I don’t think that is correct. The autopilot and auto thrust disconnected in response to invalid airspeed readings, the flight control logic also went to alternate law for the same reason. I don’t fly Airbusses either, but it’s my understanding that this is what caused the autopilot disconnect, not the stall itself.
I doubt the Airbus stall recovery would be any different from other aircraft. The protections are there to protect against things going wrong, they’re not a “get out of jail free card”. Also, the flight control logic had gone to alternate law, and the stall protections would not be available any more.
I don’t know–it sounds like an idea worth trying to me.
IANAP, but I have driven submarines. As few as a dozen or two people moving all the way forward or aft will definitely affect the trim of a submarine–in fact, the prank is referred to as a “trim party.” (It’s commonly pulled on newly qualified Diving Officers of the Watch.)
Anyway, you state that the plane weighs up to 514,000 lbs at takeoff, or 257 tons. A modern submarine with just 140 personnel or so on board weighs 6,000 tons or more, and the movement of just a few people affects the trim. In fact, I’ve experienced trim parties (in which a sub was proceeding slowly) that it could not maintain depth control (i.e. the control planes could not counteract the moment produced by the people moving forward and aft). The diving officer had to request additional speed from the Officer of the Deck to maintain depth control.
Back to the airplane–I think you could cram enough passengers up forward to make a difference. Fill the aisles, put people in each others’ laps, etc. It’s worth a try, especially when you’re out of options.
I should have had the pilot tell only the people at the back of the plane to stand up and move. (The crush develops when people at the front are out of their seats and blocking the aisles.) Or … is there a way to dump the rear cargo hold?
What was the plane’s pitch attitude as it plummeted? That may be a key variable to help understand if any of these ideas had any chance.
The idea of getting folks to run to the front is sound. I know of one Delta L1011 which took off SAN with the trim mis-set & jammed. They saved the day by moving the two-legged ballast full forward and artful use of their low & high mounted engines to create enough nose down moment. And a couple of other neat tricks not relevant to the current discussion. The story is passed around as an example of excellent outside-the-box thinking.
In the case of AF443 moving the two-legged ballast would have been problematic due to deck angle. And A330s only have low-mounted engines.
And no, cargo in the belly holds cannot be jettisoned inflight.
The L1011 incident you mention is not included in Wikipedia’s incomplete list. Do you have a link to an outside account of that incident?
I’m not a submariner, but I would imagine things on a submarine happen slowly enough (and the occupants, being trained/disciplined sailors, will comply readily and rapidly with orders issued by the command staff) that people can move fore/aft in a timely manner so as to make useful trim corrections before the sub reaches crush depth.
Likewise, I’m not a commercial airline pilot, but I would think that if the pilot is unable to correct for a weight imbalance, things are going to happen far too rapidly for passengers to make corrections. Moreover, the passengers are untrained, inexperienced civilians who are probably not going to respond quickly or effectively to such out-of-the-norm instructions, especially when they’re already panicking because the aircraft has become a rollercoaster.
A search of the NTSB archives came up with this incident report which is probably the flight in question Page not found
As is commonly the case with non-injury events, the NTSB report has zero useful information.
The fact the airplane was recovered successfully with no major damage & no injuries meant there isn’t / wasn’t much other news coverage. I did not try searching the whole internet with the additional factoids from the NTSB report.
This article: Stalled AF447 did not switch to abnormal attitude law appears to answer the question about the full nose-up trim. Under “alternate law” (as prevailed on AF447 after autopilot & autothrottle disengaged) pitch trim remains under automatic control. The 13-degree nose-up trim happened automatically in response to the pilot’s nose-up control inputs. This was obviously a significant problem, the more so since the pilots did not directly command it and gave no indication they were aware of it.
Under “abnormal attitude law” auto-trim is not enabled - but this is entered only when extreme attitudes are detected:
I’m guessing that Airbus will be re-thinking the persistence of automatic trim, and possibly expanding the circumstances that result in abnormal attitude law.
I disagree, the pilots did command it, the auto trim is just doing what the pilots ask it to, it will not fight the pilots. If the pilots pushed the stick forward to command a nose down attitude, the autotrim would have trimmed nose down.
But not directly. Which I think means they were less likely to be aware of it.
Apparently this takes some time - the BEA report says that it took about a minute for the trim to go from 3 to 13 degrees up.
But the fundamental point is that the pilots used nose-up control inputs through most of the 3.5 minutes, which indicates that they were never aware the plane was in a deep stall. You’d think that the combination of high pitch-up angle, (occasionally) high engine power and rapidly decreasing altitude would have suggested this.
Yes, I agree with this. Part of the problem may be due to the way stalls are trained in large aircraft. There has been a trend to teach recovery from an * immanent* stall as full power and hold the nose up to prevent height loss. This works when the aeroplane hasn’t actually stalled yet, but it seems that some trainers aren’t adequately stressing that this is a recovery from an immanent stall, and not an actual stall recovery.
I’m guessing this stall training is done on the assumption a ‘immanent’ stall will be most likely be at low altitude during takeoff or landing, when a sudden altitude loss could be catastrophic in very short order.
But at high altitude losing a few hundred feet is not particularly dangerous at all, correct?
So I guess the question is – are pilots trained to recover from stalls or imminent stalls differently, depending on the altitude of the aircraft? One would think that at 35,000 feet, they should be trained to push the throttles AND push the nose down. Well, one untrained person like me.
From what was posted earlier it sounds like this is how the training works.
Apparently high altitude stalls have been given little attention. Stall recovery has trained them to do it at low altitudes. I assume this is because a plane flying at cruising speed at altitude is not likely to be in a stall situation while when landing the plane is sort of on the edge of one so a stall is more likely.
Again, from earlier in the thread, it sounds like more attention is now being paid to training recovery from high altitude stalls. I think you are correct that at high altitude proper recovery is to trade altitude for speed to recover. I do not think this is particularly dangerous although you may freak the passengers (which is to say you still do it but it might be a wild ride so best not to go there in the first place if you can avoid it).
My question remains why these pilots did not seem to think they were in a stall. We established they almost certainly could look at the instruments to see which way the plane was pointed (nose up). I also assume the altitude indicator was ticking down alarmingly fast. I’d think that any pilot with that info alone would figure they had stalled whatever the air speed readings.
I guess the issue is that, by the time they looked at that info and determined what was happening, they were in an unrecoverable situation.
The question to me then becomes did doing what they were supposed to do when things started going wrong lead them into it or should they have been able to discern the problem but missed it (understanding they had faulty air speed info)?
I guess I cannot understand why the pilots did what they did between the autopilot shutting off and getting themselves into an unrecoverable situation. I understand it happened relatively fast and the cockpit was a confusing place given all the circumstances. Still, did they apply their training and this resulted or did the plane’s automatic systems screw them (albeit unintentionally)? Or did the pilots just fuck up badly? (Not to blame them overly much due to the unique situation but can we expect them to have performed better anyway.)
My post was in response to Richard Pearse’s most recent post saying that the pilots did know they were at least entering a stall, but their training still dictated they pull the nose up to avoid altitude loss. Hence my question about whether their training for stall, or imminent stall, recoveries would lead them to different sets of responses at different altitudes.
If this question has already been answered, can you point me toward the relevant post?
Of course, your question still stands – how is it that they seemingly never figured out that pulling the nose up was making things worse?
I think something to remember is that airline pilots are essentially systems managers, that is their primary job and that’s what they get good at. Some of them will have an aviation background that has armed them with good “stick and rudder” skills, they will be what most people would consider “good pilots”, but many of them don’t have that. They’re not bad pilots at all, it’s just that their experience has pretty much been focussed on airline flying and pure flying skills have not been required. They are fine airline pilots but they may not have the experience to cope with some of the more abnormal situations. Like any profession, most of the workers are average, some are really good and one or two are bad. When an exceptional set of circumstances is encountered by an average pilot, it is possible they won’t cope with it. The best defence is to avoid getting into that situation to begin with, that is where the airline pilot wins or loses the battle.
ETA: my girlfriend flies in to London from Central America on Sunday. And I’m now in a ridiculous groundless panic about her flight because of this thread.
I think you are right. But autopilots, autothrottles, etc. are now significantly better than humans at managing aircraft systems under routine conditions (which prevail better than 99% of the time). And the disparity between human and computer performance is increasing with time.
Beyond PR, the only strong justification for having human pilots aboard is to deal with non-routine situations. To have a good chance of success they need good training and aircraft that don’t tend to fool them or create traps.
Most old-school pilots would probably view some good stick-and-rudder skills as a definite plus, but I think this receives low and declining emphasis. Airlines want pilots that do exactly as told (and trained) - independent thought is not encouraged. As a flight instructor in gliders, I’ve flown with experienced airline pilots that had excellent stick-and-rudder skills - and some who were really quite bad. Many of the latter fall into the category: “I have no feel for the aircraft, the air, and what these are telling me - show me instruments from which I can read numbers that will tell me what to do.”
It will be interesting to hear about the extent to which the AF447 pilots were trained for the situation they faced. A layman’s view says that 3.5 minutes is plenty of time in which to figure out that nose-up control inputs aren’t helping and another idea is needed. But perhaps their training somehow called for this reaction.
I have been reading this thread and just couldn’t wrap my head around the idea that the pilot would just keep trying nose up inputs. I started to question just how much or little info has been released. So I went and read the BEA note for myself. I think it is important to keep in mind that they refer to this data release as a note. We have no idea whether it is comprehensive or all inclusive for the time periods describe.
However, in light of the above discussion I think it is important to note a couple points.
Not exactly sure what is meant by “pitch-down inputs”, but it certainly isn’t a nose up input. This would have been around 35 seconds after the Captain reentered the cockpit.
I interpret the above as the Captain being given control. Though that is pure speculation, it seems plausible given the timing of events. I have a feeling this is may just come down to a “perfect storm” of events. A number of critical things happening at the unfortunate time when the most senior and presumably most proficient pilot is not on the flight deck. I’ve seen how subordinates can respond when stuff goes awry and senior leadership isn’t present. Sometimes they step up and take authority and sometimes they waste precious time trying to validate that they must take authority.
I think it unlikely the captain ever took control during this emergency, as it would have required one of the two pilots already seated to unbuckle and get up in a crowded and unstable cockpit. Scuttlebutt is that the authorities don’t believe the captain was ever at the controls during this emergency, but of course we, the people on this message board, don’t have all the information.
I just want to clarify that no one gets to even the lowest co-pilot position without being competent to fly such an airplane, and even the newest hires are typically well past the minimum requirements.
The BEA report uses the convention “PF” or “pilot flying” to indicate who is doing just that, flying the airplane, as that is more relevant to accident investigations than who is pilot or co-pilot. Who is captain is based on seniority with the company, not necessarily who is most “proficient” however you may define the term. A less senior pilot may also have skills particuarly revelent to an emergency situation as well - the incident that springs to mind is the Gimli Glider incident where the more senior captain turned the controls over the co-pilot. While to the non-flying public that might at first glance seem peculiar there was sound logic behind it - they had just lost all engines on the airliner and could not restart them, and the co-pilot just happened to have a hobby of flying gliders, and competed in international competitions doing just that. Although he had total fewer hours flying airplanes than the captain he had vastly more experience flying unpowered airplanes… and thus, for an airliner suddenly without working engines was arguably the better choice to be at the controls. (By the way - he safely landed the airplane, just in case you were wondering)
You also shouldn’t assume the captain is the most senior and/or most experienced. These days, it’s entirely possible for the co-pilot to have more hours in more airplanes than the captain does. That is, after all, why various agencies are getting away from the terms “captain” and “co-pilot” and saying “pilot flying” and “pilot not flying”. Yes, the most senior employee will, under most circumstances, have final say and is expected to assume more operational responsibility, may be the one to delegate certain duties, and so forth, but in an emergency whoever has the controls is in charge.
So, really, the Pilot Flying should be assumed competent unless proven otherwise, and while he is flying the airplane has full authority to deal with any situation that comes up. He will NOT be waiting for the captain to show up and given orders in an emergency. In aviation, there is often just not enough time to wait for direction from above, the guy at the controls has to be able act and react without needing direction. True, this is at odds with other industries but that is the norm in aviation.
This is also part of why accident analysis takes more than just a week or two - it’s not at all unusual to review the flight history of all pilots involved.
