I didn’t like this article’s focus on Lion Air’s presumed negligence or lack of safety procedures and training while completely ignoring the presumption of negligence on the part of Boeings manufacturing procedures. The plane is only “3 months old”! In my experience, sketchy, unsafe transportation businesses, like bus companies, limo services and even small commercial airlines, don’t buy “new” vehicles let alone “new” aircraft. Applying the same logic, why would Lion Air hire incompetent or less than stellar pilots to fly their very expensive air craft? Did Boeing vette this article first? Why aren’t we looking at sloppy or errant production policies and procedures at the Boeing plant with equal scrutiny? There seems a greater probability that manufacturing, and not operating error is the cause of this tragedy. Therefore, the article should reflect that difference.
5
Colgan Air Flight 3407 crashed after the captain reacted incorrectly to a stall warning and then overrode the stick pusher system. He raised the nose, added insufficient power, and continued to pull back against the stick pusher. The plane stalled.
5
The article states that the feature, which applies nose down stabilizer trim when the angle of attack is too high, is new to the 737Max. This is incorrect. Older model 737s, which have flown for many years and millions of flights, have the same feature.
7
Piloting experience in terms of total time and time in the specific aircraft are the most important variable in any accident in my opinion. While the 737 is hardly a new aircraft itself , the Max version is very new. While I have 22,000 hours of total flight time at the age of 55 (military and civilian)and 4000 hours in the 737-8 aircraft, I have exactly 3:02 minutes in the 737 Max. So the 22,000 hours of total time becomes critical. I will fly the Max more as my airline purchases more as replacements but there is a period of “vulnerability “ with any new aircraft type.
Runaway stabilizer trim is not a new phenomenon and every training syllabus addresses it but it might total 30 minutes total simulator training on the subject in the specific aircraft (that’s it). Checklists exist but runaway trim requires immediate attention without reference to checklists. The 737 Max is not highly automated (relative to any Airbus product) and “stick and rudder” skills are still a big part of piloting the Max. Immediate recognition of the “runaway trim” condition is critical.
Pilots like training and we are rarely happy with the amount of simulator training received when learning a new aircraft type. A typical training cycle on a new aircraft is 9-10 simulators (at 4 hours per session for 2 pilots). Since the Max is part of the 737-8 series it is considered the same “type rating”. So how much extra training to fly this newer version?? About 50 slides on an iPad. Zero simulator training.
26
"Called the angle of attack sensor, this instrument tells the pilot if the nose of the plane is too high, which could cause the aircraft to stall. In the Max 8, if the data indicates the nose is too high, the aircraft’s systems will automatically pull the nose down."
One would think there would be another sensor that would countermand the angle of attack sensor if it detected the jet hurtling at a near vertical angle towards the ground/sea at 400+ miles per hour.
7
Pilots needing to turn off the electricity to the rear stabilizers to gain control of their plane doesn't sound like good design to me.
9
I know this would rely 100% on perfect communcations but, maybe in a case like this a very experienced pilot could chime in and either give advice (as his sole standby job) or he could take over the plane in an extreme emergency in this case (tho that would present its own security risks). I think at the end of the day, pilots need more simulation time as well as knowing the ins and outs of his aircraft (whether it may be a software defect he needs to know how to handle or something else). Exp is king and when to take the computer out of the situation and fly that bird to the best of their human abilities...
4
I am haunted by the Air France 447 crash into the Atlantic Ocean in 2009 (my friend from Germany lost 2 of her coworkers on that flight, who left behind orphaned children). When aircraft become too automated, flight crews don't know how to respond to abnormal aircraft behavior. This Vanity Fair article about that crash was one of the best pieces ever written about modern aviation, and how quickly things can go tragically wrong: https://www.vanityfair.com/news/business/2014/10/air-france-flight-447-crash
13
Wondering if the computer couldn't give a display warning telling the pilot what it is doing to take corrective action and give instructions on how to cancel the corrective action.
5
With the ever growing complexity of gadgets and procedures in such aircraft, maybe a pilot ought to suction cup a $100 retail store GPS onto the cockpit window and fly manual once in a while. Might even work better than crazy-thinking million dollar avionics.
3
This reminds me f an Airbus accident that was linked to the same sensor. This is the problem with "automation" and the assumption that it will never fail. And pilots not being trained to overcome this kind of failure.
7
What, stick shakers weren't effective enough for the crews to respond to? Fine... they're no longer holding onto a yoke, but the audible and other attention directing devices surely could do the trick with less drama?
I have to wonder how much the industry is being influenced by the desire to go with less payroll costs, thereby offering up increased automation.
4
For seven years and 6,500 hours, I've worked as a pilot for a major US airline that is also one of the largest operators of the 737.
The Times did a nice job communicating the realities of a fraught operation, and there truly is a lot to talk about when it comes to technology and our ability to adapt as pilots. The recreation is especially hard to watch.
Automation is no longer simply "laid on top" for the convenience of the flight crew. We are working with tools that have evolved in the same way as the world we have helped to connect -- our airplanes are fully integrated. Every safeguard, redundancy, and protection is part of every normal system too, branching out in loops of logic that can recognize, assess, and respond with a quickness and confidence that is genuinely impressive.
And it works, too. It really does. Our automation can make a routine flight bo-ring, and our job is to keep it that way. But we do now have rules beyond the ol' "Aviate-Navigate-Communicate":
1. Garbage In, Garbage Out. We do a lot (a lot) of computer programming, but the computer is only as smart as its user. Verification is a core tent of airline training these days.
2. It's a Tool, Not a Crutch. Use the automation, but do not rely on it.
3. Stuff Happens. It's an incentive to the industry to think different, but the reality is this: my former turboprop ride was very capable of blowing up or falling apart without a fraction of the technology we are looking at here. Mitigation is the key.
13
21st Century aircraft are amazing pieces of technology. However, major fatal crashes are usually a collection of events that come together to cause it. The article clearly outlines many potential factors that may have contributed to this crash: new aircraft, not enough training for the pilots on the new aircraft, maintenance department with lax oversight, pilot fatigue, company culture that punishes pilot for calling out defects, and pilot error. But a rush to judgement without a through investigation is simply speculation nothing more. Aircraft manufacturers are one of the most scrutinized and regulated industries on the planet. If they messed up it will come out. Airlines are another story but in time the facts will emerge and the story will be known. In the meantime airline travel is safer now than any time in its history. Automation now plays a big factor in airline safety but life is not without mishaps. For the souls lost in this horrific accident none of this matters. They are gone and those of us left behind must wait for an answer - why did this happen?
6
Thanks for this article. It covers the situation with greater clarity, breadth, and context than others I've read. It is reassuring that some news organizations still value and produce proper journalism.
25
If it turns out that manufacturing or design defects contributed to the cause of this disaster, Boeing and any other responsible parts manufacturers better pay up, and pay big! I wouldn't fly on a Boeing 737 Max until this mess is cleaned up.
5
I very much doubt that there’s a line in Boeing’s autopilot software that looks like this:
if (angle-of-attack-indicator > threshold)
push-yoke(all-the-way-forward)
Does the system not have access to an air-speed indicator? How about an altimeter? Maybe some redundancy? Six journalists collaborated on this article, but it doesn’t look like they consulted anyone who’s a legitimate expert on aircraft automation, let alone someone who can comment on the system used by this plane. In time, we may yet learn what caused this plane to crash.
8
@Rick
Problem is, if the plane is diving, pushing the stick forward isn't what you want to do (unless your flying upside down) you want to pull the stick towards you. But that is much easier said than done. The heavier a plane is the heavier the controls are. Absent hydraulic assistance a pilot would be very hard pressed to one, get the plane off the runway, and two, just to keep it in the air, it would take several very strong men.
Essentially, the sensor, tells the pilot the attitude of the airplane, and it more than likely can't be overridden, it would have to be shut off, and when your screaming towards the water at over 400 MPH, you have a very short window to level off the plane, or face over speed, and you would rip the wings off the aircraft, due to the gravitational forces of trying to pull out of a dive. Which is why dive bombers of WWII had speed brakes on the wings, (flaps with holes) in order to keep the aircraft from over speed. Still the pilots face 8 times the force of gravity when pulling out of a dive, it's very hard to move when you are fighting g-forces.
6
Why does the plane's automation automatically push the nose down? Seems to me an audible warning would be much safer. Audible warnings are already used in the cockpit for many other dangerous conditions.
12
@Celeste I believe there is a warning first. The emergency system activates if the pilots don't correct the stall.
2
After reading most of the comments, here's my blame list, in order of blame (my last list was submitted unfinished by mistake):
1. Boeing Air for essentially test-flying its planes in flight with hundreds of passengers onboard. This is preposterous: (“……..Boeing this week issued a global bulletin advising pilots to follow its operations manual in such cases…..”)
2. Indonesia’s National Transportation Safety agencies for failing proper oversight of the airline industry
3. Lion Air should have grounded the plane after issues with the two previous flights
4. Lion Air may not have trained its pilots properly on the new equipment
5. Lion Air for creating a culture of fear, insecurity, and unaddressed safety concerns
6. The pilot should have refused to fly the suspect plane
7. The airline industry as a whole for lack of foresight and safety (“…….While Boeing and its European rival, Airbus, are producing planes as fast as they can, the number of experienced pilots, aircraft engineers, mechanics and even air safety regulators has lagged.”)
8. The FAA for lack of proper oversight (“……On Wednesday, the Federal Aviation Administration of the United States warned that erroneous data processed in the new, best-selling Max 8 jet could cause the plane to abruptly nose-dive.)
11
So the airline doesn't have enough money to do proper maintenance, train the crews, or fix recurring problems... but it has enough money to buy new aircraft after they crash them? I want to have a bank like that.
8
I have always thought that (like the cruise control on my car) an airplane would relinquish automated controls when a pilot manipulated the manual ones. Is it now a case of humans having to fight the AI because sensors are the authority? That's what pulling back on the stick with all your might sounds like — the pilot fighting the rear controls which are saying and doing otherwise.
14
November 9, 2018
My God there is enough systems aids in our advance technology engineering to avoid such killer crashes and all must demand all areas of confidence in the operations to never having to experience such air industry failure - only with transparent analysis and best efforts for conclusions would I ever take to the air again -for now on its a no go even if my business interest will stick to teleconferencing and materials by air freight.
jja
1
I fly a lot, but know nothing about how to fly a plane.
I drive a lot, an automatic, of course. But I did learn to drive on a manual transmission vehicle with a manual clutch back in the day (I'm that old), and can still drive a three-on-a tree as well as a stick shift from instinctive muscle memory.
Don't tell me that even owning an automatic transmission vehicle today that is complete with electronic blind spot alert on both sides, front distance warning and automatic braking, rear view camera, automatic low and high beam regulation, traction control, stay-in-lane assist, and heads-up dashboard display that I am not a better driver having learned the old fashioned way.
The same analogy should apply to all young pilots these days learning to fly only state-of-the-art models. OK, it might not have applied to this unfortunate young pilot given he was only at 5,000 feet and likely didn't have enough training on this model aircraft. But it just might...might...have made a difference of a few seconds, enough to pull up in time.
4
@bozoonthebus
A huge difference is that driving is a two dimensional world and hence much easier to make sense of something than if it's in three dimensions like flying.
5
Did I miss the announcement regarding the immediate grounding of the 737 Max aircraft?
4
Did I miss a reason why this should be done?
2
It would make more sense to find the actual cause of the crash before grounding an entire type of aircraft. The fact that similar problems occurred on earlier flights suggests a failure of maintenance.
5
@Mike- Seems as though there is a great deal of uncertainty regarding the root cause of the issue. Erring on the side of caution and not worrying about flight revenue seems wise to the flying public.
4
There are many comments here lamenting the loss of "stick and rudder" experience in pilots who fly modern airliners as a result of increased automation in the cockpit.
However, the statistical evidence shows that as automation has increased and pilots' roles have been reduced, flying has become more safe, not less.
So while a fully automated aircraft may not have given us The Miracle on The Hudson -- instead resulting in a crash with no survivors -- the increasing automation has saved far more flights (and lives) from pilot error than have been saved by human pilots.
3
@Celeste
You make an interesting claim, that automation has saved more lives than human pilots - do you have a link you can share with us? I know that is the claim of manufacturers, but I have never seen hard data to back that up.
2
@Celeste
And while I grumble when flights are delayed due to a mechanical, I am also well aware that these kinds of problems are what we are being protected against here in the US. Which is likely why it's been 10 years since a US disaster of this magnitude.
5
@Kim from Alaska- Don't grumble when they de-ice a few extra times also.
1
Normally articles about transportation failures dwell in the mode of blaming the pilot, ship captain, or train engineer. Fin.ally, we have an article by a journalist who analyzes from the way a trained investigator would think.
Interested readers would do well to look at the work of James Reason, the British psychologist. He found that in complex systems such as aircraft, nuclear power stations and other critical infrastructure, the systems are so complex that no one person can understand all of them. This holds true particularly with computerized systems that 'automatically self-correct.'
The factor of human intuition then becomes critical, if time and the fundamental laws allow the thought process.
9
After reading most of the comments, here's my blame list, in order of blame:
1) Lion Air should have grounded the plane after issues with the two previous flights
2) The pilot should have refused to fly the suspect plane
3)
1
It sounds like the suitcases in the luggage compartment probably went through more stringent product testing than did this Boeing jet. And/or weren't prematurely rushed to market. Boeing should be held responsible for rushing it's "workhorse" to market without sufficient testing, training materials and protocols to assure it's safety. And Lion Air - why is it still in business?
What caused the Titanic to sink? : a combination of 1) pilot incompetence/errors 2) corporate greed 3) engineering ignoring these factors - because engineers and the companies they work for may be 1) arrogantly blind to faults in their products, 2) greedy and 3) human factors or planning for human systems. Engineers are handmaidens to machines, and consistently overestimate their functionality. I say this having worked with and closely observed engineers in their attempts to design human/machine interfaces over the course of many years. Nothing involving machines and humans is foolproof, and we still design and operate systems as if perfection can be achieved. Witness: self-driving cars.
"The recommended response issued by Boeing and the F.A.A. this week would not be a pilot’s natural reaction. The flight crew is instructed to switch off the electricity powering stabilizers in the tail of the aircraft that are propelling the downward pitch of the nose. But without specific training on this anomaly, what pilot would think to turn off part of the plane?"
Ya, who woulda thunk?
3
Said it before and will say it again. I worked across Indonesia, took many flights knowing that the profound corruption in that country meant I was playing the odds about every plane's safety.
The NYT states that Lion Air is politically connected. That's all it takes in Indonesia; pay off the players and ignore the rules. If Lion Air pays the price for what will surely turn out to be it's criminal malfeasance, it won't be the politicians on its payroll who see that justice is done.
10
I don't see how it would be possible to correct this error if the plane was only at 5000 feet and most likely near full throttle.
The Qantas pilot was lucky to have altitude as well as the skills of a fighter pilot.
19
Planes shouldn't be built -- or allowed to fly -- with known anomolies. And this craft was only 3 months old! This is on Boeing, all the way.
6
@Radicalnormal
The carrier bears the burden of maintenance and not dispatching a plane with safety and mechanical issues.
4
I've been flying all sorts of aircraft for 54 years, including B737's. I came to the conclusion several years ago that automation is severely eroding pilot skills. Auto land, auto level off at a selected altitude, auto ILS to landing, climb speed select, etc. The list goes on and is constantly being added to. Captain Suneja should have disconnected everything, looked outside, and flown the airplane.
Unfortunately, his training is to rely on all that stuff that Boeing puts in there to make flying supposedly easier. Even our military is doing the same thing. My experience in 3rd generation fighters was the best thing in my long career. You learn how to actually fly a high performance airplane. I grieve for all those upcoming pilots that will not get a chance to actually fly a complicated airplane.
32
What a novelty that the way to recover from a possible failure of the angle of attack sensor is for the pilot to actually take control of the airplane! So, some think that in the future we will be happy to board an automated plane without pilots? Speaking for myself, not a chance.
10
@David Godinez You May not be aware that the automatic stall protection that is being implicated only occurs when the autopilot is off. The captain may have had two issues. Erroneous airspeed and erroneous AOA. If he was hand flying because of the airspeed isssue then the faulty AOA may have triggered the stall protection. The steps to disable that - once recognized - would have taken up to a minute. I don’t think this was recoverable and I feel the solution offered by Boeing is insufficient. Prior 737s didn’t have this and they are flying safely. I think Boeing should have the feature disabled until a better solution is found.
8
@Thane
Automated stall protection only works when autopilot is off? So you are saying that when the autopilot is on, the plane can stall and not automatically recover from it? That does not make sense to me.
And if stall protection only works when flying manually, this "protection" over-rides what the pilot is doing? That sounds crazy - it's like having two sets of hands on the controls, fighting each other. Seems like such a system should sound an alarm so that the pilot then makes the manual adjustment.
2
Wonderful array of sensors on this brand new max 8 jet.
Is there anything preventing this computerized marvel from double checking certain readings - oh - perhaps the plane is plummeting at 100+ m/s as read on GPS and/or altimeter - flash warning instead of giving full control to a stuck weather vane.
Another reason why I will never fly again. I drive - everywhere.
6
You surely know that driving is much, much more dangerous.
12
@Mike- And you are probably in a lot more danger while on the road. Each to his/her own.
6
@Mike
Do you ever want to visit Japan--or even Hawaii? Not sure how you'll drive to those places.
2
I will never, in a million years, fly on a third world airline. Ultimately, the proper care and maintenance of mechanical objects such as cars, boats and airplanes is rooted in a worldview that holds human life precious. I suspect that the pilot of this airplane was told to suppress any concerns he might have had after glancing through the logbook.
9
@Frank J Haydn since this jet was only 3 months old, I think some blame should also be put on the 1st World Manufacturer.
4
In 1999, NIST (National Institute Science and Technology, formerly National Bureau of Standards) changed its primary standard for air speed measurements from Pitot tubes to fiber optic laser Doppler anemometer (LDA)
More accurate, less prone to problems
At the minimum, and LDA should have been a back up on this brand new Boeing plane
https://www.nist.gov/sites/default/files/documents/calibrations/ncsl_055.pdf
5
Amateur here....wouldn't the plane's behavior, its response to controls, allow one to distinguish between a stall and other causes of unintended descent, such as a malfunctioning angle-of-attack sensor mentioned here? Also, shouldn't the pilots (or autopilot program) interpret the *altitude* reading, and other instruments, to mean that the plane was plunging toward the ground? A stall clearly cannot be recovered by a dive at low altitude. Didn't the Air France crash (pitot tube failure) highlight an effective response in such situations, something like high power and a nose-up attitude (which could be ascertained separate from the angle-of-attack). Thanks in advance for any info.
3
This is a problem we are going to have to address. Automated systems bring many benefits and as those benefits solve problems, they are moving from the financial & entertainment worlds to controlling critical real world systems. Yet they are only as good as their programming. A real weakness in our system is that the programming is written by people who do not know or understand the systems they are writing code for. I know it is reviewed by experts in the field, but they in turn do not know code. As we have seen (look at the current Windows 10 mess) even experts can produce buggy code. When this happens in real life, on critical systems, 189 people are killed.
I don't know how to solve this problem, but I see that we have a critical weakness in our systems, and people are getting killed as a result.
6
@Bruce1253
Just "buggy code"? You've never experienced computer glitches in your car, laptop, phone or any other system reliant on manufactured parts?
What amazes me is that people act as if we expect all technology to be infallible. Glitches do happen. Like lifeboats on the Titanic, a "Plan B" needs to happen if we are ever to feel half as secure as our horseback-riding ancestors who relied on creatures they had trained to be dependent on them. Frankenstein's error comes to mind.
1
@Quite Contrary
There is a difference between a 'glitch' and a coding error that kills 189 people. In the medical field they have something called a 'zero occurrence error'. Coding on critical systems will have to adopt a similar philosophy.
2
"They [the pilots] would have had to recognize that a problem with the readings on the cockpit display was causing the sudden descent. Then, according to the F.A.A., they would have had to grab physical control of the plane...Captain Suneja could have braced his feet on the dashboard and yanked the yoke, or control wheel, back with all his strength."
I really hope what is described above is not what happened.
1) You don't need a display to tell you you're in a sudden descent. You can feel it.
2) When you go into a sudden descent (especially at low altitude) the first thing you do is pull back on the yoke. It's an automatic reaction that a pilot shouldn't even need to think about. If you are driving a car and the road starts to turn, do you spend a couple seconds deliberating with you passenger as to whether you turn the stirring wheel? No, you turn the steering wheel without even thinking about it.
Either the pilots were so incompetent that they never even attempted to pull the plane up, or there was something wrong with the plane that prevented them from pulling it up. No sane pilot would sit around for 10-20 seconds contemplating their cockpit display and debating possible courses of action while their plane takes a dive from 5000 ft.
12
Pilot error is a lot easier for me to understand than a computer going rogue and a pilot left with the choice of putting their feet on the dashboard and pulling back on the yoke with all their might trying to wrestle back control of the aircraft.
1
@Chris Mchale As explained in previous posts it is almost impossible for a pilot to pull back on the yoke to overcome a malfunctioning elevator trim system.
2
This is my aircraft of choice. Need to figure this out right away Boeing!
2
Are planes are going to be computerised to a level when pilots are no longer in control of events?
If so maybe better they are flown by “Drone “Pilots/Computer Engineers with on-board pilots as “back-up?
The inventory of experienced military pilots who've shifted from combat to commercial flying is dwindling fast, and fast coming to an end. The number of commercial pilots needed is twice as many as the number of military pilots retiring.
But to say that "less-desirable" airlines are "scraping the bottom of the barrel" in terms of pilots dishonors men and women who have made a commitment to their field.
Less experienced? Yes. Bad people? No.
Just this last April, the Washington Post ran an article after Shultz, using her military training and experience, safely landed her plane (see URL below). Remember that Sullenberger with his military training and experience, safely landed his plane in the Hudson River.
For a number of valid reasons, there is a shortage of military trained pilots shifting from one service to the next. The military wants its highly trained and paid pilots to remain in its service.
As many times as I've flown, I've never known if the plane's pilot's flying history included military service, but I've always believed that a former military pilot was preferable to a simulatedly trained pilot.
https://www.washingtonpost.com/news/wonk/wp/2018/04/20/the-odds-youll-be-flown-by-a-former-fighter-pilot-like-southwests-shults-have-plummeted/?utm_term=.dc89d3681850
7
@LivingWithInterest
Before becoming an airline pilot, my father was a Navy pilot for 10 years based in Florida, Maine, Spain, Anarticia, & Bay Area. He loved it and said it was best training and experience that he never could have gotten anywhere else.
4
"Boeing this week issued a global bulletin advising pilots to follow its operations manual in such cases." How about fixing the problem so this does not Happen in the first place?
8
@frank monaco
No doubt they will but at this stage we’re still in investigation mode.
2
Right - “we’re plummeting toward the ground, let me get out my users manual.”
4
This is far from the first time an (almost) new passenger plane, built by one of the largest and most well known companies in the world, has had engineering and/or design flaws. Regardless of the pilots' decisions which were made, they are not alone. As Airbus was mentioned in the article, I would like to point out they had a serious design flaw upon delivery of their super jumbo jet, the A380. They were grounded temporarily, until the problem was finally narrowed down to a flaw in the Rolls Royce engines.
6
NTSB investigators should ask Boeing for the specific details of the software routines that have been programmed to correct an incipient stall without pilot intervention.
How many degrees of downward pitch is the automatic routine programmed to engage? For how many seconds is the automated system programmed to engage downward pitch before the automated system disengages and reverts to manual control by the pilot? Why does the system automatically perform a piloting function without input from or control by the pilot? Why did engineers determine that the previous safety precautions against stalls, such as stick shakers or warning sirens or "reduce angle of attack immediately" recordings were insufficient to address the problem?
It is, of course, possible that in designing this system, Boeing engineers may have been responding to NTSB recommendations following the Colgan Air crash, which resulted from a stall into terrain.
Nobody should doubt the good faith or competence of the Boeing engineers or NTSB officials trying to identify the causes of air crashes and develop safeguards for the future. No matter how well designed, no system is infallible. Automated systems, while often helpful, can make mistakes under circumstances not envisioned by the designers. Learning from accidents is a costly way to do things, but failing to do so would be even more costly, with the cost measured in human olives more than money.
7
The article focuses in depth on every possible cause but the most glaring one. How did a brand new plane from Boeing delivered only three months ago have so many serious technical and sensor issues ? And has the company reduced its adherence to its own legendary standards ?
9
When the airline owners are complicit in disregarding the aviation regulations and recommendations from experts, this will happen; and continue to happen. 20 years ago in China, I would never have flown within the country. Airline accidents were relatively frequent, regulations ignored, and the government complicit in cover ups. Sounds like Indonesia has some actions to take. Will they?
5
Go to a hardware store. Buy two bubble levels. Install both in the cockpit, one fore to aft, the other port to starboard.
4
@Rupert. Neither of the two would tell you anything about the plane’s angle of attack, which as the article makes quite clear, is the angle between the wing and the oncoming wind, or slipstream. Bubble levels tell you the angle of a surface with respect to gravity, not the slipstream.
7
'...strapping ourselves into a metal tube...'? With that many moving parts, that's exactly what it is.
Let's not forget that the final accident report will be written by the Indonesian ATSB. That's international law. Australian, American and Singaporean accident investigators are assisting the investigation - working on an 'advice only' basis.
It's very unusual for the assistants to object to the principal investigators conclusion particularly since their reach extends only to producing a non-legally binding 'minority report'. The last time this significantly occurred was with the dreadful 1977 Tenerife accident - still the worst in aviation history. Apart from the appalling loss of life, there were huge liability issues. The Spanish ATSB severely criticised both KLM and TWA whilst exonerating air traffic control. Both Dutch and American investigators violently disagreed, but, when it came to the payouts, the Spanish report was the one used by the courts.
3
Inventing planes that are so inherently unstable they cannot fly without complex sensors working conflicting algorithms might be a place to look.
2
Perhaps the airline industry, whether through legal action or otherwise, should allow passengers via a government TSA portal or similar, be able to review maintenance records from airplanes searchable by itinerary date? Such a portal could color code each airplane green, yellow, red for extremity of maintenance issues, and allow passengers to decide if they want to board or change a flight. Is this too much to ask for when we place our trust and lives in an airline and plane, when we pay for this risk...?
2
Yes, there are design flaws, or at least specific areas that require learning of new procedures with the new aircraft.
Yet there were also a series of blaring warning signs in this case which should have been addressed with an abundance of caution. And were not.
This is strikingly similar to the pattern of thinking and action that led to the Colombia shuttle disaster. We've seen it before.
The story would be different if all Airline companies' top management were required to fly on their own aircraft on a continually rotating basis. Then an incident such as this would induce sudden and major change in greedy corporate culture.
I would propose an international agreement that mandates the above: All top airline executives, including those at the very top, be required to fly (regular, not first class) on their own scheduled flights on a regular, rotating basis.
Then we would likely see a new level of import given to passenger safety and comfort.
16
Lion Air knows that only when an airplane is in the air they can make money. That's why even after problems of erratic speed they still allowed the plane to fly. The pilot obviously felt pressured after seeing the logs of prior problems. It's still better to be alive to fly with another company another day than to die on a questionable aircraft.
7
@Wayne each time the issue was addressed in a different way. This pilot thought it was fixed.
2
Can these planes even be flown analogue these days?
I for one am terrified that we are all flying in things that are just one computer glitch away from disaster.
19
That plane gave ample warning by almost crashing while flying into Jakarta - all staff and passengers departing that plane then should be interviewed.
3
More evidence, after the San Francisco crash of Asiana Flight 214, that artificially-intelligent airplanes are leading to undertrained flight crews, with catastrophic consequences.
The expedience-at-the-cost-of-safety mindset has implications not only for the passengers in vehicles assisted by AI, but public safety in general. It stems from persistent attempts to reduce prioritized problem-solving into binary code - to automate responsibility. Whether that feat will ever be possible is a question yet to be answered.
6
@BobMeinetz
Well said, Bob. With these flying computers we call airplanes, and the consequent implications you correctly enumerated, God save us all. Really scary.
2
This sure sounds luke a faulty plane to me that should be immediately recalled by Boeing.
5
The reporters who wrote this are to be recommended for their very informed and clear presentation of complicated technical and ethical issues.
19
@weekend with one caveat: they didn't mention how high the aircraft was when the problem occurred. Recovery would only be possible if the plane were very high.
@Morey
They did; 5,000 ft
4
@Morey 5,000 feet.
If you have ever wondered what an entirely robotic future would be like. This is a preview.
3
An excellent article, yet I feel it minimizes the role and responsibility of Boeing in this accident. Why has the fleet of 737MAX been grounded until the true cause has been identified and corrected? Seems it should go beyond a recommendation for the pilot to pull a fuse
4
@Peter Berner
Agree. I wondered the same thing while reading the article. If Boeing is selling planes that pilots don't know how to fly, we are in big trouble.
5
@Patricia
I get that but Boeing and Airbus sell planes. While they assist with training it is the responsibility of the purchasing air.ines and government regulatory agencies to ensure that pilots are skilled enough. When you buy a new car, GM or Toyota are responsible for telling about its features but it’s up to you use them properly and know how to drive.
4
Boeing issued a warning, not a recall. Check what aircraft you'll be flying on before you board.
6
If Lionair is such a poor airline, why is Boeing equipping them? I would think it's in Boeing's self-interest not to sell to carriers who do not properly manage and maintain their vehicles and could bring a bad name to Boeing.
On the other hand, how many of the 216 Max8s currently being flown are with highly-respected US and European carriers? Or, are the guys who "scrape the bottom of the barrel" the guinea pigs?
4
@Tough Call American and United both fly 737 Max8s and I'm pretty sure Southwest has both the biggest current fleet and the most 737 Max8s on order.
2
As always a story written by folks who do not fly, way too soon.
Patience we will find out why and correct it. its why flying is safer than ever now.
3
I’m a private pilot, licensed in 1986, before even little bug smasher planes had a lot of automation. I also drive a manual transmission car. What I perceive, flying with younger pilots who learned on automated systems, or driving an automatic, is that you are allowed to distance yourself from the moment-to-moment subtleties of both the machine and the environment. If something goes wrong, it takes measurable time to realize there’s a problem and to catch up with what it could be. That transition is dangerous. Automation is here to stay, but we need to find a way to keep the human operator fully engaged all the time if he/she is responsible for handling emergencies.
33
@RSS- Everyone is entitled to an opinion on this disaster, but I would like to point out that a "younger pilot"- my son, a 23- year old First Officer for a well-known domestic US airline, has been judged competent by the flight school granting him his private pilot's certificate, the University granting him a Bachelor's in Aviation Science, The FAA granting him his ATP Certificate and his employer who provided months of additional training and check rides. I believe, and many others agree, he has transitioned well.
3
Excellent article...
Now.. what is Boeing actually doing about this? testing sensors? double checking the computer's code?
And the airlines flying the MAX, are they training their pilots enough?
6
Perhaps Boeing should have a voice-controlled override that could automatically disengage computer systems so that when the pilot thinks it's malfunctioning manual control can be immediately returned to the pilots.
9
@Scott J. But what should the pilots have done once manual control was returned to them? If they weren't getting accurate information about the airspeed or angle of attack, then it was essentially impossible to fly the plane, either manually or automatically.
@T They could look out the window at the horizon, and set the power at 70 percent with 3 degrees nose-up on the attitude indicator to maintain level flight. Hopefully they also practice partial-panel when obtaining their instrument licenses. Good Grief. And these guys had ATP ratings?
@T - I'm a former Cessna pilot myself. If I wasn't sure about the instruments showing my horizon/altimeter/ etc. I was trained to disregard them and fly by 'seat of my pants' to the nearest airport.
An analogous situation in a Cessna would be if you had a drunk sitting in the co-pilot's seat and fighting you over the controls.
1
"Even as the politically connected company, which owns several airlines... "
Ahhh, I believe we've found the basis of the original problem:
Bad maintenance.
The pilots had at best approximately 9 seconds from start of the uncommanded dive until impact considering an altitude of 5000 ft. After the first few seconds correction probably was not possible to arrest the dive. These airplanes need to be immediately grounded and not fly again until investigation is complete to include review of technical risk assessment .
4
@PJR
Just my mental math...but ignoring air resistance, a free fall from 5000 ft would take between 17-18 seconds. However, in real life there is air resistance and the plane almost certainly did not go into a 90 deg dive straight into the ocean, so assuming the controls worked properly (they might not have) the pilots had plenty of time to address the issue.
There was a recent report out that the airline/aircraft industries are projected to need over half million trained and certified personnel from pilots to technicians for maintenance.
Asia and the Middle East showed the greatest need I believe.
Airlines already send planes to non-US countries with FAA certified facilities for inspection because it's cheaper, but the rate of FAA certified techs or those who are well versed in English is poor and the inspections questioned.
It has been pointed out that any FAA officials going to inspect such facilities in China or El Salvador say, must announce their plans and/or get visas...another layer of compromise and disinformation.
So what are the odds that a maintenance tech in Bali installed a sensor correctly?
If one sends a newly designed plane costing multi millions of dollars overseas perhaps it should come with more than manuals or apparently an i-app?
Perhaps a living, thinking person for a year, trained in practical measures, not just theory like the referenced bulletin for Superpilot?
Someone here wrote economics is the problem, sadly when one gets to the bottom line that's correct.
3
@Rosiepi
Not quite so simple - the FAA isn't the only aviation regulator in the world. Consequently, they all have to work together. The FAA reaches formal agreements with EASA, Japanese and Australian aviation authorities (for example) for mutual recognition of standards.
There are many hundreds of Airbus aircraft flying within the US, made mainly in France, but the certification of US Airbus repair stations is done by the FAA not the French DGSA. Similarly, in the UK, Boeing repair stations with FAR145 certification are regulated and inspected by the UK CAA.
It's quite possible that, by agreement, the FAA has been no-where near Lion Air.
2
Why on earth isn't this article laying the majority of the blame on *Boeing*?
This plane was brand new, and should not have had serious, no: fatally, faulty components on it. The blame is less with the pilot or the maintenance crew and 90% with the manufacturer.
I'd this was a brand new Airbus diving into the earth at high speed, it would have been grounded by the FAA, but it's Boeing, which means US stockholders losing out, so it won't happen.
15
@Narikin Indeed! Why is Boeing even selling planes to this airline, knowing full well that Lion does sub-par training and maintenance.
5
@Narikin
Surely we don't yet know where the blame lies?
A bit premature to be blaming Boeing and the FAA.
3
The design concept of "fail-safe" would preclude any automated system from sending an aircraft into a sudden high speed dive into terrain. This should never happen regardless of the condition of pitot tubes, angle of attack readings, or slack maintenance logs.
4
This is a case on over reliant on technology and insufficient pilot and maintenance crew training. The Angle of Attack sensor is a new technology being introduced to alert pilots that the plane is being flown at an angle that is too steep which may lead to the stalling of a plane (aerodynamically the plane is losing lift when it stalls due to slow speed). Coupled with an autopilot that automatically corrects for this situation, the plane is put into a nose dive which leads to a crash.
The lesson here is that pilot training is critical since on this case the pilot only had seconds to recognize the situation, take the plane off autopilot, disconnect a runaway tail stabilizer that is causing the nose dive and level off the plane. I very much doubt a low cost carrier would put the time and money to train their pilots adequately.
2
@Bruce WIt has already been established that pilots were pressured into flying even when issues had been highlighted. This plane should never have been cleared for a passenger-carrying flight, but it was.
5000ft in a matter of seconds, the one saving grace is that those poor passengers will have known precious little. Lion should be grounded and all planes thoroughly checked.
1
@Bruce W
I don't know where you're getting your info, but Angle of Attack sensors have been on airliners for decades and on military aircraft for ages.
1
@Bruce W
Bruce, a few questions:
1) Why does a plane on autopilot fly too steep in the first place?
2) Wouldn't there be warnings that a plane is approaching too steep an angle, alerting the pilot?
3) Why would the autopilot over-correct and send the plane into a nose dive?
Just trying to understand.
3
One must question whether the limited liability of airlines to pay damages under the Warsaw Convention makes it cheaper to take the risk of flying a questionable plane rather than lose income by grounding it.
Perhaps it is time to revisit the Warsaw Concention treaty, an international convention which regulates liability for international carriage of persons, luggage, or goods performed by commercial aircraft.
8
A substantial minimum required training time in a plane simulator to train pilots how to operate this new plane could have prevented this accident. If no minimum training time for pilots is required, this will happen again.
Automatic control "off switch" accessible to the pilot could have provided the pilot the control and time to save this plane. The evidence suggests that the pilot had about 60 seconds to correct. No time to get out a manual and look at options.
As someone experienced with computers, engineering, science and technology, risks to human life can and should be minimized. The aviation test should always be, "would you let your wife and kids fly on this plane?"
4
Perhaps we're reaching a tipping point in automation in which a so-called safety device can instead cause a crash. Makes one leery of the systems designed to let the cars that transport us drive themselves ... in the name of safety
8
Computers are solving real world problems all the time now. When the autopilot detects a possible stall on two sensors then we would expect the controlling program would check other sensors (gyros, altimeter, etc.) before sending the plane into a dive at low altitude.
In general as we cede more control to computers humans will be less inclined and lack experience to quickly regain control of planes, cars, industrial plants, etc.
Eventually computer control will most likely be as safe or safer than human control. But we will continue to be outraged when a computer (programmed by humans and/or "learning") makes the wrong decision and kills hundreds.
2
Sounds that angle of attack reaction loop is a potentially deadly safety feature. Pilots spend a good portion of their training on how to avoid and get out of stalls. Not sure that this new gadget is making their job any easier.
6
Maybe we need to allow pilots to have more, rather than less, control of the plane they fly? There is such a thing as too much technology and not enough human interaction.
21
“...possible shortcoming in the Boeing 737 Max 8 that could affect other fleets operating the popular jet.”
Interesting article with good reporting that seems, however, to put too much emphasis on pilot error and not enough on the sentence highlighted above.
If authorities wanted to, they could cut the investigation time of these crashes exponentially as to cause. Instead, the reports are issued many months or years later when only the victims’ families still remember the crash.
6
If, in fact, it is difficult or takes time to recover from an automation bug, the control system should weight the other sensor inputs, such as altitude, radar, maybe satellite to decide (very quickly) whether or not to take such an action. I’m guessing that the concept was sound, but incomplete in its structure. If a perceived low altitude stall response is a steep drive, there are no remaining options. In addition, the plane should have some memory of its own repair and instrument log, just as we humans are aware that, for example, our car has new tires or was making a noise yesterday. Unlike our car, the airliner has new drivers every day who might or might not know or be able to interpret recent issues.
9
i am no Luddite, either, but a couple of things should be painfully obvious. First, there has to be manual kill switch that will immediately shut down any autonomous vehicle's computer system. A big, red kill switch immediately accessible to the pilot or driver. No need to figure it out or go into HAL's CPU and pull a circuit board. Second, there has to be a non-digital back up system to allow for emergency manual control. Would these make the vehicles too expensive? If so, then a rational society would rethink the rush to autonomous vehicles. In this case, at least the pilots were trained and experienced. Consider the self-driving car, or worse, semi-truck, without a driver, or with a driver playing a video game, completely oblivious to the surroundings. Consider a nuclear power plant that has been hacked in spite of the claim that its systems are isolated, and the operators cannot scram the reactor because the manual override depends on some digital system that has also been hacked or disabled. We are in for a rude awakening. Last thought - let's ask Captain Sully or the pilot of the Qantas flight, both former US military pilots, if they could have controlled this plane with the erroneous readouts. Let's put them in a Max 8 simulator. If they can't control it, it should not be flying.
41
@Paul
Having a big red kill switch, as you describe, can have its own dangers. A few years ago, a big amusement park in Hong Kong found this out the hard way. An employee in a control room accidentally hit the emergency stop button for the park's train as it was running full tilt from one end of the park to the other. The train stopped almost instantly, people inside were thrown violently forward and an elderly passenger was hospitalized in serious condition. The amusement park, Ocean Park, responded by immediately installing a plastic cover on the stop button, so that it could not be accidentally pushed.
Boeing currently has two kill switches for each pilot, and both must be flipped by a pilot to turn off the power to electric motors connected to the stabilizers in the tail if the motors are operating in a dangerous manner. The kill switches also have covers that have to be opened before the switches are flipped.
It is possible there should be one switch with a single cover. But the Ocean Park example, a transport example although not in aviation, shows that no one wants it to be too easy to turn off or stop something important.
17
@Keith Bradsher The Ocean Park example doesn't apply here. A trained pilot is not comparable to an amusement park employee. If automation can put a plane in a death dive in seconds, as apparently happened here, two things need to be re-engineered: (1) the automation system, so that the chances of that happening are near-impossible, and (2) all pilots *heavily*trained in a very quick and very easy way to abort the death dive. If both (1) and (2) can't be achieved, it will almost certainly happen again. Since it happening again at any time is too soon, not implementing (1) and (2) means the plane shouldn't fly, period. That is how aeronautic safety should be addressed, not your concept.
Additionally, it's apparently likely that this new automation system is what proved fatal, and arose because of stall disasters like the Air France jet that "death-stalled" and fell miles into the Pacific a few years back. The dive would have aborted the death-stall. Get it?
9
The Boeing aircraft was delivered 3 months before-and already had identified serious issues on previous flights. How can the pilots be blamed?
It's like buying a car that at 70 miles an hour "suddenly" turns left-unless you "overcome it with your own strength" in seconds.
42
For me, Boeing is being given far too much of a pass here where too much of the post-tragedy blame game dialogue involves distractions centered almost exclusively on these poor pilots being asked to do near impossible five second problem-solving of quite possibly very shoddy Boeing technical equipment, and less than two months out of the Boeing build plant.
There are also barely 216 of these brand new technologically new-age planes in service now, all of them with less than a year and a half of flight hours logged. What we need to remember, sadly, is, that like automobiles, it never fails that there is nearly always a lemon or two that eventually emerge from the entire bunch. With automobiles however, you send it back to the manufacturer/seller for either a refund or a viable fix. Much more problematic with a $50+ million dollar aircraft if it is indeed a defective craft to begin with, even with proof.
RIP to the victims, including the pilots of Lion Air flight 610.
"The 737 MAX gained FAA certification on March 8, 2017.[8] It was approved by the EASA on March 27, 2017.[35] After completing 2,000 test flight hours and 180-minute ETOPS testing requiring 3,000 simulated flight cycles in April 2017, CFM International notified Boeing of a possible manufacturing quality issue with low pressure turbine (LPT) discs in LEAP-1B engines.[36] Boeing suspended 737 MAX flights on May 4,[9] and resumed flights on May 12.2017. "
Max 8 Wikipedia
14
@RLC: Your saying Boeing is being given far too much of a pass is highly credible.
Recollect the same lax oversight of FAA granting permission to use Lithium Ion on Boeing 787.
Boeing’s engineers, failed to consider and test for worst-case battery failures
The Federal Aviation Administration, that failed to recognise the potential hazard and did not require proper tests as part of its certification process.
And now this, make me compelled to recollect Hamlet quote Something is rotten in the state of Denmark
4
Whether airplanes, automobiles or home sensors, when we put too much confidence in our electronic systems, without reliable override, we do so at our own peril. "Open the pod-bay doors, HAL."
7
Sure a young, inexperienced, poorly trained pilot can be over whelmed by issues. And experienced one might have put the plane in manual mode and pulled up, not be confused in these circumstances as to where the ground was.
3
I am getting more frustrated with Boeing not doing enough to find the manufacturer side of the problem. All the focus so far has been on the airline which has a pretty bad safety record. But it is more important, and I’d say more crucial, to find out if there are design and manufacturing flaws and errors. Boeing must do more and fast because the delivery and operation of this new type of aircrafts are of serious concerns of air travel safety for millions of people around the world.
61
@Bob
Your shortsighting Boeing. No sooner than Lion Air crashed, Boeing was informed. With more than a handful of these Max 8 and 9's flying with other airlines, they're already in the loop as we speculate. While your concern is palpable, please consider the fact that these models are flying as we discuss this situation. You may consider the ones in the air are not having the exact same issues and every crew one board are well aware of this anomaly. Balanced with the fact that Lion Air has demonstrated less than professional handling of overall management, this may be the lynchpin of a perfectly good airplane flown into the sea by gross mismanagement. Boeing and the FAA have the responsibility to decide whether or not to ground this new model or allow it to fly with special emphasis to all flight crews of this pending issue.
6
@Bob
If the maintenance crew broke the sensor when they installed it that day, before this flight, how is it Boeings fault and what can Boeing do about it in the future? Ship a maintenance crew with every airplane? Unsafe airplanes should not be allowed to fly. This airplane had repeated safety violations and there was no history of detecting the cause and fixing it. It flew with known safety issues. The pilot and the company that pressured him to fly are to blame.
3
Understood that aircraft, through possible faulty sensor information, could result in putting it into this “death dive”. But why can’t some ultimate self preservation over ruling system kick in that senses “plane is hurtling at near 400 MPH with its nose pointed towards the earth” and over ride the erroneous automated trim correction to allow the plane to pull out of the dive? You’d think they could instill something into the avionics that would assess what was would be the ultimate outcome given the aircrafts immediate situation (nose pointed at the earth) and level out the plane
17
@George Warren
Unfortunately, human "self preservation" systems (stimulus-response) often work much more slowly than do automatic (electro-mechanical) ones.
5
@George Warren
The only way the system can sense anything is by evaluating the information it's getting from things like the airspeed sensor and the angle of attack sensor. If those sensors are sending inaccurate information, the system has no way of sensing that the "plane is hurtling at near 400 MPH with its nose pointed towards the earth".
2
Airliners, unlike most drones use analog instruments as their primary means of measuring altitude, airspeed and angle of attack. The 3 “A’s” of flight.
Analog systems have limitations and situations where they are capable of providing deadly inaccurate information to the pilot flying the aircraft.
Today’s aircraft fly primarily by feeding the inputs of these same analog systems and the human pilots into a computer network that judges whether the human is doing the “safe thing” to fly the aircraft. Most of the time both the computer system and the Human agree on the chosen course of action.
“Most of the time” is not the same as “all of the time”! There have been a number of crashes where the computer received bad information from its analog instruments and made or caused the pilots to make fatal mastakes.
When you get something in your eye you know that and try to remove it. When an aircraft gets ice or a insect stuck in its pitot tube it has no way to recognize that there is a problem past comparing its information to a 2nd system. Digital systems suffer unique interference problems. Something as simple as a loose or missing ground strap can cause Problems that are undetectable on the ground but cause the computer flying the aircraft to receive bad information from all or part of its inputs when in flight.
This can and has caused life threatening problems, sometimes so suddenly that the pilots lack the time to correct the situation. This could be such a case.
7
This brings to mind a Sept 1994 crash of a 737 near Pittsburgh, PA killing all aboard. This crash was determined to be caused by a defective hydraulic rudder control which commanded the aircraft's rudder to put the craft into a dive. Previous to this crash there were some similar situations from which he pilots were able to recover. see
https://en.wikipedia.org/wiki/Boeing_737_rudder_issues
7
The whole idea of more automation for flying is to reduce the need for highly trained and highly paid pilots.
18
I flew Lion Air several times on a trip to Indonesia a few years ago. One of the planes I flew on from Jakarta to Bali was extremely old, a Boeing 727 or 737. Some of the overhead bins did not shut correctly, and constantly rattled throughout the flight. Some of the interior side panels of the plane were not attached correctly to each other, and were coming apart. The carpet was coming undone, and was completely filthy with stains and tears, rips. Many of the seats were torn and ripped. If the interior of the plane looked like this, what did the mechanical parts of the plane look like ?
34
@EH
I had the exact same experience on a flight from Bali to Jakarta. It was an extremely bumpy ride, too, with many rapid changes in altitude.
My future wife and I actually ate the cost of the return tickets to Bali, and purchased a flight on another airline, rather than fly with them again. The only time I ever felt like I was flying in an unsafe aircraft in all my years of travel. Not sure if they still use this in their advertising, but their slogan at the time was “Flying is Cheap!”
13
@EH
While that aircraft may have looked old and battered, Lion Air also has a lot of new ones. They have been a big buyer of brand-new Boeings.
1
Well, this aircraft was less than 3 months old, so, probably pretty new.
What if pilots were required to hand fly aircraft below 10,000ft
and only use Autoland when visibility is too low?
Why use any feature of an auto flight control system at 5000ft?
Why not when closer to cruise level where you have more altitude to deal with technical issues?
10
@juliet lima victorjust to add: it only takes a 737 five minutes to climb to 10,000 ft!
If their cruise altitude was only 5-6000ft, they could have hand flown the short hop.
2
@juliet lima victor
Although automated systems aren't perfect, planes today are much safer than they were in the days before automation. Having pilots hand fly aircraft would result in an increase in crashes.
1
@juliet lima victor
You seem knowledgeable about flying and aircraft control systems. Spend the time to learn what it takes to become an airline pilot, spend years at acquiring experience then answer the question. You may be surprised how complex your questions are with even more complex answers....
1
Let's not give boeing and their new plane a pass and blame it all on poor maintenance, poor training, etc. If it was a US based carrier, I am sure we would give all factors an equal consideration till the investigation is complete. Let's keep our prejudices aside and hope for an impartial, thorough investigation to determine all possible causes.
70
In chemical and power plants, instruments that feed critical safety controls are required to have "two out of three voting" logic. That means that even if a single instrument fails, there's redundancy and the control system doesn't rely on a single reading. The logic is that instruments are very inexpensive, have a small but significant failure rate and the consequences of failure are highly significant.
IF the cause of this accident was indeed a failed pitot tube (or other such instrument), its amazing this was not part of a 2 out of 3 voting logic. Boeing was trying to make the plane more safe by automating the reaction to stall (identified in the article as the leading cause of airline failures). In doing so, it created additional failure routes. Seems that Boeing (and I assume Airbus) need to improve their quantitative risk analysis procedures - which is a standard engineering method.
All responsible chemical plant owners require quantitative risk analysis completion prior to plant acceptance. I wonder if airlines have the same requirement. Is that one of the differences between the majors and the low cost airlines?
15
@deedubs Boeing and Airbus aircraft do have redundant instruments. Depending on the system, as many as 4 (hydraulics) and always 2 to three. For example the airplanes have multiple angle of attack sensors and pitot tubes, One set gauges for the Captain, and another for the copilot, plus a standby gauge in the center of the console. The problem sounds very complex. Air New Zealand lost an A320 over the Mediterranean a few years ago with a similar pitch problem at low altitude. The issue was the airplane had been washed with a fire hose and water had gotten into the angle of attack sensors, and then froze at 10000 feet. Both failed at the same time.
12
@deedubs
How would you explain Chernobyl? Union Carbide Bhopal? Three Mile Island?
Your assumptions of a standard engineering method falls short. Flight management systems evolved into (my knowledge is limited) two schools of thought. Automation ruling against the pilots and pilots over ruling automation. Airbus decided to have automation overrule pilot commands. American manufacturers allow pilots to overrule automation. This was brought up when American fight 587 crashed over Howard Beach, N.Y. It's more complicated than an engineering exercise.....
3
@fdryer. Chernobyl was done by deliberately bypassing the safeguards. Bhopal was the accident that made this all important in the first place. Like it or not, risk analysis techniques we're found out the hard way, so the reason they could not stop those accidents, was because they weren't even there and the first place!
1
There have been conflicts between computers and pilots that have had devastating results. This technology is making its way into cars too. A decision needs to be made as to who has the final decision-making authority , the captain or the computer?
10
@Buoy Duncan
All that an automated system has to do is prove that it is an order of magnitude better at resolving a problem than its human counterpart.
Computers have saved numerous lives in aviation and in automobiles (accident avoidance systems) and so far they are proving just as successful in self-driving cars
No system is ever going to be perfect including an aircraft AC.
3
@Buoy Duncan
All that an automated system has to do is prove that it is an order of magnitude better at resolving a problem than its human counterpart - not perfect.
Computers have saved numerous lives in aviation and in automobiles (accident avoidance systems) and so far they are proving just as successful in self-driving cars
No system is ever going to be perfect including an aircraft AC.
As an engineer, the theory that the faulty angle of attack alone caused the plane to nosedive until its demise is suspect. When designing complex control systems, engineers use a process called FMEA (failure mode engineering analysis) or equivalent, with the purpose of making the system robust to any possible system malfunction. It would be mind-boggling if Boeing engineers would allow a control system to nose dive a jet going ~400mph, knowing it's altitude of 5000 ft. If you do the quick math, that's about 17 seconds to the ground at a -45 degree AOA. No one in their right mind would ever design a control that thinks a plane is stalling knowing you have two healthy engines propelling an aircraft at that speed and at that altitude rate of change. A system that is critical to safety, has to have redundances in place for actuation and control. There is more to this crash than just blaming a faulty sensor and a computer gone HAL.
55
@NeoAce Air New Zealand lost an A320 a few years ago with a similar pitch problem. When the computers alerted the crew to take manual control of the aircraft because they could not reconcile the discrepancy, for some reason the crew did not, and did not manually adjust the pitch. All were lost. It seems the problem may be when does a crew switch off the automation and manually take control? A complex problem with so much automation. And redundancies.
5
@TVM So poor pilots is the real reason for the crash, put say Sully there and folks are scared but alive.
1
@NeoAce
My thoughts exactly. It would be interesting to see the FMEA and what considerations were given to low altitude multiple input failures.
1
No doubt Lion Air is a little sketchy. But you also unfairly tarnish them and its pilot.
The article itself identifies an Airbus flight to Australia where the US pilot had to “fight” before “finally” bringing the aircraft back under control after two automated dives. They had the benefit of altitude. This Lion Air flight was operating at only 5000 ft.
47
@Jay Lincoln
I wondered why a US pilot was flying for Australia's national airline.
Aussies have been flying around their huge country since the Wright brothers. You'd have thought they'd have enough pilots of their own.
1
@nolongeradoc
The Aussies are actually trying to recruit American citizens to fly in their Air Force. If you are a prior US military pilot, they will put you through a quick training program, and within a year or two you're sitting in an Australian F-18. I assume they fast track your Australian citizenship as well.
1
@guwinster
Sitting here in Brexit Britain, the idea of Australian citizenship sounds very attractive.
Pity, I can't fly an F-18.
1
“Captain Suneja could have braced his feet on the dashboard and yanked the yoke, or control wheel, back with all his strength.”
Really?
This looks like a bad catastrophe B movie. My old car has assisted steering, perhaps aircraft manufacturers should give that a try.
11
These systems should warn the pilot to correct a problem, rather than correct the problem for the pilot.
27
@Dana Charbonneau — Unfortunately a lot of crashes have happened after pilots ignored stall warnings, reacted too late, or reacted incorrectly.
1
Self-driving car proponents, take note. The same principle holds true on the road as in the air: in the aggregate, these highly automated systems make "the roads" or "the air" safer. But when something goes wrong, it goes really wrong, and you're plunging off a cliff... through no fault of your own!
35
As when you try to turn into your driveway and the car wants to keep you on the road, etc.
As at fault as Lion air may be, it seems to me that there is a bit of an attempt in this article to shift the blame away from Boeing, which after all may have created this situation to begin with.
45
I have wondered about the F35's flight systems since the plane was first introduced. From the reports I have read, this plane is controlled by something like 20 million lines of code. Even accounting programs have programming errors, and this is a combat aircraft. This relegates the pilot to the role of test crash dummy.
Back to Lion Air. From the early reports, it seems that the pilots and the crew were also test crash dummies. Pilots shouldn't have to fight their aircraft for control.
13
If the reports are true that the FAA warned that erroneous data processed could cause the plane to abruptly nose-dive, it’s in credulous that the agency has not directed the grounding of the aircraft, till the problem is resolved.
Instead it advice pilot in command on noticing the issue to manually take control the aircraft diving at 400 knots, meaning “braced his feet on the dashboard and yanked the yoke, or control wheel, back with all his strength or undertaken a four-step process to shut off power to electric motors in the aircraft’s tail that were wrongly causing the plane’s nose to pitch downward”.
Though a lay person I know a pilot is a human being with an average strength. Unless pilot happen to be a Herculean it’s well nearly impossible to even to move a finger when the aircraft is sharply diving at 400 knot speed with multiple G forces at play.
Pilots are no Hercules nor do they wear anti G suits while in flight.
Seems that the issue is serious enough to warrant FAA mandate immediate grounding of the aircraft till the issue is sorted out.
31
Reading once again that pilots aboard a doomed craft had been in radio contact, in this instance to request clearance to return, raises the question of procedures.
While black boxes can help determine many things, and the cockpit voice recorder may hold some information, the logical changes in rules would also encourage if not mandate that when things seem to be going wrong that the pilots radio what their concerns are. But seldom does this happen.
It should be required that this extraordinary vital information be communicated, not to be an argument to have a request considered or overruled by ground controllers, but to be additional information that could aid investigators in the unlikely but real event that a flight won't make it back.
In this case the request to return could have been made with additional statements as to the aircraft being difficult to control, contradictory data being displayed, engine malfunction and so on.
Clearly there is something getting the attention of the crew, and the lack of further communication is always a baffling part of stories like these. Guidance from those last minutes to allow investigators to possibly restrict other aircraft of the same type may save other lives.
3
@reid
The mantra is;
Aviate, Navigate, Communicate.
You fly the airplane first and above all else.
5
@reid
Are you familiar with the concept of "task saturated"?
I suspect the pilots were very busy trying to figure out what was going on a struggling to control the aircraft.
Aviate, navigate, communicate - in that order.
3
@reid
Aviate, navigate, communicate. It’s the last one on the list. Other than the fact that they needed to turn around, no one he was talking to could have helped him fly the plane. ATC isn’t going to tell him “put your feet on the dashboard and yank the yoke back as hard as you can.”
“Communicate” is last on the list for a reason...because in the moment it’s the least important of the 3. Sadly, that’s what the CVR is for.
3
A really good, well-written article. Sad and scary story.
10
Seems something as basic as a kill switch might have prevented this unnecessary crash. Appears from reports so far the 737 reacted to faulty data and the pilots tried to solved the problem. A kill switch, pure manual control by pilot, would have aloud them to fly the plane to higher altitudes and broadcast a mayday signal. Simplicity is something airplane manufactures don't pay much attention to.
9
From the descriptions in this article, this model should be grounded here in the USA. I'm not interested in flying on the 737 max.
9
It seems now the investigation may ultimately place blame not on the maintenance procedures of the airline, but on the new and perhaps overly automated airplane...
9
@GeorgeS, as a student pilot with some hours under my belt, you’re referring to go old fashioned pilotage.
My ground school instructors and flight instructor drills this concept into our heads with every class.
Computers can be an aid but they can’t replace a seasoned pilot’s knowledge and skill.
10
After reading this, I just can’t wait for fully autonomous cars.
14
Me? I'm not flying any new-fangled aircraft until it establishes a track record of having no problems and flight personnel familiar with the plane's characteristics and quirks.
2
@MIKEinNYC
I understand your concerns but how are you going to make sure that you don't fly any "newfangled
aircraft. Most of us book our flights well in advance, when it is difficult to find out which model will be assigned on any particular flight. That's especially true of airlines that fly multiple variants of any particular plane. Southwest's fleet is made up of nothing but 737's, but how can you tell a month in advance which variant of the 737 you will be flying?
8
@R.F. Southwest tells you exactly what plane is scheduled to fly the route when you books. For example, on Memorial Day weekend, I'm flying to 737-700s to and from PDX for the weekend.
That said, these can change at the last minute. The MAX is still a pretty rare bird in the southwest stable, so it would be unlikely to be replaced a -700 unless it was a really last minute emergency as they're needed on certain long routes.
Most other airlines will tell you as well, though some might be a bit more cryptic, like 772 means a 777-200. All subject to change.
2
Is this really a fair comment?
“the trust we display each time we strap on seatbelts and take to the skies in a metal tube”
Just about anything we do comes with a greater risk
The article does not clearly explain the difference between the horizontal stabilizer and the elevator.
The elevator is the small movable surface on the trailing edge of the horizontal stabilizer. The stabilizer is the big 'wing" on the tail. The stabilizer is moved up and down by the trim system and is many times more effective than the elevator which is moved by the pilots by using the control column.
It takes a great deal of force and strength to overcome an "out of trim" stabilizer using the elevator and at some point it becomes impossible to overcome it.
11
I am not an aerospace engineer, but why would an airplane have the capability for an "automated nose dive"? In what situation would a pilot want his airplane to descend at an angle greater than 30 degrees? or 45 degrees? These are commercial jetliners, not fighter jets.
16
@Scott
I forget which one, but in the not too distant past, a similar incident occurred where one of the airspeed sensors essentially froze and told the pilot they weren’t moving forward anymore (even though they were). The pilot spent minutes in a flat stall because he was trying to pull up to gain altitude. They basically did a belly flop into the ocean. IN THEORY a (properly functioning) sensor would have fixed this and pushed the nose down. In that case, had the pilot just taken his hands off the yoke physics would have corrected the problem. But yes, a computer handing you a 30+ degree pitch angle at 5,000 feet ain’t gonna do you no favors.
4
@Turgon That was Air France 447 in 2009, and it was doomed once it descended through 13,000 ft. Even the best pilots in the world could not have saved themselves once they were below 13,000 feet.
1
For citizens who don’t think an enemy hacking our computer systems will do much more than interrupt a computer game of solitaire or cause their lightbulbs to flicker, this article should provoke deeper consideration of the potential consequences. There is a lot at stake if we allow vulnerability to persist.
5
Who's the Pilot in Command, the computer? Parts fail, seemingly more often of late than people, so if there is not one readily accessible switch that completely turns of the autopilot, there ought to be.
And the notion that the most frequent cause of an accident is a stall might be true for a low-time pilot, but for an Air Transport Pilot with thousands of hours of flight time? Aircraft also provide various warnings to the pilot that a stall is imminent so that the PILOT can take action before a full stall occurs. Isn't that enough? I wouldn't want my autopilot pushing my nose down beyond my ability to override it knowing that stall indicators are not reliable.
16
@Ronald Aaronson People "fail" all the time, see the recent tragic evidence in California on the front page.
1
@vulcanalex Well, we are talking pilots vs. airplane systems. An airplane system will never be able to compensate for a truly "failing" pilot. I would just like a pilot to be able to overcome a truly failing airplane system.
1
@Ronald Aaronson — Experienced pilots have ignored stall warnings (and terrain warnings as well).
1
This is a bit like Westworld -- in the air. A 'highly automated' plane malfunctions and suddenly 140 people are dead. Perhaps the solution would be to give the pilot's more control, not less. Why do we always assume machines can do the job better?
22
“But without specific training on this anomaly, what pilot would think to turn off part of the plane? “
This article refers to training as some sort of optional and obscure requirement. All planes are not the same. Training for a specific aircraft is required. Training for emergencies is required. Boeing’s largest fault in this may be having agreed to sell an aircraft to Lion Air.
34
Runaway stabilizer trim is one of those problems every pilot of a plane with electric trim is aware of; it would be pretty obvious what the issue was almost immediately to any competent flight crew. And the solution is indeed to cut power to the electric trim, which in most aircraft means pulling a fuse. In this case it would mean manually overriding the trim (the trim whee spins forward, you just physically grab it and push it the other way and hold it there with your hand) until you could then turn it off. Takeoff would be a pretty rotten time for this to happen, but the fact that they were airborne for 15 minutes means they had some control of the plane. To me it says they fought the problem without fully realizing what it was nor how to solve it.
My gut feeling is training: if you had 15 minutes to try and figure out a problem like this, runaway trim should have been pretty high up on the list of potential causes. So why wasn’t the problem identified or addressed? Or perhaps they took the proper action but there is indeed a problem with the aircraft, ie the checklist doesn’t solve the problem.
My guess is that this will turn out to indeed be an issue with the plane, but one that a properly-trained crew could have handled. And like other procedures in aviation, the correction and training plan written for future crews will be written in the blood of the passengers of this flight. So it goes in aviation.
80
As a aviation professional myself and a pilot I like you I'm a little dumbfounded that they could not handle what was basically a runaway trim scenario though I know it was potentially caused by the flight control computers actions.
If I had been handed that aircrafts logbook with all the previous write ups I would have had a long conversation with the local mechanics and probably maintenance control. I would have wanted to know why they thought it was fixed and why, then I would have reviewed procedures with my First officer of what we would need to do if it was not fixed. I'm guessing none of this happened.
13
@Chris There do seem to be some allegations of two log books, which if true, means they may know have known about the prior issues. If true, people should be in prison for a long time.
1
You report that the plane was assembled in Washington state. What I’d like to know is where did the parts originate, especially the mechanisms responsible for the downward nosedive. The quality of each of the parts of the whole is vital for a successful flight. Lion Air should be put out of business for their negligence.
20
The passengers would in all probability be alive today if economics had not been the determining factor in giving that plane clearance for take-off.
Lion Air according to the article had at least 15 major safety lapses. Major lapses. Sending passengers into the air at 450 mph in an aluminum tube on anything less than the best safety record a company can maintain is tantamount to manslaughter.
They should be sued out of business and/or prosecuted criminally.
121
@Len
Well what is more important to a capitalist owner, profits or safety? Please don't take too much time coming up with the answer.
4
My father flew a Liberator during WWII - he said the majority of accidents were attributable to pilot error. He would say a multitude of issues would arise during flight - mostly takeoff and landing. Inevitably, it was always pilot error. I would imagine there are issues with today’s planes as well - how pilots deal with those issues is the difference.
5
@JMS First, I thank your father for his service.
While pilot error is a frequent cause, today’s planes are highly automated and frequently do their own thing. So while flying a Liberator - with possible battle damage - would be a difficult thing, imagine flying a plane without battle damage doing one thing while the pilot is trying to do another. Then multiply that by 1,000 given the number of automated systems. And all at 5,000 feet. I’m not sure if the pilots did the right thing, they’d even have the time to correct it.
25
I think @Paul has the idea right. Older planes allowed pilots more control actually demanded it. For the pilot to be dealing with instruments spewing misleading information, and an automatic system doing corrections which are throwing the trajectory off, that pilot would have had to have concluded that everything but his eyes were misleading him - immediately.
If this turns out to be what they find, maybe there should be a simpler way to opt out of the computerized system when necessary. Training. of curse, but perhaps the systems were too smart for human operation in less than perfect situations...
3
A great report and excellent sketches on AOA.
Thanks.
I lived in Indonesia over a decade in mid 80's and 90's and flew extensively - but never flew any of these low cost airlines. And they were propping up every other day.
I still remember an incident at Meean airport when a really old looking airplane belonging to a low cost - was being serviced. A ladder took a mecanic through cockpit windows - we all wondered what was going on.
Garuda - it's national airline - with a checked safety record - was the safest choice. But I never flew Garuda outside Indonesia because it was not permitted to many foreign destinations. However, it's in flight service was always excellent.
If I remember right - even Singapore had stopped Garuda for some time.
The scenario you described in the final minute of this doomed flight - simply frightening - and I love to fly.
My heart goes out to these victims and this young pilot and come pilot.
God bless.
36
We’re the crew flying in IMC (cloud) which is fundamental to understanding whether they could visually make sense of what was going on.
8
@David Pearn Even in IMC there should have been other instruments in the cockpit revealing the problem. There should be a simple procedure for disabling automated control, aircraft are just not that hard to fly. Of course we may learn of another issue as the investigation continues.
3
@David Pearn
It was a clear day.
2
@David Pearn
According to reports, the skies were clear.
1
”In the Max 8, if the data indicates the nose is too high, the aircraft’s systems will automatically pull the nose down.”
Boeing (and Airbus) makes great airplanes, and modern aviation technology has been a boon to safety. But, as with cars, we are moving from computerized helpers to HAL type devices which can decide on their own to take significant actions with the flight controls without, or even despite, the actions of the pilots. For all of the progress we’ve made and our faith in how clever we are to create these systems, the fact remains that at times our oh so clever computers, created by flawed humans, can be just as flawed themselves - and those flaws can prove lethal, whether running down pedestrians in autonomous cars or forcing a plane to dive into the ocean.
I, for one, prefer human agency to be able to overrule a controlling computer in safety matters. We also need to reassess how willing we are to risk lives, ours and innocent others, in the adulation of ever more aggressive automation. We are going to pay an ever higher price if we make the wrong choices.
125
@George S In the long (and even medium) term, computers will always be better at performing an automatable task than a human - better doesn't mean perfect, just better on average.
We don't know what exactly happened here, but until we also consider how many accidents and lives have been saved because we had computers to rely on to make the decisions and balance the two, that sometimes automation will include flaws in the automating process itself.
I'm okay with any decision that is based on evidence. Are you really saying that if evidence shows that 1/10 deaths due to flight accidents are caused by an automation error, and 9/10 crashes are caused by human error, you'd still prefer a human? How about 8/10 caused by human error? 7/10? 6/10 in a statistically significant manner is enough for me to go with automation.
I'll also say that if it's even that humans and automation are both prone to causing an equal amount of deaths, I'd go with the humans (for now).
43
Very well said. Manual control is a must. Pilots must have complete training that builds discipline adds confidence in their knowledge, skills and abilities. When the flight systems fail the human hand takes over. This is true crash prevention.
22
@RamS
I am not a Luddite and I embrace the many advances automation has wrought. But as a society we are coming to be too dependent on them and/or believing that it is inherently perfect. In this instance - and acknowledging that it is still too early to know precisely what happened - we must examine not just the fact that the computer may, under certain conditions, react to its perception of flight (which may or may not be correct) but the degree of autonomy the system possesses. Does it warn first or act in its own? Can it be readily overridden or restrained without going through multiple steps?
There have been other similar issues 8 recent years. The Air France crash off of Brazil with faulty attitude data, and a computer system that gave misleading feedback to the pilots, precluding their saving the plane. A Qantas A380 engine failure near Singapore where certain automated actions interfered with the pilots efforts and, at the end, which wouldn’t allow engine shutdown.
Technology is great, but we should master it, not the other way around!
24