy" since you don't have all the artificial stability that the computers were providing before.

Quoting kimon (Reply 18): In the kind of turbulence Flight AF 447 is believed to have experienced just before the disaster it may have been impossible to keep the aircraft upright, overstressing the airframe and causing the aircraft to break apart.

This would have nothing to do with FBW...if your aerodynamic loads are enough to break the structure, it doesn't matter how the flight controls were being commanded. FBW is very good at making sure you don't do a structural overload, something which non-FBW isn't very good at (e.g. A310 vertical fin separation in New York).

Tom.

Quoting kimon (Reply 18): Airbus pilots are essentially systems managers, and if successive systems fail, then the plane will eventually reach Direct mode, leaving the crew just about able to fly the aircraft using only the rudder, elevator trim and differential engine thrust.

System Managers??? Man, U need to sit in the flight deck of an A320 in bad weather, and see them fly it... A/P off!
Anyways... Direct Law on Airbus still gives you controls over the elevators, ailerons and rudder. The difference is that the stick inputs are now directly translated to a control surface movement... it is no longer processed (which is normally G-demand on pitch, and roll rate, rudder is direct)... The newer built frames of the 330/340 don't even have mechanical linked rudder pedals to the rudder anymore! There is something below direct law... which is mechanical... (which is elev trim, rudder and differential engine thrust only)... that is something for total (ie: triple) HYD failure... Now, one does not need to be FBW to can only use the rudder, elevator trim and differential thrust... try a triple HYD fail on a 767... how do you control the aircraft? Elev trim, roll spoilers, thrust... that's it...

Quoting kimon (Reply 18): While redundancy normally makes the odds of getting to Direct mode remote, some pilots complain that FBW does not give them full control in an emergency, comparing Direct mode to trying to drive a car without power steering.

You can have full control even when not in emergency... so, if full control means you can do barrel rolls... yes, you can do that on the 320... just press the right switches and it'll let you... and flight control degradation means less protections from doing barrel rolls... so, when in emergency, it's the other way around... the pilots have MORE control...

If you want an aviation comparison to "driving a car without power steering" when power steering means hydraulics..., then try a 737 with HYD A and B fail... alias... manual reversion...

Direct law, is still like flying a conventional aircraft with the HYDs on... loose the power steering (all 3 HYD systems) in a 767 or 747... and your yoke is useless... just like the airbus' sidestick would be the same... useless.

Airplanes that require 3 HYD sys, is what I call, unlinked flight controls... which means, no HYD means no control surface movement... as the yoke controls the actuators... 767, 747, DC10 is like that... if it's power steering, then it's the 737 where there's a cable link connecting the yoke to the elevator and aileron (the HYD is just for power assist), and on the 737, the rudder is unlinked (rudder pedals to the hyd actuator, not the rudder itself).

Try an "artificial feel computer" failure on a 767... if it's not for the spring load to center on the elevator axis on the yoke, a 767 experiencing triple HYD fail would be uncontrollable ...

Quoting kimon (Reply 18): In the kind of turbulence Flight AF 447 is believed to have experienced just before the disaster it may have been impossible to keep the aircraft upright, overstressing the airframe and causing the aircraft to break apart.

The aircraft came down most likely on 1 piece... the aircraft is believed to not have entered direct law, but Alternate law (pitch still in G-load demand, roll is direct)... now turbulence up there and turbulence down here, are generally different on its effects on the aircraft.

Quoting kimon (Reply 18): Modern aircraft like the Airbus rely almost exclusively on fly by wire (FBW) control systems to keep them in the air, and to take off and land safely.

Totally incorrect. The Airbuses are not fighter aircraft with built in dynamic instability which requires FBW to keep it in the air... Now, direct law means... stick input x% forward and y% left means elevator moves down x% * total movement available, and left aileron goes up by y% * total movement available, and right aileron goes down by y% * total movement available... So, if it relies exclusively on FBW to keep it in the air, then that should also mean, it should fall out of the sky in direct mode... the airplane wouldn't be certified if that was the case!

Quoting tdscanuck (Reply 19): Direct mode still provides use of all flight controls, as far as I know. It certainly does on a Boeing.

Ditto Airbus and Embraer...  

Mandala499

Can anyone tell me what flight that was at 3:10? Holy crap that is scary.

I searched and the closest I could find is TAROM 381 (A310).

http://aviation-safety.net/database/record.php?id=19940924-1

Quoting tdscanuck (Reply 19): Quoting kimon (Reply 18): Airbus pilots are essentially systems managers I don't agree with this charachterization; Boeing and Airbus FBW systems are designed to mimic the flying charachteristics of their non-FBW breathern. Done properly, the pilot shouldn't realize it's a FBW airliner.

   I'm always puzzled by this allegation. Hand-flying a FBW Airbus isn't all that different from hand-flying other types and, as far as I know, autopilot usage is about the same on all similar types.

Quoting tdscanuck (Reply 19): Quoting kimon (Reply 18): if successive systems fail, then the plane will eventually reach Direct mode, leaving the crew just about able to fly the aircraft using only the rudder, elevator trim and differential engine thrust. Direct mode still provides use of all flight controls, as far as I know.

   As Mandala499 goes on to say, there is a degraded mode in serious situations that wouldn't be much easier to handle in other types.

Quoting mandala499 (Reply 20): Direct Law on Airbus still gives you controls over the elevators, ailerons and rudder. The difference is that the stick inputs are now directly translated to a control surface movement...

... which is exactly what the doubters want, isn't it?  

I've still heard nothing to change my opinion that criticisms of Airbus FBW come from those who don't know them. As to not allowing the pilots enough control in an emergency, when was the last time an airliner avoided disaster by being banked more than 67^o or by being stalled?

Quoting KPDX (Reply 21): I searched and the closest I could find is TAROM 381 (A310).

I suspect it is. Not FBW, of course,

Quoting David L (Reply 13): I may well be wrong about hauling on the controls not being a recommended method of disengaging the autopilot.

The other disadvantage of this rather agricultural method is this. In the case of manual over-ride, there will be a number of pounds/kilos control yoke force at which the autopilot will disengage. I have no numbers to hand but lets say a 40 pound pull on the yoke disengages the A/P. At the moment of disengagement, you have a manually controlled aircraft with a huge force pulling on the yoke. An immediate forward control input will then instinctively happen as the surprised pilot reacts to the initial pitch-up. Clearly undesireable compared to a concious, pre-meditated, intentional disengage using the correct button.

Yes the electric elevator trim will do it, but why bother, when the A/P knob is two centimeters away?!

A look at pilots in the sim shows that go-arounds aren't the easiest thing to do accurately, in terms of procedure or flightpath. One might do only the mandatory two or three every six months in the sim, and none until the next sim. Hardly conducive to maintaining proficiency. Pilots who are good at immediate recall and execution of rote memory sequences usually do well. For those of us without that particular strongpoint its high workload, especially when its unexpected. Suddenly a routine approach turns into a high intensity, high workload procedure for which the memory actions, and the flightpath, need to be as simple as possible.

Its often said at all levels of aviating that one should fly each approach as if its going to end in a go-around. This is valid for Cessnas as well as heavies, but how often does a pilot on final approach mentally rehearse the G/A route, the vertical profile/level-off altitude, and the handling actions? Its probably been discussed in the approach briefing 20 minutes ago, but it may only be in the case of an approach to minimums, or a limiting crosswind, or realisation that the preceding aircraft may not clear the runway etc. that a G/A is actually on the pilots' minds on finals.

The point is that it might be a surprise when it does happen, there might be a "deer in the headlights" moment while the pilots make the mental transition from complacent, relaxed, routine approach mode, to G/A mode. This combined with a small gap in the pilot's knowledge of exactly what the A/P is going to do, as was apparent in the Airbus cases discussed above, makes a mountain out of a molehill.

Quoting tdscanuck (Reply 19): A310 vertical fin separation in New York

A300. The A310 incident was an Air Transat rudder seperation (most of it anyway).

Have to go to work. I'm on airport standby...............

regards - musang

Quoting Musang (Reply 23):

Much as I suspected, plus additional details - thanks.   

I just read the A310 FCOMs for this reply... and before I start, I'd like to say, gimme a Boeing (even the 732 AOM is better written) or A320/330/340 FCOMs or E170 AOMs any day!!!!! Bloody hell!!!! (personal opinion though   )

Quoting Musang (Reply 23): I have no numbers to hand but lets say a 40 pound pull on the yoke disengages the A/P.

Here it is for you mate   (A310 FCOM1.03.33)
Below 400ft
- LAND green or GO AROUND mode announciated in the FMA
-- 20kg/45lbs nose-down
-- 46kg/100lbs nose-up

- with any other vertical mode (eg: V/S or GS green) announciated on the FMA:
--15kg/33lbs nose-down or nose-up

Above 400ft
- With GO AROUND or any other vertical mode announciated on the FMA
--15kg/33lbs nose-down or nose-up

Quoting Musang (Reply 23): Clearly undesireable compared to a concious, pre-meditated, intentional disengage using the correct button.

Very nicely put... very nice...  
When you want it off, you switch it off... When you want it to do something... let it be... if something's not right... switch it off...

But then, we go to... "what if you forget to switch it off the right way?" Its an elaboration of:

Quoting Musang (Reply 23): Suddenly a routine approach turns into a high intensity, high workload procedure for which the memory actions, and the flightpath, need to be as simple as possible.

So, I want to correct the A/P's trajectory and put in control inputs... I push the yoke, the autopilot will not resist the nose down through the yoke, but will instead go against it in the trim... Would I press the trim switches? Probably not... I'm preoccupied with "what the??? why is it not reducing the pitch?"... so I press the yoke further... until... it disconnects... if I hear the disconnect horn, I'd go "Oh, it's gonna dive if I keep the damn yoke full forward... so I let it go, forgetting that the trim is giving me a HUGE nose up... so, now it... "the A/P's off... but why is it doing this?"

Simple, switch it off first before getting into the mess... make it into a habit whenever one wants to disconnect the autopilot... better than the "agricultural method"... LOL! Gotta put that phrase down in my book Musang!

Now, I don't know if they've modified the 310's AP yoke input disengagement since the incidents/accidents (whichever it is, one thing's common, "since the roller coaster rides"... which, by the way, includes more examples but I can't recall off the top of my head)... but reading the A310 FCOM, I find it one hell of a quirky plane. It's like trying to describe a Boeing in Airbus language, and then, write about the Airbus features (such as the protections etc) in Boeing language.

The rollercoaster cases of the A310, is NOT a problem on the FBW system... the A310 is only partially FBW... The slats and flaps, spoilers (roll, ground, speedbrake). Primary roll and pitch, is still not FBW... it's fly by HYD (just like the 767, 747, DC10, MD11)... it's uncoupled flight controls (which belongs to the 737)... the cables from the yoke, goes to the actuator controls, and the actuators are the ones moving the control surfaces. (This is different from 737's power steering on the roll and pitch).

Even non FBW aircraft can have computer related controllability problem if you let it in the loop. Any aircraft can have computer commanded envelope protection, or autopilot overrides, or pilot overriding the autopilot... For non-FBW aircraft, this gets a bit complex as you have a mechanical device somewhere within the control cables to facilitate that (and the A310 diagram for it, is a nice one, but raises my eyebrows!).

The Autopilot will not fight you on the yoke input on the pitch, it'll do it on the trim! Hence, if something's wrong and you need to control it / override it quick... switch it off!!!!

Here are some excerpts from the A310 FCOM on automatic flight system:
"In CWS, there is no automatic disengagement by stickforce on the control column or control wheel (in CWS mode, the pilot cannot work against the AP but
works through the AP)."
and:
"Overriding the AP pitch actuator is not possible. Any attempt to override the AP in pitch results in the automatic disengagement of the AP before or when
reaching the break-out threshold of the AP pitch actuator override detent."

"The AP automatically disengages if the pitch trim motion is stopped by holding the pitch trim wheel."

So, the A310, still has the priority logic on pitch control like any other conventional aircraft... that is... pitch trim first, if more, then the command the elevators (and the yoke moves too as there's a mechanical junction with it and the A/P command servos). Stop the Yoke and the trim wheel, and it'll cease to function and disconnect because it's saying... "I can't do what I want to do, so here, have the plane back." And just like any other conventional aircraft, the electric trim switches overrides any autotrim... (and disconnects the A/P if you press the yoke trim switches)

Now, with conventional aircraft (albeit uncoupled/FBH controls), a large change in thrust will cause some nose up/down movements... Here's more from the FCOM: In go-around:
"PF should be ready to override or disconnect the autothrottle function in case of thrust asymmetry and to counteract aircraft yaw.

Rotate the aircraft at a rate of typically 3° per seconds.
Set a 18° pitch attitude then follow the SRS orders if available (not exceeding 18°) or maintain 18°.
As thrust increases, be prepared to counteract the thrust pitch-up moment.
Trim the aircraft nose down as required. The pitch attitude should not be allowed to develop beyond 25°, as such a pitch attitude would result in a significant speed loss.
An immediate and firm elevator nose down command (together with a nose down pitch trim order) would allow to recover the target pitch attitude."
Note it doesn't say yoke only...

But then, here's a quirk that got me thinking... now unless this was modified after the roller coasters:
"If both AP were in CMD for the approach, both stay engaged during the go-around, as long as GO AROUND mode remains engaged."
With the above, I can understand why the A/Ps would remain engage, but if both AP1 & 2 were in CMD. The question is, looking at the Tarom footage in the video, they were definitely not in CAT II or CAT III condition... and both APs on are only required in LAND 3 situations... Did the crew engage BOTH APs???

Others may disagree with my opinion, but I find the A310 (and A300-600), a bit quirky... here's something from the company training manuals of the A310 from an airline...

"Flight Envelope Protection
A further anomaly exists with the speed protection system that is not explained in any of the Airbus manuals, and has only been discovered by experience on line operations. When descending in SPD mode with the throttles in idle and A/THR armed, i.e. showing A/THR blue, (such as when conducting a descent in LVL CHG mode), and if the FPV is now selected, the ATS system will no longer capture the target SPD set on the FCU. Underspeed protection at Vls will also no longer be available, and Alpha Floor protection with THR L is then the only underspeed protection available. In such cases, pilots should ensure that a
SPD mode is reselected. SPD GREEN will then be indicated in the thrust column of the FMA."

I think I'd stick with planes other than A300-600/A310. The Airbus FBW family (320 and later), was set up based on the experiences and quirks of the A310, and in my opinion, are much better aircraft. Envelope protection (beyond Autothrust and into flight controls) on conventional aircraft, can yield strange results. The A320 and later, does away with those... if you're on A/P, any sidestick input will not retrim the aircraft or put it in a different bank, until you disconnect the A/P through any means, so it won't catch you! On a 777, the trim the pilot know is to speed (processed trim to give neutral bias for a given speed regardless of everything else)... so, yank the yoke anywhere you want to break out of the A/P (for those preferring the "agricultural method"), and when it disconnects, it'll just follow what the yoke is doing... no trim upsets, etc.

This is why we do not have envelope protection through the flight controls (via A/P or not) on conventional aircraft (ie: where the aircraft is in direct mode all the time, be it FBW or not), and it's why there's no envelope protection for FBW aircraft flying in direct mode. The A310 as the halfway house solution, isn't my cup of tea.

Kimon, one thing I do not like about that particular episode, is that it is biased towards the "tractor driver" mentality. Whilst "tractor driver" mentality has nothing wrong with it, that episode went beyond into the "anti-technology" by confusing (deliberate or not is a different matter) one to think that those upsets were FBW... and that "computers on controls = FBW"... something totally *censored*. Even the 732 has a computer to help you controlling the aircraft!

And enough of this rant... I don't even know what I'm ranting about!   

Quoting mandala499 (Reply 25): Now, I don't know if they've modified the 310's AP yoke input disengagement since the incidents/accidents (whichever it is, one thing's common, "since the roller coaster rides"..

The relevant document is here: http://www.faa.gov/other_visit/aviat.../all_infos/media/2008/FSAT9513.pdf (1995) Seems to me the lowest height for a/p disconnecting by yoke force has been changed to 400 tf from 1,500 ft.



Airbus Industries issued Service
Bulletin (SB) A-300-22-6021 which provides for a modification to
the flight control computer to change the software control laws
for the A-300-600. This modification provides for the
disengagement of the autopilot when a force of about 33 pounds
is applied to the control column in the land or go-around modes
above 400 feet AGL.

also:
The lack of a stabilizer-inmotion
warning appears to be unique to the Airbus A-300 and A-
310.

Cheers for that... shocking... 1500ft before, then to available above 400ft, and now, at all times (see the "below 400ft" in the post above... taken from 2007 version of FCOM 1, A310).

Nice find !  

Quoting mandala499 (Reply 25): "In CWS, there is no automatic disengagement by stickforce on the control column or control wheel (in CWS mode, the pilot cannot work against the AP but works through the AP)."

We're introducing a complication here! Control Wheel Steering [its called other names also, like TCS] on the 737 Classic at least, is where the A/P simply holds an attitude. A bit like I imagine an Airbus to behave in that the pilot pushes or pulls the column to attain a desired pitch, lets go, and the A/P maintains that pitch. Its the same basic idea in roll, but with protections built in. I use it in the climb sometimes, as our 737s are atrocious at holding the selected speed (in Level Change and VNAV Speed, where they should pitch to maintain command speed) and seem quite happy to get up to 15 knots off without correcting the pitch.

Anyway, the way CWS works is by accepting pilot inputs against the A/P. You wouldn't want the A/P to disengage in CWS. So I'm surprised the Airbus manual says that!

Quoting mandala499 (Reply 25): they were definitely not in CAT II or CAT III condition... and both APs on are only required in LAND 3 situations... Did the crew engage BOTH APs???

Our company procedure is to use both A/Ps for any ILS approach even in visual conditions. I suppose its to establish a common procedure for all ILSs so we won't need to do anything different when its Cat III.

Regards - musang

Quoting Musang (Reply 28): Our company procedure is to use both A/Ps for any ILS approach even in visual conditions. I suppose its to establish a common procedure for all ILSs so we won't need to do anything different when its Cat III.

OK, coming from the land of no Cat II, I guess your answer explain why some procedures differ between companies... I was wondering about that because in the current version of the 310, No A/P disengagement on G/A requires both APs on, but I remember nothing from those incidents covering that... until AverageUser came with the SB...

Quoting Musang (Reply 28): I use it in the climb sometimes, as our 737s are atrocious at holding the selected speed (in Level Change and VNAV Speed, where they should pitch to maintain command speed) and seem quite happy to get up to 15 knots off without correcting the pitch.

Better the deviation is caused by you and not the A/P... the 737's reputation on the A/T varies from OK to atrociously sloooooow...   

Quoting Musang (Reply 28): So I'm surprised the Airbus manual says that!

Well, my eyes nearly popped out when I read that...    As I said, one of the strangest FCOMs I've read is the A310... "conventional/FBH with A/P protection" just sounds weird to me... I understand why they did it as they wanted to push the technology paving the way for the full FBW 320, but, the halfway house is... well... confusing...

Skip to 08:23"
http://www.youtube.com/watch?v=Adv-XtxOjxQ&feature=related

"The controls suddenly feel completely different to the pilot," says flight expert Hüttig. The sheer complexity of the Airbus' systems makes it difficult to control in critical phases of the flight. It would be easier for pilots if they could simply switch the computer off in critical situations, as is possible on Boeing planes"says Gerhard Hüttig, an Airbus pilot and professor at the Berlin Technical University's Aerospace Institute.

Quoting kimon (Reply 31): The sheer complexity of the Airbus' systems makes it difficult to control in critical phases of the flight.

I would argue that a modern Boeing is just as complex as a modern Airbus.

Quoting kimon (Reply 31): It would be easier for pilots if they could simply switch the computer off in critical situations, as is possible on Boeing planes"says Gerhard Hüttig,

They can? News to me. In any case it would certainly not be better. There's a critical situation and you want to increase pilot workload by turning off automation?

I think Mr Hüttig is missing the point. The Airbus flight controls are not supposed to feel different. AFAIK, pilots are told to fly as they do any other aircraft.

Quoting kimon (Reply 31):

Any chance of providing a source to put that post into perspective?

Quoting kimon (Reply 31): "The controls suddenly feel completely different to the pilot," says flight expert Hüttig. The sheer complexity of the Airbus' systems makes it difficult to control in critical phases of the flight. It would be easier for pilots if they could simply switch the computer off in critical situations, as is possible on Boeing planes"says Gerhard Hüttig, an Airbus pilot and professor at the Berlin Technical University's Aerospace Institute.

Does such a "panic button" exist that would revert all controls to basic or manual law instantaneously?

Faro

Quoting kimon (Reply 31): "The controls suddenly feel completely different to the pilot," says flight expert Hüttig. The sheer complexity of the Airbus' systems makes it difficult to control in critical phases of the flight. It would be easier for pilots if they could simply switch the computer off in critical situations, as is possible on Boeing planes"says Gerhard Hüttig, an Airbus pilot and professor at the Berlin Technical University's Aerospace Institute.

He was talking about the AF447 when it reverted to ALTN 2 Law????
OK, what is ALTN 2?
Pitch under Alternate Law... which is basically, Normal Law and the only protection is excess G-load protection. This feels the same as under normal law.
Roll under Direct Law... which is basically, your stick input = the degree of aileron deflection... this feels slightly different from Normal law... plus no protection... the only difference is that under Normal law, your maximum roll rate is 15deg/sec... under Direct your maximum roll rate is 20 - 25deg/sec depending on your speed.

Yaw under Alternate Law... which is basically the same as Normal Law, the only difference is the damper authority (to protect against dutch roll) is reduced to +/-4degs flaps up (+/-15 all other config) and that turn coordination is not provided in flaps up situation (only slips and skids affected).

Ask our resident Airbus flyers here (Zeke, Pihero, Wing) on Altn 2... see how different it is from normal law... (they'd likely say... "not much difference".)

Quoting Starlionblue (Reply 32): I think Mr Hüttig is missing the point.

He is... On the 777, there is a switch for you to kill/reset the primary flight computers... on the 330, the switch also exist, but there are 3 of them. Switch the primary flight computers off, and the plane goes back to Direct Law... be you on a 777 or 330.

Quoting David L (Reply 33): Any chance of providing a source to put that post into perspective?

It was on Spiegel's article on AF447... mentioned in this topic: Fuel Consumption/Flight Managment System (by Luftfahrer Feb 25 2010 in Tech Ops)
The article itself:
http://www.spiegel.de/international/world/0,1518,679980,00.html
See part 2... it's there... which is in my opinion, pure scaremongering on the Airbus FBW control systems.

The deepstall due to icing isn't only an Airbus FBW problem... it is a risk on ALL "g-load pitch demand manouver" FBW aircraft... which includes 777, 787 and the E170/190...

Mandala499

[Edited 2010-02-27 04:55:11]

The article has some serious factual errors:
http://blogs.crikey.com.au/planetalk...-and-too-much-freight-in-the-hold/

There is one thing I agree with Crikey... which is:
Der Spiegel basically holds the leading editorial franchise for bashing Airbus in Europe, and would disappoint many of its readers if it didn’t run a scathing indictment of the company or its wicked and evil parent EADS at least once every season. Those rants are usually full of obvious nonsense. The problem with Gehirnschlag im Cockpit is that only some parts of it have to be true for Air France to be in very serious trouble.

Quote: In summary, the German article, which quotes sources in the Air France pilot’s union, among others, says that AF 447 was overloaded prior to departure from Rio de Janeiro to Paris late on the night of May 31. Specifically, that it came in at 237,757 kgs, or a mere 243 kgs below maximum takeoff weight or the flight before another ten tonnes of freight was loaded.

If there was the extra 10 tons of freight loaded... it shouldn't have brought the aircraft crashing down... It went through the bloody storm before it sent automated signals of the failures. Plus, if the crew knowingly had a little over 9tons exceeding MTOW when it took off, they would be eyeing the "In Cruise Check" page on the QRH/FCOM 3 and monitor the Estimated Fuel On Board at destination all the bloody time... to make sure they have enough to get there...

But then, it's always too damn easy to blame the dead!!!! I don't read the English version of Der Spiegel for a reason...

Why is Der Speigel so anti-Airbus?
Sensationalism to boost circulation or some other agenda?

Quoting kimon (Reply 38): Why is Der Speigel so anti-Airbus?

Dunno... anti French and anti-Hamburg perhaps?   

Quoting kimon (Reply 31): The sheer complexity of the Airbus' systems makes it difficult to control in critical phases of the flight.

Quite the opposite...all that complexity is in there precisely to make the plane *easy* to control in critical phases of flight.

Aircraft flight dynamics are very complicated, and it gets worse as you try to eke more performance out of the aircraft. In days gone by, you'd simplify the flight dynamics down to where you could actually work with them to implement (relatively) simple analog control laws and that worked pretty good. With modern FBW, you can work with the true airplane dynamics and make them behave much closer to an "ideal" airplane. For example, getting exactly neutral spiral stability with a non-FBW system is almost impossible, with FBW it's ridiculously easy.

Quoting kimon (Reply 31): It would be easier for pilots if they could simply switch the computer off in critical situations, as is possible on Boeing planes

They can do that on both Airbus and Boeing FBW. Although, at least on Boeing, you're not really turning the computer off, you're taking it out of the control loop (and moving to a much simpler other computer).

Quoting faro (Reply 34): Does such a "panic button" exist that would revert all controls to basic or manual law instantaneously?

Yes. But it's not a "panic button" in the sense that you should do it in the event of an airplane upset or something like that...it's there only to protect against a failure of the flight control computers themselves.

Tom.

Quoting mandala499 (Reply 35): Quoting David L (Reply 33): Any chance of providing a source to put that post into perspective? It was on Spiegel's article on AF447... mentioned in this topic:

Cheers - I haven't read that thread yet. All I could see were some words that could have come from anywhere.

Quoting kimon (Reply 36): The article has some serious factual errors:

Maybe it's just me but I'd probably have done that research before posting rather than a few hours later.

Quoting tdscanuck (Reply 40): Quite the opposite...all that complexity is in there precisely to make the plane *easy* to control in critical phases of flight.

Some people seem to be under the impression that the crew has to re-code the computer programmes every time they want the aircraft to do anything.  

Quoting David L (Reply 41): Some people seem to be under the impression that the crew has to re-code the computer programmes every time they want the aircraft to do anything.

Nah. You just hit "Ctrl-Alt-Del" on the keypad, kill the offending process (fltctrl.exe), then restart it and wait for the little hour-glass icon to stop spinning.

Tom.

Quoting tdscanuck (Reply 42): Nah. You just hit "Ctrl-Alt-Del" on the keypad, kill the offending process (fltctrl.exe), then restart it and wait for the little hour-glass icon to stop spinning.

And then get a blue screen error...   

This should solve it:http://ccollomb.free.fr/unlocker/#description

Would it solve it? Those airplanes don't run on windows for the FCCs...

Quoting mandala499 (Reply 45): Would it solve it? Those airplanes don't run on windows for the FCCs...

My original comment was firmly tongue in cheek.

The only thing I know of on a large airliner that runs Windows is the some of the EFB's and some of the IFE. Flight critical functions run on DO-178B operating systems, like Integrity from Green Hills or VxWorks from Wind River.

Tom.

Quoting tdscanuck (Reply 46): My original comment was firmly tongue in cheek.

I know it was... I was actually concerned about this reply...

Quoting kimon (Reply 44): This should solve it:http://ccollomb.free.fr/unlocker/#description

Mandala499

Quoting mandala499 (Reply 47): I know it was... I was actually concerned about this reply...

Whoops! My bad. I need to make better use of the emoticons...

Tom.