That Mr. Wise is a private pilot and has written a book about MH370 indicates he has an abiding fascination with flying. We have that in common.
But both his stated qualifications and the content of his article betray some lack of expertise and perspective.
There seems to be nothing that inflames emotions and reactions more than an airline crash, with the exception of two in a row. Data on the second crash in particular is entirely inadequate at this point, but it’s quite likely that both crashes were easily preventable without and prior to any changes to the aircraft systems.
The 737 MAX system issues and the safety questions they present were and are easily and completely addressed, pending permanent software / engineering changes, with simple procedural changes for pilots. And after the first accident, most 737 MAX pilots were probably already employing these slight procedural changes on their own initiative even without a regulatory mandate. Accordingly, I believe the worldwide grounding of the airplane and the disruption it has caused was an unfortunate and unnecessary overreaction.
The NY Times is one of the two or three best news outlets on the planet, but its coverage of these accidents and related matters seem to me as though they should be in a Rupert Murdoch publication.
5
There is something Boeing is not copping to on these Max planes. "We are upgrading the software"....Gee Thanks for that. I wouldn't get on this type of plane even if my life depended on it.
5
still the only way to prevent this kind of accident is to have pilots use autopilot system as little as possible, and use it as a backup.
even Lebron James not playing basketball for a year, when he steps into a game against Houston Rockets, he is going to miss all kinds of shots...
Nothing beats practice.
4
Promoting and selling an aircraft which is a radical modification of an earlier model without cautioning the customer about all changes to its operating systems and flight characteristics, and emphasizing when additional training using online and classroom presentation, simulator, and actual flight training with a type-certified check pilot is required, is an inexcusably irresponsible act. It becomes more so when it is done, as in the 738 MAX case, as a key part of a sales strategy to convince the airlines they could avoid the cost of retraining and recertifying their 737cockpit crews in an effort to establish a compelling edge against Airbus competition for replacement of aging 737s and A320s.
Boeing certainly should have been aware of the role that muscle memory of experienced piston engine airline pilots played in the disastrous introduction of the 727 series in the early 1960s. In that case, the longer spool-up times of jet engines and the instinctive reactions by piston-trained pilots when unexpectedly high sink rates were encountered during the final stages of approach resulted in deadly landings short of the runway threshold.
Boeing had an obligation to fully inform the FAA, the airlines, and their pilots about any basic change in override from automatic to manual control protocols and the need for intensive training to ensure their muscle memories have been reprogrammed to accommodate the radical changes presented by the MCAS system. Their failure is indefensible.
18
Listening to some representative for the FAA on the radio yesterday, the gentleman was explaining why this plane was not "grounded" sooner. "We're a fact based agency," he repeated several times. Well, I got a fact for ya. Two planes crashed, killing everyone. You can find out "why" just as easily and as soon with the planes on the ground as in the air. Unless you are waiting for a third.
9
Dear NYT:
It would serve the public interest, and the investor interest, to understand the Boeing corporate economic efficiency drivers in the decision process that has led to these failures.
Sounds like profitability was put above public safety.
Please consider this side of the story in your reporting.
MB
12
"MCAS is designed to automatically reduce the pitch in manual flight without pilot input. The system is constantly fed data from two synchronized wing-like two synchronized wing-like devices called Angle of Attack sensors, located on the plane." Source: Yahoo News, Boeing, FAA, USA Today
If the above statement is correct, how does turning off the auto pilot correct the problem?
4
It makes no sense to speculate further, since the altitude and rate of ascent data that have been shown to the public are incomplete. The last 3 min of the flight are kept secret.
Why?
It should be obvious that those are the most informative. Releasing only incomplete data reeks of manipulation.
What is being hidden?
9
Training for contingencies is often insufficient unless continual - almost daily. Those who've been trained can forget quickly, even when the adrenalin isn't pumping. As for software upgrades, it took Microsoft years before they finally solved the old 'blue screen of death' problem. And despite many software upgrades to my 2013 auto, the manufacturer still hasn't been able to solve the fundamental transmission problem (but at least the car won't fall from out of the sky).
8
U.S. pilots have reported problems with the 737 Max descending rapidly upon application of the autopilot. The MCAS system operates when the autopilot is off. There is no reported mode where an automated system cannot exert control. If there is a component failure or a software problem, that can negatively influence both systems (autopilot and MCAS) the pilots are in a Catch-22. Their only option is manual disarming of the rear trim, and manual manipulation of it, if that is sufficient.
It could also be a component failure effects one system, and a software problem the other, and combine.
8
The article headline seems to blame "autopilot" systems. However, the article rightly explains that automation is good for the aerospace industry, and has decimated accident rates over decades.
"Autopilot" systems have advanced beyond what humans are capable of. If one looks at SpaceX and their successes, they have built software-based autopilot systems that can do amazing things. One system can land a rocket, "backwards", onto a floating ship in the ocean. They have tested this numerous times and is considered safe and reliable. Last week another of their autopilot software was able to perfectly dock their space capsule with the International Space Station - with no human intervention. Testing of both will continue.
When autopilot systems are designed and tested thoroughly, they work reliably -- and can be 100% trusted. When such systems are used in ways they were designed to operate, they work flawlessly.
When we mix people with such systems, failures can occur.
"[Failures occur when] pilots find themselves having to suddenly figure out why a complex system is no longer acting as expected and then take quick action based on that analysis. Moments of high stress are precisely those in which the human brain is least equipped to figure out complex situations."
Systems intended for use by people must be designed with the limits of the human mind. Our minds can get confused. Systems must be designed to work with people, and make it easy to temporarily turn them off.
4
@IN- !00%? Sure. We can "trust" 100%, however things never work 100%. Every piece of technology we own and operate needs repair or replacement 100% of the time. That is why we have all those repair shops and stores. And although seemingly contradictory, those human minds built this stuff.
2
Can't help myself:
"Inadvertent" is a generous word to use to describe the AF 447 co-pilot's actions. The plane went nose up because the dude pulled back on the stick for minutes on end for truly no reason, all while the computer was barking "stall". It would be the equivalent of reacting to your a sudden power outage at home by sprinting as fast as you can. There may not be a ton of great options, but that one is certainly not great.
Interestingly, in that case, while the pilots didn't feel they could trust many of their readings, they *did* trust the altitude, meaning their (correctly) believed they were falling. Hard to understand what they thought was happening. In any case, I would say that pure panic was what brought that one down, which is obviously a human problem, not an automation problem like in the case of the 737 Max.
3
When flight problems emerge, and the pilot needs to take control of the aircraft, a single emergency button should allow the pilot to override automation. Sufficient old tech instruments should be part of the cockpit display so the pilot can properly guide the plane if automatic systems display errors.
10
"Unfortunately, in the case of the 737 Max, it seems that Boeing has built a plane with a fundamental aeronautical issue that it thought would be resolved by adding a new automated system. But the company convinced airlines and the F.A.A. that the planes were essentially interchangeable with earlier models of 737, and therefore pilots who were already trained in flying older 737s would not need comprehensive additional training on the new system. The F.A.A. agreed with that conclusion. This would, among other things, save money on crew training and maintenance costs. That has proved to be a terrible miscalculation."
Knowingly building a "ready to stall" plane by taking an existing air-frame and dropping a heavier engine onto the wings is a decision by incompetent managers who have been promoted for reasons other than capability.
Boeing is showing symptoms that American companies with highly incompetent management have shown (Kodak, Xerox, GE, Enron, this list is long and growing).
Also, Boeing now outsources its code writing and most of the careful work necessary to get quality. But, by outsourcing there are now no overall system tests performed before the airplane is sold. Which, hey, its OK because it is cheaper to outsource and not test than it is to do all the work carefully, in house, then test it.
Management bonuses are way bigger with everything outsourced and no system test.
The stock is way up. That's all that matters. In America.
17
@Michael
They added larger engines, which required a slight re-positioning of the engines on the wings.
While this needed to be better disseminated within the using community and there should’ve been better training to the pilots than the 3 hours on a tablet and some reading material, so they know how to react it has nothing to do with “America.“ The plane is fundamentally sound (look at the hours/incident ratio), but has a serious issue that needs to be better addressed.
1
So Boeing built a plane that could nose up into a stall after take off automatically. Then to counter that flaw they added MCAS to automatically reverse the nose up to counter the stall but that system has automatic sensors that are flawed and could pitch the plane into a dangerous nose down situation under maximum thrust right after takeoff. Now Boeing is adding new software to somehow automatically counter the flawed workings of the MCAS which was added to counter the flawed workings of the underlying aeronautical stability of the plane at the exact time when that stability is most critical. And we are now expected and are being told to believe that this software won’t be flawed as well?? Meanwhile almost 500 people have died and Boeing and the FAA still say this is a safe aircraft and there is no need to completely decertify the entire fleet because the main flaw - the take off flight characteristics that lead to a potential stall under certain conditions can never be fixed without a complete engine/air frame redesign can be effectively overcome by a cascading series of autopilot computer controls, sensors and software fixes and when all else fails we’ll now train the pilots to pull it out of the fatal dives the Rube Goldberg systems throw the plane into. Is this an appropriate approximation of what constitutes passenger safety being the number one priority at
Boeing?
22
The false idols of otherwise godless modernity are all starting to show their feet of clay: not only is artificial intelligence and automation to blame for these debacles, but aviation itself is a prime cause of manmad climate change. Belching obscene tons of soot-laden semi-combusted aviation fuel that is indistinguishable from kerosene, each flight sullies our atmosphere in a way that will take eons for nature to somehow correct. Our brattish insistence on going somewhere just because we can, when will we examine this and attribute our planet's destruction to our insupportable indulgences?
10
If Boeing can point to the specific flaw in the software and replicate it for public view, and then explain the fix, this would go a long way to restoring public confidence.
If they just say it is patched, then the public shall remain wary, and have little confidence.
4
Perhaps some clarification is needed here.
The core issue is what is called a “common type rating”. To save training costs, especially at small airlines that may have many different aircraft types and a small pilot staff, Boeing (and Airbus) try to make new more efficient models fly and have systems and controls similar enough to current models that regulatory agencies (like the FAA) will certify the aircraft with a common type rating. Additional “differences” training would be at the discretion of the operators in this case, a significant cost savings. This is, for the airlines, a deal buster.
Boeing’s view of the Max is that even with the MCAS system installed and pilots being uninformed about it, what it does, how it can malfunction and what to do when it does, it doesn’t matter as long as pilots use the same pitch control malfunction procedure. Here it is. 1) Disconnect the autopilot (If it’s on), 2) counter the pitch excursion with the control wheel, 3) disconnect the two pitch trim disconnect switches, 4) extend the manual trim handles on the trim wheels and trim as required. The autopilots and autoland with be inoperative.
So unless there is some mystery involved that’s the whole program. The procedure is a “memory” checklist along with a very few others like Engine Fire and so forth. Like most airline pilots, I want to know how everything works, but increasingly, that’s just not realistic and in many cases, including this one, unnecessary.
12
@Confusedapotamus -- Could you also address how it might be that two apparently experienced pilots did not know this, and did not do this?
Also, do we know how many times pilots have had to do this, because they system went bonkers on them?
4
"Unfortunately, in the case of the 737 Max, it seems that Boeing has built a plane with a fundamental aeronautical issue that it thought would be resolved by adding a new automated system. " My mechanical englineer husband had read an article about how Boeing tried to copy the engines developed by Airbus, and put them on their old 737 model plane, but they didn't fit exactly, so they moved their position. This made the aircraft less stable, so they added the software to rectify the mechanical problem. I don't think it's a good idea to use software for this purpose because software is too susceptible to glitches.
11
VN, I didn't state that correctly. You are right. The larger point is, they tried to use technology designed for a different aircraft and make up the difference with software. I myself am not an engineer, but that doesn't seem right.
4
@V N Rajan
What you say is beside the point. It is Boeing who is responsible for designing its aircraft, where ever the individual parts come from.
Its seems (an many experts have said) that they tried to squeeze one more facelift's facelift's facelift out of a 52-year old plane model, and went too far. And then tried to use software to fix the design flaw, fingers crossed.
And AAT let them, because they (and the US politicians) are in Boeing's pocket.
8
@Nancy:
You are trying to find the root-cause, though you don't understand how systems (entire airframes) and software are designed. When we do not understand an area, it does not help to pose blame.
Whether the 737 MAX series airframes are "less stable" is problematic. Aircraft have parts/sections (eg, engine placement) that reduce stability in some ways but also improve stability in other ways. This should not be a problem if Boeing has fully tested each area affected by their new engine/airframe design.
You also say that "software is too susceptible to glitches." This is a generalization, and is not correct. Software is no more or less susceptible to glitches than any other area of engineering design. Any problem you've seen with software was in commercial non-safety software that is rarely tested thoroughly -- for example in consumer products and business software. Such software have "glitches" because those manufacturers made business decisions to not fully test.
In space, medical, aeronautics, automotive, etc. applications, software is heavily used and tested thoroughly. We rarely see software "glitches" in these industries. However, even these industries have software failures when it is not tested adequately.
Boeing did not properly design, test and educate pilots on the MCAS system (which was designed/added to counteract effects from the larger engines of the MAX series).
We may find that 189+157 people died due to Boeing using poor quality in design/testing.
2
I am neither a pilot or an aeronautical engineer so I would defer first to pilots as they have all these lives in their hands and then to the theoretical world of aeronautical engineers as we know that engineers have to listen to the money bosses.
It appears to me, Boeing made a huge mistake and the FAA bought it to help profits, listening to lobbyists, ".. the company convinced airlines and the F.A.A. that the planes were essentially interchangeable with earlier models of 737, and therefore pilots who were already trained in flying older 737s would not need comprehensive additional training on the new system. The F.A.A. agreed with that conclusion. This would, among other things, save money on crew training and maintenance costs. That has proved to be a terrible miscalculation."
I agree with Andrew J. Chisenhall Cologne, Germany. a pilot, we constantly need pilot training to do without autopilot and how to interact with it. Pilots need training in the worst of situations on the "Actual automation that the plane will use, nothing less." Nothing is foolproof but practice gets us much closer to errorproof. No errors usually means fewer accidents hence all pilots use check lists.
7
@WorldPeace2017
It was pretty two-faced to call the 737 Max a great innovative and brand new plane in marketing and just the same old 737 when communicating with the airlines' technical staff.
5
It’s hard to know where to start here. While it is probably true that automation along with improvements in aircraft engine and systems reliability has improved flight safety, “automation” involves a great deal more than what you describe and some of what you say is just plain wrong or oversimplified to the point of nonsense.
First, no airplane of any type can make an automated takeoff. Some aircraft can make automated landings some of the time at some specially equipped runways. Usually, that’s only one or maybe two runways at really big international airports. Not all the time everywhere, that’s wrong. As for pilot proficiency, they make all the takeoffs and the vast majority of landings and highly realistic flight simulators and frequent recurrent training keeps them sharp.
Aircraft automation encompasses many more things than just flight control. Systems automation has replaced Flight Engineers who used to monitor the aircraft systems, manage normal and irregular procedural checklists, some ground communications and many more things now handled much better by systems, communication and checklist automation.
Flight control automation has improved the precision of onboard flight control, planning and information, but this has been offset by increased complexity in traffic control procedures.
Modern airliners, especially Airbus aircraft are extremely complex aircraft. Confusing, blurry and inaccurate descriptions of them aren’t helpful.
12
The author states the following: “Pilots on at least two flights in the United States reported similar problems, but in those cases they were able to disengage the autopilot system and recover control of the plane.”. If the problems were, in fact, similar to Lion Air and Ethiopian, then this statement is misleading. The M.C.A.S. system operates to push the nose over when it detects an impending aerodynamic stall even when the autopilot is turned off and the aircraft is being flown manually. If the problems were, in fact, similar, the pilots of the two flights in the U.S. saved themselves and their passengers by not only turning off the autopilot, but most, importantly, turning off the power to both the main electric trim and the autopilot trim inputs to the horizontal stabilizer, and proceeded to stabilize the situation flying the aircraft manually and using the manual trim wheel, things with which any rated airline pilot is quite familiar.
16
@John Clifford -- Thanks for the clarification.
However, if as you write, "things with which any rated airline pilot is quite familiar," then how could it be that in two recent situations experienced pilots, each nearing 10,000 hours, did not do this?
Perhaps these are only two of a great many incidents? Then a small % failure rate would be expected. Is it that bad, that constanta problem? Or is this just harder than you've made it sound?
4
It appears that according to the article and quotes from Boeing people, the new engines in the airplane "could cause the nose to pitch dangerously skyward, Boeing has added a Maneuvering Characteristics Augmentation System, or M.C.A. S., that would kick in and push the nose down if necessary". Did Boeing create a design that is aerodynamically unstable, like modern fighter aircraft, that requires a computer to enable it to fly normally and continuously compensate for an inherent instability? Should not a commercial aircraft, especially one designed to carry people, be designed to return to straight and level flight in "hands-off mode"? If so, how and why did the aircraft become F.A.A. certified?
16
@Lawrence Abbott -- Except in very old aircraft we'd now consider slow, low, and under powered, engine power settings, weather, fuel consumption, and other things change the behavior of an aircraft more than that. Those times are gone.
Once stabilized in flight, with weather conditions stable, many aircraft can be adjusted so that for a time it remains straight and level at that altitude and power setting. However, even then things can change, as with the weather.
2
"...it seems that Boeing has built a plane with a fundamental aeronautical issue that it thought would be resolved by adding a new automated system. But the company convinced airlines and the F.A.A. that the planes were essentially interchangeable with earlier models of 737, and therefore pilots who were already trained in flying older 737s would not need comprehensive additional training on the new system. The F.A.A. agreed with that conclusion. This would, among other things, save money on crew training and maintenance costs. That has proved to be a terrible miscalculation."
"Terrible miscalculation" seems an overly charitable gloss.
13
The problem in this case is not automation, but badly designed (even, as said below, arrogant) automation. The underlying fault of MCAS's anti-stall software is that, if the pilot overrides its action (pulls back to resist its push), it repeats its action cumulatively, on the underlying assumption that it (and it's sensors) must be right and the pilot must be wrong. I spent 40 years of my life writing software, for functions much less critical than MCAS's. I learned to examine my design, and its underlying assumptions, by exhaustive thought experimentation - asking of "what if". The MCAS software seems not to have been subjected to such examination. That would have been failure of good practice, perhaps reflecting a modern overly "technological" and insufficiently "literate" approach to software design, based on the perception of software as mechanism rather than expressed intention.
27
@Joseph Grant
Excellent observation. I have a related background.
Boeing wanted to build a highly efficient plane that could save customers the need for extensive crew re-training. Gaining the efficiencies let to an aircraft that was only conditionally stable in normal flight (like many high performance military planes).
Commercial aircraft have used auto-pilots for decades. In their design, the pilot is assumed to be the one who flys the aircraft in difficult times, so an auto-pilot failure just puts control back with the pilot--and is easy for a pilot to recognize.
The new "maneuvering characteristics augmentation system", or M.C.A.S is something else. It's job is to fill in for pilots by addressing the conditional stability of the aircraft and making it fly more like the earlier 737's. This obviously works pretty well--when the sensors are working properly. The problem is that the MCAS is still treated like an auto-pilot when it fails, with apparently no indication of MCAS failure except a terrifying urge to push the nose down and speed up.
The MCAS needs to be fail-safe--or at least self-diagnosing, and it is not. In addition, the pilots must know what it does, how to detect inappropriate action (prior to loss of the aircraft) and how to correct.
Finally, why is the angle of attack sensor failing after take off? While much of the discussion concerns the flawed automation, there has not been as much looking at the failure on take-off prone sensor.
14
@Joseph Grant -- "I learned to examine my design, and its underlying assumptions, by exhaustive thought experimentation - asking of "what if"."
Of course you are right. But that is in conflict with the sheer volume of code to be examined in the newer designs. With everything automated, there is code written for everything, and it all needs that attention. Just one thing overlooked becomes like this.
2
@Mark Johnson -- " there has not been as much looking at the failure on take-off prone sensor."
On planes, cars, every piece of machinery with a sensor, the sensors are a constant source of trouble. It is not a surprise that a sensor or its connection fails. It is routine that they will.
Of course, that brings us to another part of what you've said, designing for the certainty that sensors fail.
4
Here's the problem: "the company convinced airlines and the F.A.A. that the planes were essentially interchangeable with earlier models"
The FAA had no clue about, well anything.
The FAA should have the ability to independently review and test each addition such as this.
Otherwise, what good are they?
12
@luket FAA has been widely seen as a captive agency for decades.
5
@luket -- The FAA are very good people of vast experience.
What we see is "regulatory capture" in action. They are not free to act as their judgment tells them. It is politics.
3
Boeing deserves to be suspected of malfeasance unless it comes clean about what seems to be the cause of both 737 Max 8 crashes.
Supposedly a sensor provided erroneous readings to the autopilot, deceiving it into believing the plane's nose needed to be lowered.
Why were there erroneous readings? Was it a software problem, or a hardware one?
If it was (is) a software problem, why hasn't Boeing been far more proactive and urgent in rolling out and installing a fix? Boeing has been curiously dilatory.
If it was (is) a hardware problem, Boeing needs to explain how a software change will fix faulty hardware. Any refusal on the basis of trade secrecy should be met with the truism that if Boeing won't explain, the 737 Max should stay grounded.
Unlike our predictably credulous president, the rest of us will not take Boeing's CEO's unsupported word for anything.
9
Jeff Wise makes a good point that should be taken seriously by self-driving car enthusiasts. Substitute the words "drivers" for "pilots" and "driving" for "flying", and then omit the phrase "stick-and-rudder" from this opinion piece:
“…DRIVERS find themselves having to suddenly figure out why a complex system is no longer acting as expected and then take quick action based on that analysis. Moments of high stress are precisely those in which the human brain is least equipped to figure out complex situations.
The best way to deal with emergencies is to train and practice beforehand, so that response becomes automatic. Ironically, by turning over the mundane [...] aspects of DRIVING to computer automation, DRIVERS are deprived of the opportunity to continually practice. This leaves them without the mental automation that might save them in a crisis.”
13
@Amy -- I find this happens to me with something as simple as cruise control on my car, despite decades of driving without it.
We need to keep the human element factored in. Reliance on tech has its dangers in this aspect too.
Automated driving certainly will produce accidents of automation failure.
However, right now we have a shocking number of traffic deaths. It is a leading cause of death in America. Would automation be better or worse than that.
Then we can consider how to accommodate the human factor to reduce that too.
Now with airplanes, pilots are not just some guy behind the wheel doing something routine. They drill and do simulations, and check rides, and checklists, and have backup with a co-pilot. It is serious business taken seriously.
When the human factor is failing twice in six months in that more rarefied setting, it is a problem of a different order.
2
Hacking your control away from you.
Is this a myth?
3
Few years ago Boeing CEO was heard saying that he was not interested in another “Moon shot” (deaigning a new plane from scratch) and instead modify and update existing plane.
Herewe are with inherently unstable aircraft. Would you just modify Ford Taurus instead of designing a new vehicle for more passangers?
4
Boeing is finding the downside of trading expensive solid design based on known principles for (apparently) cheap Rube-Goldberg in-flight real-time corrections.
Marketing, not engineering!
13
@John✔️❎✔️Brews And you know this how?
You should change the word "Autopilot" to artificial intelligence. But in the case of either one, money covers the problem.
2
I believe transport pilots should hand-fly at altitudes under 10,000 ft. AGL.
4
@Jambalaya. Many do, but sometimes autopilot usage is recommended for complex arrival and departure procedures.
2
The arrogance of technologists is lethal.
14
How many software devs go to boot camp, essentially getting their training from a Cracker Jack box?
2
I don't want to fly on an airplane that's wearing a "patch".
12
Too bad GPWS, CFIT, TAWS, GCAS, EGPWS, TCAS, T2CAS, and similar alphabet soup din't override the 737 MAX's MCAS from gaining speed, due to a possible stall, but into a CFIT - or more accurately the SCFIT. At least a stall gives you one more chance than flying faster into the ground.
Maneuvering Characteristics Augmentation System (MCAS)
Ground Proximity Warning System (GPWS)
Controlled flight into terrain (CFIT)
Terrain Awareness and Warning System (TAWS)
Ground Collision Avoidance Systems (GCAS)
Enhanced Ground Proximity Warning System (EGPWS)
Terrain Collision Avoidance System (TCAS)
Terrain and Traffic Collision Avoidance System (T²CAS)
Controlled Flight Into Ferrain (CFIT)
Software Controlled Flight Into Terrain (SCFIT)
Was Boeing too cheap to redesign the wings, fuselage and engine position to have a plane with inherently stable flight characteristics like previous 737 designs that do not require a software hack to compensate? - in the MAX's case to prevent it from causing the nose to go up inappropriately thereby increasing the wings' angle of attack at a critical stage - taking off, slow but increasing speed and a nose up (naturally) configuration, near the ground.
Indeed, the 737 is not a stealth military aircraft like the completely un-aerodynamic F-117 (or B-2) that requires essentially FOUR REDUNDANT computers to enable it to fly - else it is a flying brick. 737 MAX's MCAS used only one sensor? Cheap!
Will future design flaws be covered up with software hacks?
12
Hacked the controls right out from under you.
Is this a myth?
2
Too much reliance on technology thereby eroding piloting skills.
There is no substitute for hands on training in the cockpit.
(Simulators are fine, up to a point.)
"Know your airplane, fly your airplane." The words of my sainted instructor, a WASP veteran.
Apparently some carriers have directed pilots to manually fly their aircraft during "severe clear" weather conditions. After the Lion Air accident and the recovery of information indicating they had been On autopilot, I am surprised that carriers did not direct 737 Max pilots to fly their aircraft manually to cruise altitude. I would wager that many savvy veteran pilots did just that.
The comments speculating that the FAA may have colluded with Boeing re certification are troubling indeed. I hope that is not true but we are living in a time when "it's all about the money".
Condolences to all those affected by these two tragedies.
Let's press for a thorough understanding and solution to the problem.
Expediency be damned!
5
I am asking again to anyone who may know.
Are the Boeing flight control systems susceptible to being hacked in the same manner that computers can be hacked?
NYT - please pin this to the top of the list so all readers will read it and maybe someone with knowledge can answer.
2
@LivingWithInterest
No. The central nervous system of today’s aircraft does not connect to outside flight controls to computers. It can be done but right now the pilots control the plane with imputs by hand or control of the flight systems through an autopilot.
1
My automobile’s throttle is controlled by computer software that, when it goes awry, can create a stall or surge from a standing start, which is no fun at a busy intersection. Lifting the hood to determine the problem is no longer an option for anyone but a trained mechanic with a plug-in. It is a software rather than maintenance problem.
And ponder for a moment the black holes into which we all descend when our computers and tablets refuse to behave, often the result of a power surge and router failure. You diddle around in search of a solution and hours have passed. I can’t imagine being on a flight deck with 150 souls in my hands, dealing with essentially similar problems. Patch be damned. Return to the tried and true systems that made the 737 one of the greatest airliners ever built.
11
"Many modern airliners are capable of taking off, flying and landing without any human assistance."
Without ANY human assistance? Which ones?
If the author is a private pilot with no airline experience, he's not qualified to make some of the judgements he has made here. It's irresponsible. Mr. Wise, this is a disservice to the public who think you know what you're talking about.
4
I want to draw a cruel parallel between the VW emissions scandal and the 737 Max 8 tragedies. At some point, someone at a desk-a semi-important middle manager who speaks code gobbledygook, identifies a horrible error that he/she is positive can be patched. Against the backdrop of heavenly sales, great economic benefits looms a fatal decision not to confront the truth, but to paper it over with something resembling a solution. I imagine that callous, cowardly, small minded idiots who maintain an elitism because of their proprietary techno-intelligence are too shallow and sufficiently emotionally adrift to admit and report error. These kind of people are the architects behind the fraud in scandals like VW's and the executioners behind the deaths of hundreds in the skies above.
4
@MARK The scenario you paint is, I think, a little paranoid. MCAS was intended to overcome a minor difference between the flight characteristics of the Max and those of earlier versions of the 737, so that the Max would comply with the existing 737 certification. There was nothing wrong in doing that, no malevolent intention to cover anything up. The trouble is that the anti-stall software was very badly designed. Instead of making a corrective adjustment just once, to serve as a warning (as, I believe, systems on earlier 737s did) it made repetitive and cumulative adjustments if the pilot did not heed its warnings. It thus took control away from the pilot, if its warning was due to false information (e.g. defective sensors). The problem with MACS arose not from greed or malevolence, but from simple incompetence - failure to ask "what if".
4
@Joseph Grant My point is speculative but compelling because at some point there had to be someone at Boeing with a horrific "oh oh" moment. And from there dishonesty, deception and delay rules until tragedy sufficiently buts the breaks on a production line.
1
@MARK "puts the brakes"...
I hate flying. Used to love it.
8
@Hb
This is "AI" at its ? worst? best? usual?
I’d bet lack of training is the culprit, and that’s on both Boeing and the airlines. All about profit over lives.
1
First things first: The CEO of Boeing should be fired and subjected to civil and criminal suits, and the head of the FAA should also be fired.
Next, comb through the FAA for any employee whose loyalty appears to be to Boeing, rather than to the flying public, and fire them.
Finally, impose a lifetime prohibition on any FAA employees ever working for Boeing or its partners.
11
@spindizzy Before you go through all that silliness assuming that Boeing was at fault, you might consider that the FAA was shut down for 5 weeks due to Trump wanting to build his wall. Those 5 weeks can be vital in times like this to properly design and test fixes which must be studied, re-tested, approved, and certified by the FAA. If the FAA finds any issues, then the whole process must be repeated. Your assumption of malfeasance is completely unwarranted without a lot more investigation. And that investigation WILL happen as part of this process. It always does.
2
@spindizzy
There is no head of the FAA. POTUS can’t seem to find anyone that will take it. Unfilled position. Could be part of the problem.
@spindizzy Man, are you dreaming. This is America, the bastion of the revolving door.
Despite recent hysteria, automation in the cockpit has made airline flying safer than it has ever been. There has not been a fatal airliner crash in the U.S. since 2009. Of course there are going to be mistakes made in any complex system, and humans will make mistakes when faced with anything new and unfamiliar. It's a tragedy that we lost the passengers and crew of two 737 Max 8s, though what may well turn out to be a SNAFU in training and in the details of the human/machine interface. I am a private pilot of relatively little experience. But even I know what to do if my 737 Max starts pointing downhill against my will: flip the Stab Trim Cutoff Switches, next to the co-pilot's left knee, and then trim the airplane manually. Boeing has probably already fixed this, will roll out software fixes as fast as possible, and will come out of this in good shape: neither Airbus nor Boeing can deliver a similar plane to an airline without waiting years, and by that time this will be a faded memory and a valuable lesson. By the way, automation not only saves lives, it makes it far easier and faster to fix problems once they occur. Tesla can fix your car, via software update, wherever it happen tso be.
3
@Dexter Ford It isn't the technology that is a problem with this situation. The problem is Boeing's dishonesty about the need to train pilots on the new airplane and the FAA's willingness to go along with it.
5
There is still absolutely no data from the flight data recorder, and in absence of that, the MCAS should not be blamed for the crash.
The head of Ethiopian Air Lines is on record as saying the crew was already trained on how to deactivate the MCAS. It would be reasonable to assume that +any+ pilot flying a 737 Max 8 would have already obtained training in some form after LA610's crash.
Is it possible that the training offered was simply insufficient? Lion Air only offered its pilots a 90-minute PC-based training class, while other airlines offered full-day simulator-based training.
Finally, is it possible that either the airline or the pilot and FO falsified their training credentials for the Max8?
3
no doubt the software code controlling this plane is very complicated probably having over a million lines. somewhere in there someone made a coding mistake. there is probably a line or two that mistakenly triggers a nose down attitude when certain airspeed and angle of attack parameters are met. this condition may have been reported by pilots a year before the fist crash. if thats true and boeing covered it up they are in deep kimchi
6
Most professional pilots are loathe to opine about any aircraft mishap until the NTSB releases their findings, and even less inclined to point a finger of blame at the pilots involved, because, 'but for the grace of god, there go I.'
All that said, here are my observations.
As I understand it, the problem with both planes can be defined as a stabilizer trim runaway. The stabilizer is the horizontal portion of the tail of the plane. It adjusts in relation to the speed of the plane. The auto stall prevention system - the MCAS - would operate through the stab trim. All jet aircraft that I am familiar with (I'm certified in only six) have switches generally called something like 'stab trim cut out' switches (the terminology varies). Stab trim runaways are the kind of emergency is routinely practiced in the simulator, at least at my company.
The same can be said of the reference in the article to excessive speed. If the cruise control in your car suddenly accelerated your car to 100mph, wouldn't you just disconnect it?
Again, I do not intend to blame the pilots of either plane. None of us know how we will react in a high stress situation such as those pilots faced. But it is possible that had they had the presence of mind to simply use those stab trim cut out and auto-throttle disconnect switches, things may have turned out differently.
It's not that the plane is too complicated for mere mortals to fly, contrary to what this article and our tweeter in chief say.
5
@Peter B. -- I've observed conflicted feelings among pilots on this point.
They do resent the automatic "blame the pilot" attitude.
At the same time, they very strongly want to think, "I could have flown it out of that problem."
They have a strong aversion to an assumption of themselves as pilots being helpless to prevent their airplane crashing. That is not only mere ego, it is a healthy "can do" attitude -- none of us wants the pilot to give up.
So they do both, "don't just blame the pilot" and "that particular pilot was a fool and I could have done it."
3
@Peter B.
Sir? Are [the Boeing] flight control systems susceptible to being hacked in the same manner that computers can be hacked?
Is it possible that the plane's controls were hacked? Ships have had collisions because of GPS "spoofing", so is it possible that this crash was due to a similar attack? This fight in particular, EA-302 had at least 22 UN workers on board. The UN has been targeting in the past. Also with18 Canadian nationals on board, Canada is now in the sights of China for the arrested Meng Wanzhou, Huawei's chief financial officer and daughter of company founder Ren Zhengfei. Besides the ongoing militarization of eastern Africa, by the US and China. Consider the disruption to Boeing, a major defense contractor, and to the international aviation industry. The possibility of hacking is growing in banking, military, engineering, education, media, politics, credit, so is it possible here?
3
@JS No, not in the sense that your computer can be hacked. The various computer systems are located in modules that are either replaced when updated, or are updated by having a specific computer device used by maintenance to update the software. Other than the on board WiFi, none of the aircraft computers are part of the internet of things
1
@Nick -- "none of the aircraft computers are part of the internet of things"
Not exactly. There is a constant flow of data outward, by data link to the ground.
Where there are plug-ins, there can be a corrupted plug-in. Ask Iran and North Korea about their carefully isolated missile programs getting hacked anyway.
So we have a data link and a potential source of corruption. Add to that every plane's elaborate radio set-up with is also computer assisted.
We mustn't get cocky about that possibility.
I don't know of any reason to suspect it this time, but I see plenty of reason to exercise a lot more caution than we do.
That would well be the next accident, or the next. It is as much a danger as a bomb, and we know that happens.
3
Well, at least just as many people as died in the two recent 737-Max accidents don't get killed on American roads every three days. Oh wait . . . actually 300 people do die on American raids every three days.
2
The next time some puffed up government hating Conservative or Libertarian spouts off to me about the dangers of government regulations, I will remind them of "Boeing Max 8".
We desperately need Effective Government.
10
There is a video on You Tube of a flight attendant landing a huge Airbus via a simulator. As someone who has flown aircraft today’s Airbus fleet is fully automated with pilots engaging them after takeoff. We expect that pilots have the experience to fly an aircraft on and off autopilot. Why didn’t Boeing leverage the experience of commercial pilots in designing this aircraft?
2
There is obviously a problem with the software of this airplane. For Boeing to put people’s lives at risk by not owning up to the fault that they had after the first crash is criminal by any means.
The company deserves to lose business and market value as a result. People have lost their lives and the manufacturer is to blame. Shame on them!!
2
@MetnPride -- Now that Boeing has captured the market in an essential monopoly of US commercial aircraft manufacture, your idea punishes not just the company but the country and entire industry.
So we see yet another danger of monopoly. It hasn't been that long since Douglas Aircaft and Lockheed produced competitive commercial aircraft. It needn't be this way, but it is.
3
Pilots are just "overseers of systems"? True, up to a point. I'll put my money on Captain Sullenberger over an autopilot any day of the week. We still need humans in the cockpit when computers and software go south.
6
@Troutwhisperer -- He was flying an airplane that would do that. We need more than a good pilot.
I hope I’m wrong, but what’s published so far indicates that if the pilot, using the switch on the yoke to move the trim, the MCAS is off for a few seconds, then resets, which means it fights back. Then, if the pilot cuts the breakers, he has no electric motor trim at all, and has to spin those hand wheels, mechanically geared down direct action, many times if the stabilizer is way wrong. Probably would have enough time if there was enough altitude, but near the ground with the stabilizer run to one limit, maybe not. Meanwhile, the plane is being pretty violent, and I’m guessing a pilot would go full throttle out of fear of a stall. Full thrust with that engine placement....
Really hope the system doesn’t work this way.
4
Earlier this week, there was a reader commenter on NYT articles about these planes, ‘Mike, NY’ who made several disparaging comments about people’s hysteria and sub standard trained and qualified ‘third world airline pilots’. I hope he is now reflecting on his assumptions and inaccurate perspective particularly in light of the number of US pilots now coming forward with their concerns. He belittled many others’ comments over those 3 days.
4
I am sorry. I do not want my airplane to be 200 person flying Tesla! All hands on the yoke and feet on the rudder by fully trained pilots with thousands of hours in the air. We deserve no less from airlines making major profits while gouging passengers for checking a simple two bags.
4
Would Sully have been able to land the 737 Max in the Hudson?
5
I am sure US-based airlines will charge passengers premium for traveling on an Airbus aircraft going forward.
2
"This would, among other things, save money on crew training and maintenance costs."
There's the root of the problem: $$$
4
And still the story changes!
Now we are being told that the altitude and vertical ascent profiles that were shown to us do in fact NOT show the final minutes before the crash, only the first 3 min AFTER the take-off. The data for the last half of the flight have not been released!
Hence, all speculation and information we thought we could extract from the published data are moot.
Obviously, a lot more was going on during the final few minutes before the crash.
One can only wonder why - AGAIN - the public is fed manipulated information!?
What is the purpose of misleading us?
Are the Russians involved in this cover-up as well?
1
An autopilot “running haywire” is not what happened on the Lion Air flight if the MCAS system was part of the problem. MCAS doesn't operate when the autopilot is engaged.
The linked article “Pilots…reported similar problems” may not refer to similar issues at all, since the reported problems related to autopilot engagement, not MCAS, which, again, doesn’t operate with autopilot engaged.
The “faulty sensor” to which Mr. Wise alludes is an angle of attack sensor of the same type found on most modern airliners. These sensors, basically wind vanes, can provide incorrect data, most often if damaged or improperly maintained. Boeing’s mistake was to program MCAS to respond to a single AOA input. The designers calculated that professional pilots around the world would immediately disable the trim motors and thus the MCAS if there was a problem of the sort that apparently led to the Lion Air crash.
Pilots at U.S. carriers, at least, are trained ad nauseam on the procedure to stop trim anomalies. It requires one finger to flip two adjacent switches within inches of either pilot's elbow.
One can learn far more from a piece in the Seattle Times (“Pilots Struggled against Boeing’s 737 MAX Control System...”) about the issues pertinent here than from all the Times coverage taken together. The public and apparently lots of government officials have been misled in many ways about these accidents and what can be known at this time about their underlying causes.
9
@RBW I have read several references to there not being a simulator for the 737 Max 8 so how are pilots supposed to train for the unexpected given there does seem to be significant enough differences between the Max and the regular 737?
4
Mr. Wise is premature in his conclusion about the 737 MAX accidents because the MCAS system hasn't been detailed by Boeing or anyone else who possesses that knowledge. MCAS is an unknown to both the press, and the pilots flying the MAX.
2
Can they ever solve the aerodynamic fault of the front edge engines? Why don’t take just make the ground clearance higher and fit the engines under the wing? Higher landing gears!
@Harri August. Which then requires changes in the wing and fuselage to accommodate the larger gear, which then has other ripple effects on the design, etc. And the result is a different airplane, greater cost of the product, new type certification for pilots with greater cost for airlines, etc. etc. Not a trivial design change. They have looked at that before.
1
It seems inconceivable Boeing would sell planes that crash due to a design fault. I sense something else at play, like cyber-sabotage, perhaps by a foreign power eager to demonstrate its capabilities, for whatever reason.
1
There is also a military version of the 737 used by our military. Should it be grounded too?
2
@Mike As far as I'm aware, the military derivatives of the 737 such as C-40 and P-8 are all based on the NG models, not the Max series
Clearly, if the plane is a rock..
No amount of electronics will make it into a stable or reliable aircraft. Especially if the electronics fail, and suddenly the pilot has to fly a plane that just doesn't want to fly.
The problem with the Max is, Boeing chose to put larger engines on an air frame that wouldn't prudently accept larger engines, without finding a different wing mounting location and upsetting stable flight characteristics. Hence the need for the new electronics.
This was a business decision.
They didn't bother to tell anybody.
7
@james smith. All real-world engineering has an element of business decision. You can’t avoid it.
@james smith
You are absolutely correct
These are a very informative set of comments.
So, an angle of attack sensor is external to the aircraft and measures, essentially, how close to a stall the aircraft might be. Clearly, avoiding a stall is a very high priority.
A gyroscope can measure aircraft attitude. Other measurement devices can measure absolute and relative aircraft speed, even the speed of the airflow over the wings.
An autopilot that depends on an angle of attack sensor must also be able to detect when that sensor is not working properly. While it may be the most direct measurement, if it indicates a near stall, and the aircraft is flying level but over-speed at low altitude (as suggested in the article), a stall would seem to be exceedingly unlikely. Yet the automation kept putting downward pressure on the yolk.
Clearly dependence on a failure-prone sensor without redundancy or "reasonableness test" (except from the crew sensing a problem) is a flawed concept.
The decision to hide a difference in aircraft performance from pilots so the training can be shortened using a failure prone sensor that can only be detected by the very pilots the automation is intended to keep ignorant seems fraught.
6
The FAA shouldn't allow manufacturers to depend on computers to fix fundamental aeronautical design flaws that exist for marketing reasons. That is a failure of policy.
6
@Mark
Bingo.
Boeing's decision to use engines too large for the air frame was clearly a business decision.
All indications are, they were trying to hide something.
6
My real concern when reading this is not about planes but cars. As we rush to embrace self driving cars think about the difference with planes. We all know computers sometimes crash, and all technology suffers occasional glitches.
But commercial jets are thankfully inspected constantly. Further, pilots are trained professionals who are working when in the cockpit.
Driverless cars, however, will likely be subject to diagnostic checks about as often as they are now (not nearly as often as planes). Moreover, most drivers will not operate as pilots in these cars, but more likely become distracted and over time even act more like passengers. And even the conscientious ones will not be as trained or capable as pilots when technology (inevitably) fails.
As we discuss rolling out self driving cars can we at least discuss these issues? Let’s demand a much higher level of diagnostic inspection than current automobiles, and be very aggressive in creating “driving” laws that insure the person in the “drivers seat” is properly trained and always “working” when behind the wheel.
As a Luddite, I of course will prefer to keep driving myself. We humans are giving up everything that tests our thinking and reflexes.
9
Actually, good design based on the laws of nature, etc. is still needed in this day, and age. To go away from all of that with larger engines, winglets, etc. leave the 737 Max crashes dangerously close to never being able to fly again more than likely, just a software fix, and new pilot training for the thousands of pilots who would need this.
1
The bottom line is that more automation and less human piloting makes the planes much safer.
Most people, unfortunately, don't have enough quantitative reasoning skills to put this in perspective, and will only focus on the lives lost when the automation fails, instead of the far greater number of lives saved by avoiding human-pilot accidents.
3
@Celeste
Automation and electronics does not make a plane that is aerodynamically sound.
Not more so than does a computer hooked up to you kids wagon, means you can operate it as a commuter on your local toll road.
Not mentioned in this article is that one of the Ethiopian Airlines pilots only had 200 hours of flight time. In the US, the minimum for these large planes is 1500 if I recall.
Also, in the US, pilots generally start flying on very small planes which lack automation and then move up to larger planes and higher levels of certification. There is a lot of space in this country that enables our US pilots to get this practice without automation.
Other countries lack the space and in the case of Korea, there is a genuine worry of flying into N. Korea.
This became a factor for the 2013 Asian 214 flight in San Francisco crash. A runway was under construction and the automatic guiding beacon had been turned off and the FAA told pilots to use visual approach instead. Whereas US pilots had the practice with visual approach, the Korean pilots did not since they really had not been flying the smaller planes.
The Air France 447 2009 flight mentioned in this article had 3 pilots. The most Senior pilot was allowed to fly even though he had only *1 hour* of sleep the night before and he was sleeping in a bed intended for pilots to sleep.
The Airbus fly-by-wire planes use independent joysticks. The very inexperienced pilot was pulling back on his joystick causing the plane to stall and the pilot with mid-experience as co-pilot did not realize that the pilot was pulling back on the joystick -- Boeing planes have a connected yoke.
6
Didn't the pilot actually in charge have more than 8000 hours? Why judge based on the new kid?
3
"it seems that Boeing has built a plane with a fundamental aeronautical issue that it thought would be resolved by adding a new automated system."
Software or pilot training cannot compensate for a dangerous design flaw. Boeing 737 Max 8 should be grounded until the flaw in the airframe, engine positions, or whatever, is fixed.
3
@JBThe designer of the winglet, Richard T. Whitcomb, was not part of any team, and many of the ideas he had came to naught. The winglet was designed for fuel economy, and so that more planes at busy airports could be parked together, not the fact that the winglet, weighed more, prevented drag, and no direct lift, and it took the industry 3 decades to accept this design. It was originally intended for the shuttle, and for gliders. It may be a total engineering flaw gone wrong, and the luck ran out in the last 2 crashes.
2
Two memories come to mind.. a sharp contrast in cultures.
1. Johnson and Johnson's response to the Tylenol poisoning in 1982.
2. The space shuttle Challenger blowing up in 1986.
J&J's response was outstanding and totally appropriate. J&J ignored short-term income and placed their responsibility and reputation to upgrade their customer's safety and welfare ahead of all other business and medical considerations.
Their Tylenol market share crashed from 35% to 0% overnight. But, they worked to discover the cause (deranged person or people poisoning bottles of Tylenol) and took immediate and sensible steps to prevent it again. Within 12-18 months their market share rebounded The consuming public trusted J&J's sensible and prudent behavior and sound judgment.
The Challenger disaster was allowed to occur because some narrow business interests and management egos were allowed to bend the rules and to allow the countdown to proceed... in spite of serious reservations by the operations "on the line" people who recognized the great risks. Established safety procedures were largely ignored.
The USA space program has never fully recovered and regained the public's confidence.
3
People should understand better what technical debt is when designing or updating complex systems.
1
I am wary of cumbersome solutions to problems rather than elegant and simple ones. In the case of the Max 8, according to this piece, one automated system was added to counteract the failings of another automated system. Now a third patch is added to the "solution" to counteract the previous two. (That may bring up the chances of more glitches in the program and the need for more patches and levels of automation, like bandages to cover new leaks)). In any case, if the pilots are expected to manage such complex systems when failures occur we may not need pilots anymore, but people trained in other skills. Boeing is building jets and products for companies all over the world not just highly trained computer savvy pilots in the first world. It has to simplify as much as possible the protocols, skill and training level necessary to run these new planes in Ethiopia, Brazil, France, the US or wherever. On top, Boing has deliberately concealed the level of training required in the marketing of its new cumbersome product. Accidents are waiting to happen. The costs, human and material, are enormous. Who pays for them?
4
So let me see if I got this right. Boeing introduced a replacement of the 737 with a plane that had fundamentally different, and more unstable, handling characteristics, therefore requiring pilot retraining to fly without autopilot. They then sold it as a plane that behaved the same as the old 737, but only if flown with new software, the MCAS, that mimicked the old 737, that the pilots were not trained on and that could be overridden in an emergency only with a new procedure buried in a manual. What could possibly go wrong?
17
So can't we be training pilots with what they needed when the new software was released? No. Boeing fears it will be taken as admission of negligence to do so now when previously said it wasn't needed.
3
"We don’t fully understand the risks introduced by the rise of increasingly automated airplanes" – au contraire, I don't trust the Boeings of the World to be COMPLETELY honest with us and to look beyond the bottom line.
4
In my light plane, I have airspeed from a pitot and airspeed derived from GPS. You’d think that in an airliner, a critical warning would come up when the two disagree by much. Icing happens quickly. If my GPS had 12 satellites and disagreed with the pitot, I would trust the GPS.
4
@ThePB
The GPS will only display Ground Speed. This means in a 30 knot headwind, your GPS speed would be 30 knots less than your Indicated Airspeed provided by your Pitot Tube-Static air system.
So if you want to fly a particular speed, trust your Pitot Tube/Static system . This measures the relative wind , traveling over your wing, and providing lift. Not your groundspeed.
ps, what kind of light plane do you have? Did you train for a Pilots License?
A lot of comments about modifying the 737 platform over and over. Saying the 737-100 and 737 MAX 8/9 are the same platform is like saying the 1964 Mustang and the 2019 Mustang are the same platform.
They’re not.
6
@Juliet Lima Victor
They are the same platform in regards to wing height above the ground, and wing location.
In this respect, mounting a larger engine required a redesign placing the engines higher and farther forward on the wing, compromising the planes center of gravity.
That is why the MCAS system was installed. To cause the plane to mimic the handling characteristics of the prior editions of the 737 and to prevent potential stall at slow speeds.
1
The fact that Boeing, and the FAA, pushed forward with a plane that has a fundamental design flaw tells me that it was not just about poor engineering but about a company and its oversight agency having cultures that are deeply flawed.
They may fix this plane this time because it's killing people, but we are all still at risk until some fundamental changes happen n these organizations.
10
If it turns out a human made a mistake nobody will care and we'll continue to trust human pilots. If it turns out the computer made a mistake people want to make it into an existential crisis. Get used to it -- the problems of the future are all robot problems because we are giving them more and more jobs. And they dutifully do them with better outcomes.
4
It's obvious that many commenters have never operated complicated machinery with very sophisticated computer controls.
Modern airplanes are an integration of thousands of complicated systems that must be precisely controlled for efficiency, comfort and yes safety. The inclusion of modern microprocessors has likely avoided many more accidents than it has caused. Even with extensive, rigorous testing it is possible for the variance in performance of individual "identical" components to induce a dangerous control condition.
My experience is in the operation of modern power plants which are similarly complex. Often some minor changes in tuning/programming in the processors, what many would call a patch, completely resolves issues. Regrettably when it happens on these planes, it is usually at a time when there is literally not much room to maneuver, with tragic results.
Adequate training is essential, as is adequate hands on flying time. But...these things are very complicated and a human just can't know everything. Computers are essential to safe flying, and more and more will be deeply integrated into the flying of these machines, just as they will with autonomous cars, trucks, trains etc. A few Tesla cars have already been involved in fatal collisions, and they will be involved in more. But, on the balance they will save far more lives than they cost.
The 737 Max issues will be resolved and they will provide countless safe trips to millions of people.
3
@J and then there are the cowboys that don’t trust the indications or the instruments and say we never run that way so how could we now.... until the inevitable trip and safeties lifting waking up the neighbors then having to deal with the aftermath.... certainly not a plane crash but sometimes extremely expensive in more ways than they can imagine
@J
Ok, let me know 'J' when you find a computer, electronics design that allows your kids wagon to operate as a decent commuter on your local toll road.
Planes need to be able to fly. Period.
And if they can not, all the electronics in the world wont correct that problem. Especially if it fails. That's the whole reason you have a pilot on the plane to begin with.
1
@james smith. Ummm... it is able to fly and has been doing so for many thousands of flights.
"The problem is that it’s hard to prove a negative. Passengers won’t have any faith in the fix until decades have passed."
Very interesting piece. The new engines apparently have a center-of-mass that is farther forward, but how much compared to the previous generation aircraft? Presumably, the control system was designed to keep the aircraft out of an unstable (e.g. oscillatory) regime, but perhaps it doesn't always. The truth will out if the FAA gets to dig in to Boeing's test data.
2
Apparently the two larger more fuel-efficient engines—that Boeing used to market and sell the plane—
Were NOT the engines originally designed for this aircraft.
And they needed to be moved significantly forward on the wing, raising stability problems and causing the aircraft to nose upward at angles that would lead to a stall (hence the MCAS system).
So there is a basic design flaw. Not just software problems.
[And somehow that was approved by our FAA.]
8
It's very common for aircraft to be repowered with newer, more efficient more powerful engines. No design flaw. The changes are accounted for and changes made to computer controls/programming, possibly hydraulic and fuel systems. That's what engineers do.
Not everything is a conspiracy to make money at the cost of people's safety. Crashes have major bad for financial markets mpact for airlines, plane and engine mags, insurance companies. Losing an aircraft is bad all around.
Once in a while they don't get it exactly right. Designing and manufacturing these things is complicated.
9
@J
According to the article: "Unfortunately, in the case of the 737 Max, it seems that Boeing has built a plane with a fundamental aeronautical issue that it thought would be resolved by adding a new automated system."
This statement certainly suggests that there IS a design flaw, one that needed a new computer system to compensate for. This is a big problem. Adding a new computer system purely to compensate for a design flaw is asking for trouble.
3
[And somehow that was approved by our FAA, in cahoots with Boeing—that’s the worst part of it.]
2
Apart from the terrible tragedy of this crash, I would like say that it’s annoying that the 737 Max 8 is continually referred to as the most popular plane ever sold. It’s only popular among airline industry executives who most likely never fly anywhere in this model of plane. They like it due to it’s ability to squeeze ever more seats into a limited space. Flying in this plane is most kindly described by passengers as ”miserable”. The seats are rock hard, there isn’t enough overhead bin capacity to hold 1 piece of approved size carry on per passenger, all entertainment has been removed, the seats don’t recline and unless you’re traveling in a group of three, I don’t know a single person that wants to sit in the middle seat. The 737 max is flawed mechanically and otherwise.
7
@Annie
Don't we have choices, Annie, unpleasant as the choices may be.
In the pursuit of "cheap fares," compromises are made in the safety and comfort of the passengers. Since the passengers prefer "cheap fares," the airlines provide them... at the compromise of comfort and safety.
We can always select to travel in "first class," but the price may be more than feels "comfortable and safe" (for our wallet).
As in grocery shipping, one may shop at fancy and fancy, specialized grocery stores such as Whole Foods and The Fresh Market, or one may shop at less expensive mass market stores such as Pick 'n Save, Wal-Mart, Target, Cub Foods, and Kroger.
@Annie I think what is really meant that the 737 series as a whole which includes the original -100 and -200 models, the 'Old Generation' (-300/-400/500) and the 'Next Generation' (-600/700/-800/-900) is the best selling commercial aircraft series ever, not 737 Max. Equipment like seats, lavatories, and galleys are chosen by the customer airline as 'buyer furnished equipment', so you get what they choose. The 737 Max series has the newer Boeing Sky Interior similar to the 787 model with bigger and easier to access overhead bins and better lighting.
@Annie
That is part of the image damage control campaigns the public will be subjected to now by Boeing and its PR agencies and trolling agents..
It's all about the Benjamin's, baby -- they talk the talk about the primacy of safety, but in the end the profit motive prevails.
Boeing stretched the 737 for greater capacity and used larger engines with greater range. More occupied seats and longer range equals more revenue per mile. The larger engines required repositioning the wing. The stretched 737 was inherently unstable requiring an automated system to compensate for its instability. Boeing cut corners on the training necessary to alert pilots to these key changes and new systems. Pilots have characterized the pilots instruction manual for the 737 max as totally inadequate.
The FAA and the airlines went along with the rush to get the new 737 in service. Pilots reported problems with the new system and when the crash in Indonesia occurred, that was the time to ensure that every pilot knew how to disable the system when an incident arose or to ground the 737 max until the software fix was ready. Neither measure was taken, and a second 737 max went down. Had Boeing, the FAA, and the airlines put passenger safety above profits, these tragedies never would have happened.
13
Insufficient testing and inadequate training equal greater profits for Shareholders and CEO's, which is supported by Lobbyists and, if not a toothless, puppet government oversite bodies (e.g. FAA). This behavior is rampant throughout most industries, unfortunately with deadly outcomes for the airline industry users. The US's lackluster performance, quality and oversite will continue to erode US products on all development and manufacturing fronts in the decade(s) to come. P
1
With Boeing's design and the FAA's certification process now very suspect, it is imperative that we find a way to assure the worldwide flying public and Boeing's customers that the design of the aircraft has been fully reviewed, to include whatever modifications may be necessary, in an open process that builds confidence and trust. I call upon Boeing and the FAA to work with the National Academy of Engineering to appoint an independent panel of distinguished experts in aerodynamics, control systems, software and verification to conduct a design review of relevant aspects of the Max series and commit to implementing whatever recommendations come from the panel. To do less would fail to honor the memory of lives lost in these two crashes. With the 737 series the bread and butter of Boeing's commercial airplane business, doing less would threaten the future of the company. Lastly, we should remember that Boeing's commercial aircraft sales are the single largest contributor to exports from the United States and thus it is vitally important that Boeing regain the confidence of all. Finally, perhaps there will emerge from this process good ideas on how to overhaul the aircraft certification process in the future.
4
If there is a man-made fault, this gotta be it. If pilots can't deal with faulty sensors, Boeing must pay for forcing on these intractable technologies.
1
Lots of technical comments here. I can't add to those knowledgeably, not being an aeronautical engineer, but I do drive a fairly sophisticated piece of engineering - my car. I have ABS, hill descent, AWD, all-wheel steering etc. but if I didn't know what I was doing, all of the fancy equipment is worthless - especially in a Canadian winter.
I would think expertise is a thousand times more important when you are "driving" hundreds of people in the air.
While there appeared to be a criminal lack of training for the pilots of this new aircraft and its software, I can't help but think that "pilot error" will be the ultimate outcome of the black box investigation.
3
All that stuff on your car is designed to enhance safety, not to compensate for an unsafe modification introduced into the design of your car to cut corners.
Boeing should have redesigned the plane or at least told the airlines the modification required extensive new training. Instead they obscured the nature of the situation to increase sales. Now, more than 300 deaths later, that decision appears wrong morally and financially.
There is a tendency to want to excuse those in charge and blame the little people. It’s better to start at the top though because it save time.
6
@Steve
Sorry - I have had all too real, and almost deadly, occurrences in cars due to cutting corners in design. You must not be familiar with the small part, and a replacement part that had the same number, that caused several hundred deaths in General Motors products. They knew about the part defect that was causing the cars to turn off completely, including power brakes, no matter how fast the car was going. They knew enough to replace the part, but didn't do a recall and didn't change the part number so you couldn't know if it was a new, improved part or the old, deadly part that was in your car.
All for profit, all to cut corners. Hundreds of deaths.
I don't believe anyone, big or small as you put it, has gone to jail for those deaths and there will be no repercussions to the Boeing negligence either, I am afraid.
2
I'm a computer programmer and a private pilot. I fly the Cirrus SR-20 which is a highly automated plane. The technology does make the plane safer, however, before I could fly this plane alone I needed 20 hours of training and have to do recurrent training every six months. As a programmer, you can test and try to break your programs during trial, but it's impossible to test for every single situation the users will actually put your program through, and it will, pardon the lingo, crash. In my opinion what is needed here is an extremely simple way to quickly disengage or override the autopilot, and yes a lot of training for pilots before they fly this plane. Maintenance costs and training costs should never be a consideration when the alternative cost is loss of life. As a society we have to eliminate from our vocabulary "...but it will cost too much".
8
@ART Agreed, but it appears all the airlines and gov wanted much more fuel efficient airplanes and Boeing didn't want to redesign the 737 from the ground up - so they basically grafted huge new engines onto the old airframe.
And it looks like those big engines affect the aerodynamics of the plane to the point, where it pitches the nose up because the engines move the plane's center of gravity back too far. They 'solved' the problem by introducing a software system that overrides this tendency but when it isn't working right, you basically have a plane that isn't airworthy.
I could be jumping ahead but it looks like there is a fundamental issue that is being corrected by software and if it malfunctions watch out below.
5
@ART
I don't think disengaging the MCAS was the problem. Apparently that can be done fairly quickly, in ~3 short steps, by the pilot or by the copilot.
See my comment below yours: I actually wonder instead whether that is what they in fact did, only to discover that the airplane had unforgiving flight characteristics that no longer were so docile, resulting in the pitch-up during the final moments of the flight.
And, in contrast to your SR-20, these jets don't have an in-built parachute.
2
Two thoughts from a stick-and-rudder SEL flyer:
The Air France accident was a clear pilot error. The last thing you do when losing the airspeed indicator while in stable level flight is pull the nose up. If I remember correctly, they combined that with a second fatal mistake: reducing power. As any aerobatic pilot knows, both actions together will induce a stall/spin.
When flying aerobatics, we do this for fun. With a passenger jet, it is deadly.
If the pilots just had not touched the yoke, trim and power levers, they could have flown all the way to Paris with the iced up pitot tube, using their GPS as their speed over ground indicator.
Second thought: I can't believe that the pilots in Ethiopia had not practiced switching the MCAS off following the Lion Air crash. I would have practiced that so often in my nightmares, it would have indeed been second nature, simulator or not.
Actually, I wonder whether the pilots did indeed switch off the MCAS, only to discover that the flight characteristics of the plane without the MCAS were radically different from the docile handling of a 737. Essentially, the pilots may have been handed a plane they had never flown, with a narrow and unforgiving envelope, at low speed and low altitude, resulting in the stall/spin that killed it. It did ascend rapidly during its final moments, which would have resulted in a stall, unless matched by sufficient airspeed.
21
@Kara Ben Nemsi
Great comments, after the first crash last year there really is no excuse for pilots to be surprised or not know how to react , they must be something fundamentally wrong with this plane, period.
2
@Kara Ben Nems
It would be VERY interesting to see powerset, airspeed and altitude plots overlaid onto the VSI.
Your MCAS "off" - power "on," pitch "up," airspeed "down," stall, spin, near vertical impact scenario is all to easy to imagine. Unless the pilots were prepared for the pitch up when they applied lots of power it would be all too easy to become overwhelmed.
Knowing to turn the MCAS off doesn't solve the problem unless you can respond to what happens next.
4
This is a bigger issue than you might realize for the society as a whole. The automation in question is a bunch of software algorithms that are not qualitatively different from what runs social media or your company's sales and marketing. We've worried too much about automation taking away jobs and not nearly enough about it outright killing us. This ought to be a teachable moment for all algorithmic driven systems while humans are still in control.
3
@Denis Well yes, automation may kill you. However, as has been proven with modern aircraft, it will kill you a lot less often than a human being.
@Melbourne Town I’m actually talking about all the other places automation has entered our lives. It might kill less but it kills different too.
Perhaps it is notable that Boeing propagated their engineering failure even in the wording of the corrective system they devised to counteract it -calling it the Maneuvering Characteristics "Augmentation" System. The problematic "maneuvering characteristic" is apparently that the new more powerful engines incline the fuselage to pitch nose up under heavy thrust. If Boeing would have more truthfully named the thing "Maneuvering Characteristics Mitigation System", one has to wonder if anyone would have bought this aerodynamic lemon in the first place? Is it a sign of the times that a software patch would be perceived as a legitimate fix for a fundamental structural flaw?
11
If these engines were known to be prone to pushing the nose of the plane dangerously up then they are not a good match for this plane’s aerodynamics. That should have been the end point. Instead they were built and the end point should now be all these funerals but I’m guessing the end point of this plane is still a long way away. And no one will go to jail.
12
@Dave
Whether the engines are pushing the nose up is secondary. A pilot who is used to manually fly that plane will compensate for that subconsciously. That would be as second nature as pushing right rudder during take-off and climb-out when flying a single engine propeller plane. It is required to compensate for the yaw induced by the propeller.
The thing is, if the pilot is only used to flying that plane with the MCAS engaged, which automatically compensates for that nose-up pitch at take-off power, compensating for it will in fact NOT be second nature, if the MCAS needs to be switched off due to malfunction. In addition, turn characteristics will also be altered. Failure to having practiced flying that plane without the MCAS in a simulator may turn out to be the reason for this particular accident.
2
@Kara Ben Nemsi Precisely, and let me reiterate a point that has been lost in some of the news story simplifications: the different engine mounting configuration, as I understand it, does not make the plane legitimately unstable. It merely makes it less stable in some circumstances. Next, understand that any airplane must be unstable to some degree, or it would be exceedingly difficult to maneuver. Hence, every aircraft design is a compromise between stability and instability, and every design requires the pilot (or computer) to compensate at times.
What the different engine placement does is make the new Max fly differently than the previous models. Boeing compensated for this with software so that pilots would not require costly retraining, but this ploy is only valid assuming the stability controls are working.
I don't fault Boeing for releasing an inherently unstable and dangerous airplane, because I don't think that's the case. I do, however, fault them for not having explained the differing handling qualities and the MCAS adequately to pilots.
5
@Kara Ben Nemsi
Then the pilots should never have been certified to fly this plane, just because it is called a 737 it is actually something quite different. Boeing pushed for no extra training required and the faa agreed, unless I’ve read something incorrectly in a previous article.
1
The fact that automation has, over decades, greatly reduced accident rates (to 1/10th of previous rates) is very good, and cannot be forgotten. But no system can be considered perfect, and for every planned software upgrade, I assume Boeing builds a software flight simulator that puts it through its paces. Such simulators must ferret out the failure modes of all control software, and these must be carefully reviewed. We cannot afford to learn in real-time, with live passengers.
Keep in mind that 100+ people/day are killed by gun violence in the US, totaling to millions dead -- a problem unchecked -- indeed aided by Congress, with even lawsuit protections for gun manufacturers. But fortunately, the Conn. SC sees it differently. Remember, gov't for the people.
Similarly, the FAA must work to serve the public, not the airlines, by reviewing all control software with their simulators before it is approved. Lives are at stake by the millions who fly daily. Don't waive Boeing's intellectual property rights to keep that hidden.
2
Or, as a commercial pilot relative of mine once said, "too many airline pilots don't know how to actually fly." The pilots of the Miracle on the Hudson and the Gimli Glider spent a lot of their careers pre-automation and were not trapped by technology. They knew how to fly.
12
@View from the street
It's easy enough to do and fairly cheap for every airline: Have a fleet of small aerobatic planes and compel the pilots to spend a couple of hours each month putting these little planes through their paces.
It is fun and the pilots would get it for free! I would sign up in a heartbeat.
@View from the street
!st pilot had 8000 hrs flying time.
1
@kimball. 8000 hours doing what? Huge difference between 8000 hours doing the same limited set of things over and over “flying” a highly auromated airliner, and 8000 hours doing many and varied things including hand-flying small aircraft to hone stick-and-rudder habits. Hours are a very poor metric for “highly skilled”.
Boeing wanted to save the money for a completely new development (Boeing 797 maybe), and decided to go ahead with an 50 (+) year old basic design. That is the outcome! When Airbus suggested a new A350 based on wing and fueslage geometry of the A330, but with 90% of all components newly developed, the airline industry wouldn't let them get away with it, but pressured them for a totally new aircraft instead. They did let Boeing get away with it! Why is that?
1
Airlines didn't let Airbus get away with it because of money, their re-designed A330 would simply not be competitive with the efficiency of the newly-designed 787. In the case of the 737, the update it's just as efficient as Airbus' own updated A320NEO.
Planes undergo thousands of hours of flight testing, even to certify relatively minor updates such as the 737MAX and A320NEO. Unfortunately sometimes it's still not enough to uncover every possible design flaw, and in the case the MCAS system it's also not hugely different from stick pusher systems installed in countless other aircraft, and predicting the possible consequences of it was extremely difficult.
1
A one in a million chance when you do it a million times, my point is they do happen.
2
Nice, informative article, but the author loses a little credibility by suggesting airliners are capable of take offs on autopilot. No, they aren’t. All takeoffs are manual until somewhere in the climb profile, after the aircraft is “cleaned up”, the pilot engages the autopilot to provide the smoothest ride for the passengers. Some pilots elect to hand fly to altitude, but In the interests of passenger comfort, most use the autopilot.
3
@BBH. I believe some are *capable* of automated takeoffs, as the author said, which is different from your statement that they are not actually flown that way. Both statements can be true.
Seems to be a matter of compounded negligence by multiple actors. Interaction of an inherently unsafe design, a callous attitude towards safety, and human error/ limitations.
1. Boeing, in convincing the airlines that no additional training is required, and not admitting to itself that there is a serious design issue regardless of MCAS.
2. The airlines- any airline- should know better than to be convinced that no additional training is required which points to the possibility they were penny pinching. After all, all airlines knew about the Air France flight and so many other major incidents.
3. The pilots themselves- many pilots unions protest for wages but when it came to these issues, no one seems to have spoken up. If one flew a few thousand hours, one must be knowing the risks involved.
4. FAA and various other governmental agencies (DGCA in India for instance) for waving it off.
1
@Rao, K. Kind of a reach to impugn pesonal integrity (callous?? sheesh...) without facts and data. You have no idea of how the design decisions were made.
Don’t attribute situations to malice that can be adequately explained by simple error.
software updates to flight control system with no accompanying announcement or info to pilots?
That is criminal corporate behavior.
Heads must roll or the lesson the vendors learn is that they can get away with murder.
5
Boeing needs to install a couple ejection seats in a 737 Max, find a “Tex Johnston” and take the plane up to 20000 feet and figure out how a crew can manually recover from an MCAS induced dive. If they can’t recreate the problem and solve it they need to redesign the plane so that it is inherently stable. 737’s aren’t military aircraft, they must be able to be manually recoverable from unusual attitudes by pilots without software assistance.
3
@Mark Me Present It says in the article that two pilots at least were able to recover from a MCAS induced dive by turning it off. It's really the autopilot system that's overcorrective - my understanding is that if there's no autopilot at all, the plane can be flown safely.
However, I support automation. The only thing I think that Boeing didn't do right here was: (1) set up an easy way to turn off the MCAS (the current process involves coordinating 4 actions which seems quite complex to me if you're not trained in it) and (2) point people to the relevant pages in the manual on how to do it.
What I don't think Boeing needs to do is outline how the MCAS system works, etc. in the manual or provide training on these changes, etc. Just on how to turn it off if it starts to malfunction, that's all that's needed. There were US pilots who said that's all they cared about (apparently there's a lot planes do in the background these days and they don't need to know everything they said).
@RamS
Not so fast. In a case like this, the pilots need to know much more than "just disable MCAS." My understanding is the other instances in which it misbehaved were steady state conditions.
In this case (and Lion Air) they also need to know what the airplane will do when it's disengaged. It seems that MCAS is Boeing's response to a basic air frame instability problem in a low speed, high thrust condition posed by repositioned, more powerful engines.
If you tell the guy holding the yoke to "just turn it off" and don't mention that "by the way, when you do if you apply a lot of power it'll try to stand on its tail" you'll still need a lot of body bags.
1
@RamS In other aircraft, pulling back on the yoke turns off the autopilot.
Pilots who don’t know how to fly the plane should not be flying it; so, too, Boeing should not design a plane that most pilots don’t know how to fly.
Boeing sought to adapt the old 737 and remake it as 737 Max. The transformation did not work and Boeing then used unproven AI to correct its mistake. Doing so introduced complications for its pilots which Boeing obviously did not want to own up too. Result: incomplete training for pilots and crew and possibly untrainable pilots who cannot understand or absorb the necessary computer knowledge.
5
"Pilots on at least two flights in the United States reported similar problems, but in those cases they were able to disengage the system and recover control of the plane."
This misleading comment is indicative of questionable reporting in this piece. Those two incidents appear in NASA's ASRS database and refer to autopilot issues. MCAS is disabled when the autopilot is in use.
If the author had taken the time to conduct even a casual search on the ASRS database for any transport category airplane, he would find dozens of examples of autopilot pitchovers, suspended climbs, failure to capture course or altitude, failure to maintain vertical speed...and on and on. Doesn't mean the airplane is bad, it just means that autopilots do that stuff sometimes.
His argument that the MAX series is flawed because it requires MCAS is weak sauce. Airbus aircraft have a stunning array of envelope protection that they require to fly and they aren't bad airplanes for having it. For decades, airplanes have had stick shakers and stick pushers as stall mitigation devices. They work.
It's a little too soon to be burying the MAX based on poorly documented speculation.
10
@Paul Bertorelli, Yes Paul, I totally agree, but NOT too soon to severely criticize Boeing, and modern air transport in general, for the [philosophy of] not documenting and training in detail, the systems and procedures, especially in 'fly-by-wire', and electronic systems. The 737 family has existed since the mid-sixties for good reason, and they are fine aircraft, but perhaps dollars and expediency have overshadowed proper training. Kudos on your AV videos.
There is so much misinformation out there, including in the comments. Would the Times consider doing a story on the actual design, testing, and trade-offs involved in these aircraft, with detailed sourcing from within Boeing? At the appropriate time, of course.
9
How about going back to drawing board and design a new plane with better aerodynamical features? Seems that this extensively modified 737 is no longer behaving as 737 and is inherently unstable.
2
" the accident rate worldwide has fallen from about four accidents per million flights in 1977 to less than 0.4 today."
The MAX has only 100s made but 2 fetal crashes already. How BA and FAA reconciled the data and still claim MAX to be safe to fly in the U.S., even other countries concluded otherwise.
Should we investigate the problem?
1
@David 100s of flights made? Try thousands per week (currently 8600+) over several months!
So it's safe to assume that AI won't be replacing pilots anytime soon.
The Boeing 737 Max 8 looks like a plane for history. I would not trust them ever regardless of manufacturer assurances.
4
These pilots remind me of the woman who was supposed to be driving the Tesla in Phoenix, but was streaming a movie. They should NOT be using auto pilot on take off or landing. It should only be engaged when cruising altitude is reached. They are hired to pilot the plane, not ride along
2
@Barb Porto the pilots likely were paying attention. From other reporting I've seen, other antistall interventions can be overcome by pulling the stick, sort of like pushing the brake turns off cruise in the car. The MAX needs to have this intervention turned off before the pilot can save the craft.
So, in the end it was the push from Boeing to have no additional training required that may have killed hundreds of people. Corporate greed knows no bounds.
1
I’m gonna guess the guy in the right seat with 200tt was the problem not automation.
3
"The problem is that it’s hard to prove a negative."
Actually it is not - the Author and the Editor probably meant that it is hard to prove a counter-factual.
To save money, the FAA allows Boeing employees to certify the flightworthiness of Boeing aircraft.
What could possibly go wrong?
10
@SKK Yes, that arrangement was made in 2005; previously the FAA did its own work.
2
The 737 MAX design failings were meant to be mitigated by the MCAS however the software only had a single data source. Critical software development requires multiple sources and calculation of data, for example, a common result that can be calculated at least two different, unique ways before action is taken. From a flight perspective, this inadequate software system was the critical deficiency from which all other failings stem.
Software will be used to advance flight's exploitation of physics more and more going forward. Removing proper software system development from aircraft development programs and it's place on the critical path will not and should not happen. However, it is my experience that the relative importance of 'right' software development is deprecated when the system designs originate within the mechanical, electrical, and aeronautical engineering specialties. Evidence of a 'working system' in aviation critical development should remain with software engineers having knowledge of the other engineering sciences. All should have a basic understanding of the theory of flight.
Also, I don't understand why auto pilot is permitted to be engaged on the approach and departure legs. Pilots can fly and should be in control, or at the very least, be able to take control easily instead of being forced into a race with death on what should be a routine departure leg as was the case in these two accidents.
11
Pilots are able to disengage the autopilot with a simple press of a button on the control yoke. Automation isn’t the problem here.
@John
Many instrument approach and/or departure procedures require compliance with very tight navigational tolerances, sometime +- .13 miles. This can be very demanding to hand-fly, and so autopilot usage is required.
@BBH When the autopilot is engaged the aircraft is stable. When the autopilot is disengaged the MCAS is active. This was intended to keep the trim correct and the aircraft stable when flying manually. The MCAS can be disconnected and the trim controlled manually, but this requires moving the two MCAS disconnect switches to the disconnect position. This is not a procedure that was included in the training for the aircraft.
1
This is not a problem that can be solved with a software patch. or more to the point, it should not be solved with a software patch. Boeing needs to transform itself, and to do so, it needs input from all the pilots--international and domestic--who fly the 737 Max8. There is so much wrong here that nothing short of transformation will do. The relationship with the FAA represents regulatory capture. The attempt to use software to overcome limitations of an old plane shows the limits of software to solve fundamental aerodynamic issues. The apparent attempt to replace pilots--experienced pilots--with very complex systems and software would seem to be a failed experiment, with hundreds of lives lost as result. An experienced pilot in Seattle was quoted on the radio as saying that the instruction manual was "criminally incomplete." That about says it all, doesn't it?
12
Has it occurred to anyone who's a aeronautical engineer that the longer the plane, the longer the lever arm and the great the potential to set up oscillations. The original 737's were around 102 ft. The newer versions are up around 138 ft, all for the sake of greater carrying capacity without having to do a complete redesign. any one have insights into validating this supposition?
5
President Trump thinks that it takes an Albert Einstein to pilot... when what we could use now is a Richard Feynman to investigate.
29
Suppose that the starting place for modern aviation was automation rather than human piloting, and a manufacturer came along and proposed introducing the human factor to the equation. The common wisdom would be bafflement that we would dare to allow a human take control when so many lives are at stake. There is just too much risk and uncertainty when a fallible and stressed-out human gets behind the controls, except in the most unusual circumstances (what engineers call edge cases). Compare the mind-boggling unlikelihood of dying on an airline to household falls (6,000 deaths per annum) or automobile deaths (2,000 children per year) and it seems this fear-mongering is grossly misplaced. Automation and autonomous machines represent a paradigm shift, and it takes a while to digest the benefits, especially when the drama of a death resulting from a self-driving car, or in the extreme, a tragic and horrifying plane crash, takes place.
2
When Boeing was designing this aircraft and the issue with the engines came up did anyone raise their hand and say “we need a redesign” or was cost of keeping the 737 the driving motive? It seems that the software addition was a “fix” and not the direction Boeing should have gone. The CEO reaching out to Trump also casts doubt on Boeing’s integrity.
14
Pilots need to be taught, first and foremost, to fly. They should learn to fly - manually - on light aircraft. They should fly manually large aircraft in the simulator. The opening to this article is absolutely right. Pilots are being trained as systems managers and not as flyers. Philosophically, every pilot should be able to turn off any automated system when it malfuntions and just fly the plane home in the old fashioned way. In a modern large passenger jet the pilot's true role is to know what to do in unexpected situations - the autopilot can manage the routine.
13
@Michael O'Farrell, Yes, but in addition, the pilots must be made aware by systems knowledge training, as well as plenty of sim training, of how these automatic systems function, and how to recognize and what type of malfunction [can]exist. This can be done, provided all parties concerned- Boeing, the Airline, the pilots union) are willing to undergo and be responsible for, such training. This has been going on for at least 50 years that I know of specifically. These airplanes, especially Airbus, are philosophically DESIGNED to be flown almost exclusively automatically. Remember the guy that said "Follow the money"?????
2
This isn't just a problem for airplanes. The self driving cars that Silicon Valley is working on are likely to run into similar issues. In fact, I can recall one fatal Tesla crash where the driver over-estimated the capabilities of the self driving system. That reminds me of the 1988 Mulhouse crash of the Airbus A320, where the pilots flew their airplane into trees when they over-estimated the plane's capability to keep them out of trouble.
Sometimes, engineers get the man-machine interface right. One reason that the Iphone was such a huge sucess was that it had a superb man-machine interface. Steve Jobs deserves a lot of credit for this. Other computer user interfaces aren't as good. I'm sure we have all had the experience of being confused about why our computer just did something. And those of us who are old enough certainly remember the 'blue screen of death', which would take on a whole new meaning if it occurred in a self driving car.
Malfunctioning automation can put operators in a bad palace. That seems to have happened in the recent 737 crashes. We need to remember that air crashes aren't always what they seem and our view of the Ethiopian crash may change a lot when the full facts are known. However, it looks like the so-called 'MCAS' system should not have been certified by the FAA. Now the world seems to have lost confidence in the US regulator. The FAA may have been subject to 'regulatory capture', which means they are no longer independent enough of Boeing.
23
Boeing, the FAA, and the flying public do not have to wait decades to find out if the fix works. MCAS should be testable in the labs and on the tarmac, repeatably, and reliably, until everyone is satisfied that it will do the "right thing" every time in every situation. Isn't that the whole point of automation?
2
There has long been the meme in aviation circles of The Dog and the Man in the future of the industry. The man is there to feed the dog. The dog is there to keep the man from touching the controls.
We're there. What I don't understand is the engines. If they were positioned so that they interfere with the basic stability of the plane, why weren't wings and control surfaces designed to compensate for that tendency? Airplanes are a complex system of inter-related systems that determine wing shape and position, weight distribution, and flight characteristics under widely varying conditions. The 737 Max was designed only to have a computer make those decisions. It's an awfully heavy reliance on fly-by-wire for a passenger jet.
11
@Stevenz
I heard that “joke” about the “dog and the man” from somebody at the Paris Air Show years ago and I remember it made me laugh at the time. I’ve been thinking about it now, and it sure doesn’t make me laugh anymore.
1
From the various published reports it would appear that the fundamental problem with the Boeing 737 Max 8 is not an automation problem, but a basic design problem. Apparently, the Max 8 modification involved bigger, more powerful engines, which had to be moved further forward on the wing. It is this that results in a tendency for the plane to take a steeper ascent on takeoff. Rather than redesigning the planes, and moving the engines, which would be very costly, a cheaper software/automation solution was undertaken. It’s all about the money!
14
You are right, it’s all about money. But the real money was made by not training Pilots on how to correctly address a malfunctioning MCAS system. The Pilots of Lion Air (and all US Pilots) were literally unaware that this system existed because it wasn’t in the operating manual. It wasn’t in ANY operating manuals.
The sales pitch from Boeing goes like this: “ we have this great new Airplane with many new efficiencies. It will save you fuel, can carry more passengers and requires no additional training”.
One small problem, it does require additional training and now the fix will be far more costly to Boeing.
2
@Cassandra. All real-world engineering design incorporates cost considerations. Can’t be avoided - it’s the nature of engineering, to find a “good enough” solution to the design problem. And yes, sometimes the engineer gets it wrong.
These planes are an extraordinarily apt metaphor for the larger issue of automation, as it applies to large-scale industrial implementation.
Why even use trained pilots in a vehicle with built-in mechanisms that can override human judgment? Seems like we’ve already convinced ourselves that we know better.
3
Everybody is blaming Boeing, and it may be deserved, and it may bear the responsibility.
Just suppose, however, the analysis of the flight recorder data indicates that the reason for the crash was for a reason not related to the plane, but the ground or flight crew - e.g. unlatched cargo door, cargo containers not properly restrained causing a shift in the center of gravity, pilot error, something else - will all the “experts” and media apologize to Boeing?
1
No, because they still should have supported a precautionary grounding and erred on the side of safety over greed. Two new model planes crashing under similar circumstances is extremely rare, those are all the “facts” needed to mandate a pause. Boeing deserves every second of the ire hurled its way, no matter what the outcome. It earned this by putting so much effort into fighting the common sense idea of erring on the side of public safety. The actions of the company and its too cozy regulators calls their integrity and level of safety consciousness into question, no matter what the crash investigations conclude. The heat they feel is well deserved.
15
@AndyW - Safety rides in economy class.
1
Some level of automation does indeed prevent unnecessary deaths, and airline travel has overall become even more safe.
Since the loss of hundreds of lives has sparked such genuine concern, I wonder if lives could be saved by other forms of finely tuned safety technology, such as biometric locks on firearms.
We lose more lives in a day in America due to guns than the 737 has lost in its entire history (since 1957). Roughly.
Yet there is outrage that the response of the US government was too slow on the 737, while demands for basic safety in gun ownership...
13
@Doug Keller - Right, except the 737 wasn't developed until the 60s, after the 727 (another plane that had "growing pains.")
1
@Stevenz
I read that the 737 was introduced in 1957. Perhaps I misread, or the article needed fact checking.
Still, far more Americans die every single day from firearms than in the entire history of the 737.
Some (not all) of those deaths would easily have been preventable, including by fingerprint technology that would allow only the owner to fire it.
We are right to mourn the loss of life, and to improve upon the technology that prevents it.
Two crashes on the other side of the planet grounds domestic flights of the 737 Max as soon as the decision is made.
Simple technology to make gun ownership safer for everyone? Too inconvenient.
2
Within a few years after entry into service in the ‘60s, there were 13 fatal 727 crashes. None were attributed to the plane. Most were caused by pilot error. Never grounded. 189 of about 1,800 planes were lost in crashes.
1
Whether we like it or not, the history in other fields has shown that the response to faulty automation is more automation: MCAS version 2.0, you get the idea.
For example, after the Air France crash, a fix would be to allow automation, upon loss of airspeed data from iced pitot tubes, to set the proper range of throttle to prevent shock induced stall. In other words, automation does what the pilot should have done.
However, automation has many blind spots which could cause it to misinterpret a situation, usually unknown until a mishap occurs.
We should really have the best of both worlds: automation at the service of competent and highly trained humans. A chess program + a master human player always beats a chess program alone. Of course, we should be willing to pay for and encourage the training of master pilots like Scully.
3
The solutions are blatantly obvious to any second-year engineering student. 1) Increase training; 2) Impose mandatory sensor and software redundancy; 3) strictly regulate and more severely test all advanced software systems that can directly place the public at potential risk for harm. Capitalism can’t work when a duopoly is so intertwined with understaffed government regulators that it can pretend that 300 deaths potentially caused by flaws related to the design of its new aircraft don’t warrant a precautionary grounding. Industry pundits are still on the air defending the idea the flying public should have remained as guinea pigs until “facts” emerged. Tell that to the thousands of friends and family members of those three-hundred. Most people understand all too well how industry arrogance and self-interest can combine to put them at extreme risk. We have all seen this scenario repeat itself throughout history, don’t lecture us about “facts”.
26
@AndyW
and 4) don't rely on software to paper over any inherent stability problems arising from changes in the structure of the aircraft.
9
Something that definitely deserves deep scrutiny. Especially true, since there are so many non-military aircraft on the drawing board that potentially rely far more on software to maintain basic flight stability than does the Max. In the least, these systems should require an order of magnitude more testing, along with multiple, independently designed backup systems.
2
„Unfortunately, in the case of the 737 Max, it seems that Boeing has built a plane with a fundamental aeronautical issue that it thought would be resolved by adding a new automated system.“ That is the key issue with the 737 max. Boeing added next generation engines with extremely large fans to an airframe from the 1960s. Delaying the development of a successor to the 737 and presumably saving billions (did not work). Boeing therefore intentionally accepted the risks of an aerodynamically unstable plane that required computerized stabilization to keep it from falling out of the sky. This is not about automation, it’s about greed, market shares and taking completely inancceptable risks in civil aviation.
68
@Carl
+1
2
And self driving cars ? Maybe on average they will be safer than manually driven cars, with some « in the news », high profile, stories of where they failed.
I would think it is part of the same story.
6
@David Martin
Exactly.
Driverless car supporters always say the human will take over if the computer gets confused. Just like this 737 situation.
9
Thing is, we don’t have stats for all the instances automation in aircrafts made adjustments that subtly prevented a tragedy. I’ve already been saved from an accident through a car’s proximity alert. No record will show that accident avoided.
We are mentally prone to be more horrified by a machine error that causes death than by the innumerable human errors that a machine would have never made. It feels worse to us psychologically and so we react more strongly. We’re not there yet but when self driving cars are on the road en masse I suspect we’ll see a great drop in accidents and fatalities. Nonetheless the horror when it makes a mistake will still being calls for a return to the good old days.
3
@David Martin
And we're talking MILLIONS of cars (and big trucks) on the road that might go haywire, not less than a 1,000 airplanes.
Yeah, the death toll will be much lower, but the feeling of betrayal when your self-driven ride veers into oncoming traffic will be at least as terrifying as the feeling those folks on the Lion Air and Ethiopian Air flights experienced as they plunged to the ground.
2
Why is this stall prevention system even needed? I looked back in a database of all commercial aircraft crashes that occurred during takeoff. I found accidents that occurred as a result of weather or engine failure, not stalls.
According to his old Army pilot, during both takeoff and landing every fiber of a human pilot's being is concentrated on what is happening. Corrections happen automatically. You don't think about it, your body just does it, both feet and both hands, instinctively and instantly. With this stall system, when it fails it is necessary to reach over to the center console, flip up a cover, and turn a switch off, while you or someone else (or both) is fighting the controls, and the plane is pitching up and down. It's said, "It only takes seconds" to switch it off. That's all it takes to loose control. The chaos in the cockpit while pilots are wrestling with the plane may mean it takes longer, very probably too late -- these planes need long distances to change direction and these takeoff dives occur too close to the ground to allow recovery.
This whole "safety" system sounds like something that was installed because it could be done, not because it should.
Use autopilot at altitude to travel long distances. Let pilots fly airplanes on takeoff and landing.
57
How do you account for the large reduction in aviation deaths? Do you have stats on how often the anti-stall function has been employed and prevented an accident?
@David I am a pilot. I agree. Flyl them off the ground, land them. Or if landing one hand on the AP disconnect and one on the Yoke. Don't trust automation. Sensors fail. Etc.
6
@David
My understanding is that this new system cannot engage if the autopilot is already on, which means the pilots were manually flying from takeoff to altitude.
2
There may be unknown risks posed by automated aircraft but the indisputable fact is that air travel is safer with automation than without. So by all means automation should not be looked at as a guarantee against a crash and must always be improved, but suggesting that we ought to reconsider the wisdom of automation in air travel is throwing the baby out (of the airplane) with the bathwater.
9
Imagine if we trained surgeons to watch robots to do the surgery and then expect the surgeon to jump in and save the patient once the robots makes a mistakes--say slice an artery or removed the wrong organ.
The autopilot and lack of training (and the flippant attitude of Boeing, FAA, and others) is appalling.
26
And yet the number of fatalities has dropped significantly with increased automation, and computer assisted surgery has delivered astonishing results.
1
@Jan Unlike the fatal mistakes you talk about in the body, there is time in the air to react and take over manual control.
I do think the AP system being turned off more easily should've been possible. Apparently on older MAXs, it was a simple button button here there's a 4 step process process to turn off the MCAS.
The only training needed was on how to quickly recognise this problem and turn off the AP. People who followed the Lion Air crash could've known how to do it and that's why the Ethiopian crash is puzzling to me: if it was the same MCAS problem, there was at least one fix (considered the easiest by SW pilots). Here, I'll Google it and tell you: "turn off the electric trim—which is used by the MCAS, and by trim switches on the yoke—and adjust the trim using the trim wheels."
Also, you see that Southwest and American Airlines do provide some extra training and require either the manual be read or watching of a video that talks about how to do this.
1
Essentially you have undependable systems and pilots unprepared to deal with malfunctions. So malfunction
manuals and constant training are absolutely necessary.
Good luck with that. How many of the world's airlines
will spend the constant training money and even know what kind of training is necessary.
Pilots don't fly planes. They just monitor systems like
automated cars. On an airplane, reaction time is
critical and different in every pilot. This makes flying
much more exciting, especially on take-off.
5
@Tom I am 71, became a private instrumented pilot at 40 just for fun. Own my own 4 seater. All the pilots in the US were trained in small planes and had to graduate to 4 engine planes to get there ATP ratings to work for the airlines. So there is a huge understanding in this country how airplanes work. Not so in the rest of the world. In the rest of the world, not as rich as the US is there is no private airplane fleet and no money for most people to train with. So they train in the class room and on simulators. They have no real hands on training unless they train here in the USA or in Europe. This is a HUGE issue. I would not dream of getting on an Ethiopian or Indonesian airplane because I have no idea who is flying them and what experience the pilots have and that also extends to the maintenance people. In the US the maintenance people have to sign the repair logbook. They need to be licensed aircraft repair men. It has been reported that we have had two similiar incidents with 737's in this country but OUR pilots knew how to kill the autopilot program.
5
Let me summarize. Boeing introduces an airliner that is "new" but not new, with minimum training onus (maximum profit margins), emphasizing to purchasers its similarities to previous 737s but barely mentioning its differences.
Defective readout in one such airliner prompts it to nose-dive, with its pilots not knowing how, or not able, to pull the plane out of the automated conviction that brings it down. See barely mentioning, above.
189 souls are lost.
FAA: no worries, guys. Update the software and rewrite the manual, and we're good.
Boeing: How does April 2019 sound?
FAA: Sure, fellas.
A reminder. 189 souls are lost.
Meanwhile: US pilots complain anonymously to NASA of sudden, inexplicable plunge in the nose of the Max 8. Discontent ripples through pilots of American Air and Southwest (per NYT reporting today) of the flimsy, barely-there mention of M.C.A.S in the Max 8 manual and the flimsy barely-there training requirements.
Next: another Max 8 goes down shortly after takeoff. 157 souls are lost.
40 other countries: in an abundance of caution, given the probability of the same (brand-new) model going down within minutes of takeoff, we're grounding this craft until the black box data reveals more.
FAA: No worries, folks. Software update coming soon and, hey, they're rewriting the manual.
Almost 48 hrs later: FAA overruled by POTUS.
This isn't about automated systems. This is about why Boeing felt no urgency about its obvious deficiencies, and neither did the FAA.
100
@Lex
It's the logical outcome of government that governs for corporations, where corporations are persons, and government agencies are led by leaders hostile to the mission and charter of the agency. It is the natural outcome of a philosophy that government is the problem, never part of the solution.
50
We are totally ruled by Capitalism. We need to back off a bit from the profit motive. Ohmygosh !!!! Socialism !!!!
3
@BBH
I'm a leftie but I don't want the Government to own all means of production which is what Socialism means.
Elizabeth Warren has it right, Capitalism works just fine as long as it's well-regulated.
2
I read the author’s article on this incident two days ago and found it to be very convincing. He points to the flawed design by Boeing as he also alludes to it in this piece. He raises an alarming prospect for this aircraft: Boeing May have a hard time to convince the public that the plane’s can be fixed by simply debugging a software. I think he may be right.
6
This is not an automation or technology problem. BOEING made a poor business decision to refit an old plane with new engines and save $$$. They tried a short cut that cost lives.
Those responsible were clearly unethical but should face a criminal activity
31
This problem started with a decision to relegate pilots in the cockpit to secondary status. The sales department at Boeing determined the way the 737 Max would be presented; as a money saver that could replace the previous version with little pilot training. Boeing decided that pilots could figure out for themselves what they would needed to review in the new plane's manual and more importantly what to do if a flight control problem developed- in flight.
This was done with full knowledge that the new plane had a potentially fatal performance characteristic. Rather than improving the design of the aircraft to compensate for a clearly unacceptable flaw, software and a system that compromised the pilot's in-flight control was tacked on like a software patch, a minor "problem" corrected with a few strings of 0s and 1s. This is nothing more than technological hubris in the service of quick profits in yet another super heated competitive market. There should be no system in an aircraft that can override a direct control command from the pilot at the controls. And there should be no aircraft cleared for commercial flight that requires a unique, tacked on combination of "sensors" (attitude indicators) and software to make up for a plane that is fundamentally flawed to the point that it cannot be safely and predictably handled by a human pilot.
36
Well said. I’m not an aviation technician, but it seems the MCAS system is akin to cruise control in a car that can’t be shut off by tapping the breaks or pushing an “off” button. The MCAS will disengage for a few moments once the pilot takes control of the yoke, but then re-engage. Even that wouldn’t be so bad if the pilots knew beforehand and knew where the MCAS “off” button was.
Did Boeing not put together a “Use Case” for the new plane and MCAS system that was derived from pilot input? That seems very unlikely, but looks to be the case here.
And the MCAS system was needed because of the larger engines and their placement on the wings. Now maybe that saves a few bucks versus creating an entirely new plane, but at what cost to human life?
These are all suppositions on my part, based on what I’ve read on the 737 Max and the MCAS system, and I could be wrong. All I know is that someone, or a group of people, made a series of incredibly bad decisions.
11
To be clear: MCAS is disengaged when the autopilot is on. MCAS is engaged when the autopilot is off. In the 737 Max original flight manual, the procedure to turn off MCAS did not appear. This is the best explanation I've seen (polish up your Spanish or read the English subtitles):
https://www.youtube.com/watch?v=sNlKWcjEB_4
10
@tanstaafl Thanks for pointing this out. I will add that, if MCAS is active, it won't actually allow autopilot to engage. So to switch back to autopilot from manual operation, the pilot must first switch off MCAS. If you do not know how to switch off MCAS, as the Lion Air pilots apparently did not, and perhaps the Ethiopian pilots also, you're stuck in a plane determined to point downward.
2
Wow, so shut off one auto pilot function and engage another, and fail to inform the pilot how to disengage MCAS? Negligent at best. Criminal at worst.
2
In every case mentioned above the root cause was a faulty sensor not the software. No-one is talking about why BOTH air speed sensors froze on Air France 447. Or why the angle of attack sensors were faulty on Lion Air (and probably Ethiopian Airlines).
Improving the software is fine but I hope they are looking at improved reliability in the basic sensors as well.
18
@Bob That's confusing to me, too. Is the revised software going to know that the sensors are faulty and ignore them? I don't see how software is going to fix all this.
4
@Bob
As a software developer we learn (the hard way) that ALL inputs to a software logic process will at some point be unexpected or incorrect data.
Therefore, reliable software MUST validate all data inputs, or be fully succeptible to unexpected and undesired behavior when the inputs are unexpected or incorrect.
This is software dev 101...
3
@kenzo
actually maybe it is software dev 201.
I have worked with tons of H1Bs that don't understand this need to validate inputs, and then blame the users when nonsense inputs they never bothered to validate in their code cause their programs to output nonsense (but
eliminating validation does speed delivery of the software and so earns them points with foolish managers)
1
There's a maxim in aviation that the first three things any pilot should do in the event of an in-flight emergency are to fly the airplane, fly the airplane, and fly the airplane. Most aviation accidents from the simplest to the most complex have more than one cause. In many recent cases, an automation failure was followed by a human failure.
Automation and advanced electronic flight aids make things easier for the pilots in almost all cases. The key is to make sure that pilots are prepared to immediately fly the airplane if the automation fails, and this requires a mindset that prioritizes old-fashioned aviating over troubleshooting. Get the aircraft stable at a safe altitude before trying to figure out why the computers aren't working. If this requires shutting them off preemptively, so be it.
Automatic flight controls should always be designed to give the pilots adamant warnings when they are not working properly or are intervening to override pilot commands.
The pilot must be able to to shut down any automatic flight aid quickly and independently, without turning off other systems first or wading through computer menus—the shutoff should ideally be a hardware switch or button.
Automatic controls are no substitute for basic stability. With all electronic aids switched off, an aircraft must remain hand-flyable by an average pilot (military aircraft with ejection seats excepted).
13
Couple points:
First, it's a bit of a stretch having a private pilot pick apart the problems of the commercial aviation industry.
Second, the author - like everyone else - is assuming they know what happened in not only one of these two accidents, but both. I would caution everyone not to assume anything, but apparently that ship has sailed. (I personally find it much, much more likely that pilot error - particularly in the case of the Ethiopian Airlines crash, which had a 200-hour copilot, who shouldn't have been allowed near that plane without a ticket - will play very heavily in both crashes)
Third, yes, automation is a problem. The great irony here is that the design philosophy of Boeing is that in the case of a problem, the pilot should be given the benefit of the doubt and given control of the plane; vs. that of Airbus, which by default thinks the plane itself is better equipped to fly itself than the pilots (sometimes with disastrous results - like the Air France crash that the author cites).
Much of this mass hysteria is unfounded, and rooted in international politics (the Chinese, who are trying to both best Trump and develop their own narrowbody jet, were the first to ban the 787MAX).
I go back to this point: never take third world transportation. I'm sorry, but it's no coincidence that these two accidents happened where they did. If there were more of a focus on the Ethiopian copilot, readers would be appalled that he was allowed in the cockpit of that plane.
11
@Mike
The third world didn’t design this plane or write the training manual. Your comment is hysterical.
2
I am a ex BOEING scientist. The problem goes to the heart of problem at BOEING that general public does not see. I say this reluctantly. With majority of the students with advanced degrees graduating are non-white then how come vast majority of the BOEING engineers are white? In fact the problem is alarming when you take a glance at middle level and upper management .... it is dominantly white and non-white like myself are unwelcome ..... I felt it so did other minority.
So no matter how hard it tries and how much money they throw at the problem it will be a second rate solution and their market will deteriorate over time. Case in point is their solution to 787, add more fire protection shields!
With China and other countries entering the market, let me assure you if adversary is willing to put the money they can start to take the market away from BOEING. There is nothing special or Holy Grail about BOEING. BOEING is white's company.
Just to level of with skeptics, I captured several IR&D grants and even captured Chairman Innovation Grants yet my biggest hurdle was these incompetent management who did not understand the very basic and more interested in swinging their stick, so I resigned.
6
The op ed did not provide any new information insight or analysis. There was no opinion, in my opinion
3
This opinion piece, like almost all of the reporting on this incident, assumes that there was a flaw in the plane's automation that was a major cause of the accident (notwithstanding the airline's fervent assertion that the pilots were trained to handle runaway trim, including from MCAS). No such conclusion is warranted, yet.
For example, consider the possibility that the plane was over weight (for the conditions) and/or improperly loaded. It might have exhibited many of the same flight profile characteristics as are shown on the sketchy radar data that's available.
I have seen no discussion of this possibility. But from what I've been able to glean on the Internet, the empty weight of the plane is just shy of 100,000 lbs. There were 157 souls on board, times 200 lbs./person (including carry-on and checked baggage) and you're up to 131,400 lbs. Add 1,500 pounds of fuel for takeoff and 13,500 lbs. for 3 hours of cruise (I'm guessing here), and you're up to 146,400 lbs (that assumes no additional cargo).
The density altitude appears to have been around 9,600 feet. At 146,000 lbs., the plane would use almost the entire 12,000+ foot runway to get off and would probably have lousy climb performance.
Maybe my back of the envelop calculations are off, or maybe this particular scenario isn't consistent with other known data. But the point is that there are lots of possible causes that don't implicate the plane's automation.
1
@Piper Driver You are absolutely correct; the eyewitness reports of smoke and unusual noises imply engine trouble, and given the factors you bring up, the margin of safety for an engine failure on takeoff was already slender on this flight.
However, given the similarities between the two 737's flight paths, I still think that grounding the aircraft was the right call. I don't want to speculate too much, but it's possible that there was an engine failure followed by an automation failure that compounded the problem and overwhelmed the flight crew. The vertical speed logs indicate that the airplane initially wasn't climbing, and the first thing a competent pilot wants to do when a twin loses an engine on takeoff is CLIMB!
2
@Piper Driver.
Do the math 157 x 200 =31400 NOT 131400. The plane was loaded at 46%.
The pilot is ultimately responsible for flying the aircraft, and even more so when a problem arises. If the accident investigations reveal that these tragedies could have been prevented by simply switching off the system and hand flying the aircraft(as the US pilots appeared to have done), then there is a case to be made for some serious revisions in pilot training with a focus on old fashioned stick and rudder skills and emergency procedures. Eligibility requirements for sitting up front should also be looked into. One of the co-pilots only had 200 hours of total experience, cmon!
5
So the way I read this is that the plane is fundamentally unstable in flight, tends to head in a dangerous direction (upward), which then requires special software and hardware to constantly monitor and correct the situation.
Imagine if your car had a natural tendency to veer toward freeway medians and brick walls. The manufacturer has installed computers and other parts to keep the car from crashing, as long as everything works fine. And as long as the driver does not misinterpret whatever these systems are trying to do at a given moment.
I would not wish to be anywhere near such a car when it was moving, much less inside it.
7
@JF
And that is why we should be very concerned about the full steam ahead efforts to develop driverless cars and trucks and let them on to the highways. Not only do new vehicles come with more and more automation -- some of them already have lane correction capabilities, at least according to their commercials -- but several companies are rushing to get a fully driverless vehicle to market. How happy would you be to be driving next to an 18-wheeler with no human inside? Hope that software works right.
2
I'm not typically a fearful flier. I spend a lot of time on an airplane. But this makes me nervous. Working in IT I know how many times code fails to take into account all outcomes. And how often it has to be fixed.
A human at the helm at least gives me a chance. Computers can't be rationalized with. They do what they were programmed to do and have zero stake in their own survival.
13
Let us be honest.
"Because the 737 Max had been outfitted with larger new engines that could cause its nose to pitch dangerously skyward..."
This is a design flaw. One should not have to create a computer program to override a design flaw.
Pilots will tell you that some planes are a joy to fly. The older model 737 flies just fine. Even the gigantic A380 has been found to be remarkably responsive and a pleasure for the pilot to fly. But a plane that naturally wants to pitch upward?
Back to the drawing board.
141
@Larry I read somewhere that this design change was to make the plane more fuel efficient.
2
@Larry
I read a comment in WAPO that stated that Boeing took an old air frame and slapped new, much bigger engines on it and to make them fit, they had to put them forward of where they were in the original design (which already placed the engines very close to the ground), mainly because they are so much bigger in diameter. This too-big-an-engine-clapped-onto-an-old-air-frame approach resulted in inherently unstable combination - Boeing's 'solution'? Add computer automation to try to make an unstable aircraft fly.
The FAA should have caught this but it doesn't have a proper leader; Trump tried and failed to nominate his personal pilot and nobody else seems to want the job at the moment.
2
@Larry: Your comment blames Boeing for what you call a "design flaw" -- yet you don't understand "stability" and "design" issues.
Boeing wanted to replace what are now inefficient engines on their base 737 airframe, with more fuel-efficient modern engines. This did not make the plane unstable. It did change its flight characteristics, especially during take-offs. So they added the MCAS software to "correct" for the greater engine thrust. This was not a problem -- automakers do similar things. They annually tweak car designs, and use software to "adjust' drive characteristics. Most modern cars limit their speed, using software, to120mph or 150mph. Auto insurers request this. So Boeing's airframe is ok.
Using automation/software to modify an airframe's flight characteristics is a common practice. The reason the "A380 [is] a pleasure for the pilot to fly" is because of its software fly-by-wire system -- the software makes it easier (and yes, a pleasure) to fly.
The problem with the 737 MAX series is not that it has autopilot and MCAS systems, making it easier to fly.
The problem is that Boeing's MCAS was not designed properly (taking into account pilot behavior) -- and most critically it was not tested sufficiently.
Boeing is at fault -- however the mistake they made is in insufficient testing of their systems.
The FAA is also at fault -- they lacked oversight and allowed Boeing to "self certify" design changes.
To see well-tested autopilots work reliably, look at SpaceX.
5
I thought the Lion Air (and possibly Ethiopian) issue was MCAS, which in my understanding is only "on" when autopilot isn't activated? Yet you said: "Sometimes a malfunction causes the autopilot to run haywire and put the plane into a dangerous state." The issue was MCAS not autopilot, right?
4
@Alex S
You're absolutely right. That was my understanding as well.
1
The problem is not so much the automation per se, but rather it's inexperienced and under-trained pilots who don't know how to recognise that they need to take-over, fly manually, and take action to recover the aircraft.
5
@PeteH What are you saying? Pilot does not need to train for the flaw in the design left because of incompetence at BOEING. I say this as ex-BOEING scientist.
Your approach to the problem is flawed.
Runaway Stabilizer Trim checklist. When the MCAS kicks in at an undesired time, it mimics Runaway Stab Trim except that reverse control column pressure will not brake the electric trim. “Stab Trim Cutout Switches: CUTOUT.” If the pilot’s of Lion Air and Ethiopian had done that, neither plane would have crashed. It was in fact performed by the pilot flying the ill-fated Lion Air plane the day before the accident, when it malfunctioned in the same way it did on the fateful following day.
2
Luckily I hate to fly and have no plans to. It’s not a problem for me. We can always go back to DC-8s.
1
@John Doe
Ironically, most older airframes like the DC-8 have horrible records compared to modern aircraft.
In particular, the 707 and DC-8 have fatality rates that are orders of magnitude greater than modern planes.
How soon we forget!
7
@SteveRR, I know, that’s what I’m talking about. We create our own Armageddons every day just for the fun of it apparently. What a drag if the real one never comes.
1
Interesting that Ethiopia refused to send the important Flight Recorder to the US for analysis. Thanks to Trump and his bullying, erratic tactics, many countries that always trusted the United States no longer do so. Trump’s insulting and belittling comments towards former strong friends and allies while embracing despots has caused a worldwide drop in admiration for America.
65
The real problem here doesn't seem to be the design, but the lack of training and information available to pilots on how the MCAS works, and what to do when it fails. From what I've read, it sounds like the MCAS functionality wasn't even described in the operator's manual, let alone their having been and emergency procedure for unwanted MCAS inputs.
As an analogy, imagine if your car's cruise control didn't disable when you hit the brakes, but rather you had to hit the little cancel button on the steering wheel. What would happen to you if traffic slowed down ahead of you? You'd probably want to have been told that the brake pedal would't turn off the cruise control. Now imagine if your brake pedal got harder and harder to push if you still had the cruise control on.
38
As and occupational accident and fatalities investigator (in an unrelated industry), my. observation is that "save money on crew training and maintenance costs" is exactly the wrong approach.
James Reason, the British psychologist long ago pointed out that in complex systems such as ships, nuclear plants, and aircraft no one person can understand the whole system.
Pilots need to practice, practice, practice for emergencies, not just in simulators but in actual flight with no passengers aboard. If anyone in a regulatory agency thought about saving money on crew training, then these two Boeing 'night-mare' aircraft crashes should put that theory to rest.
Boeing's reputation is shattered right now. They have very little 'due diligence' tp share with the public right now, and this lack of diligence they also share with the administration (not the engineers) of the FAA.
How is it that China was the first to ground the aircraft, and the United States was the last?
59
@Svirchev
Maybe because Boeing has no sway over Chinese legislators, but aircraft manufacturers' and airline companies' campaign contributions most likely influence US legislators?
16
"Many modern airliners are capable of taking off, flying and landing without any human assistance."
This is a positively wrong statement. Takeoffs are always done manually the old fashioned way. Much flying is done on autopilot. However, the autopilot is just like your dumb average cruise control. Just operating according to the pilots inputs and commands.
Human assistance is taking place all the time. Just simply because there are no such thing as artificial intelligence.
75
@Lars Chr. Svanholm
Is right. There are a lot of approaches, where you must disconnect the automation; at least the autopilot and hand fly the aircraft. Think of the Expressway Visual into runway 31 in New York La Guardia. Or the slam dunk (from the North) into San Francisco. "Cleared Quiet Bridge visual" , a tight turn outside the San Mateo bridge, and dive down, with gear out in an Airbus.
Sure, an Airbus or Boeing (I flew the 767) can do an Autoland, with suitable ground facilities, but I rarely did these in Line flying, just the Sim.
13
@skier 6 I would not state Lars is correct in his statement. He says takeoffs are always done the old fashioned way. Not always. Nor is it not possible. It’s just pilots feel more comfortable taking off and landing. I would prefer the auto pilot land and takeoff. The pilots can observe and correct if need be. Just my humble opinion.
@skier 6 Love your description of the SFO approach!
3
Yes, complex feedback systems are subject to opaque ills hard to anticipate and to fix.
But in this case Boeing elected to misuse automation to cut costs. Instead of designing according to well established principles to accommodate their larger engines, they opted to make a fundamentally unstable plane and rely to an unprecedented degree upon a sensor-software-computer kludge to maintain stable flight.
As it turns out, the system defies pilot intervention and is subject to unexpected malfunction that has yet to be tracked down.
It ain’t automation that is the culprit here. It’s marketing decisions “going where no man has gone before” and overriding engineering,
12
@John✔️❎✔️Brews As a ex-BOEING employee the matter likely went like this (trust me I have gone through this). Management guy hearing engineers multiple options, incompetent management accepted solution from incompetent engineer, who incidentally knew how to win argument (i.e. show lower cost).
I have no grudge or anything against BOEING but this is what is happening inside BOEING.
61
The FAA requires 1500 hours of "stick" time before becoming an airline co-pilot. As a result, most US airline pilots have come from the military or have many hours of fly by the seat of your pants experience in smaller planes. They tend to react and take more appropriate action such as switching off the auto pilot and taking control of the plane.
Foreign pilots do not have the same level of experience or training. Automation has been able to compensate for some of this deficiency, but contingencies happen.
21
Amen to that. You hit the nail on the head: Foreign Airlines (in emerging markets, mostly — which includes Ryanair and a few others in the EU) heavily rely on aircraft automation to mask a serious and economically-driven issue: lack of talent and experience on the flight deck. Some even go so far as to run their cockpits like a flight school, creating a kind of revenue stream with the savings in labor costs over what they’d pay for experienced and sharp pilots.
5
It appears to be more insidious than that. Apparently MCAS in only operative in manual flight. Engage the autopilot and MCAS is off. Earlier Boeing aircraft had a pair of switches to disable electric trim in the event of runaway trim. I think the MCAS is a different system and don’t know if it has a simple “off” switch. Given that the aircraft is apparently inherently unstable because of engine placement, it would not surprise me if there were no “off” switch. So... we are probably looking at a software upgrade.
2
@CC Laughable, I see a disturbing bias against foreign pilots.
3
Well written and clear as to the challenges ahead.
4
Watch some of the videos on YouTube showing commercial flights from the cockpit. After takeoff the pilots get a heading and a speed from ground control, dial them in and fly hands off for most the flight. Auto throttles, self-steering, auto climb, it seems auto everything. I could see how when something bad happens a poorly trained pilot could get in trouble quickly if they had to suddenly grab the yoke.
10
It seems extraordinarily dishonest to not mention the lack of a yoke in the airbus cockpit when bringing up Air France 447 in an article discussing aircraft automation that is critical of Boeing.
1
@max
Airbus have a side stick. Boeing aircraft have a yoke.
While the philosophy of the automation is different (Boeing yoke moves with autopilot inputs, Airbus sidestick doesn't) you get used to the different control systems. Same with thrust levers. Boeing thrust levers are "driven" by motors, so you see the levers move as thrust is modulated. Airbus thrust levers sit in a series of detents, and thrust can increase up to that max value, without the levers actually moving.
I preferred hand flying the Boeing (B-767)over the Airbus series, but hand flying an Airbus 319 is fun too.
I always encouraged my First Officers to disconnect the autopilot, and even Auto Thrust if they wished..to hand fly.
3
@max
A correction to my previous post..
Same with Auto Thrust engaged. Boeing thrust levers are "driven" by motors, so you see the levers move as thrust is modulated. Airbus thrust levers sit in a series of detents, and thrust can increase up to that max value, without the levers actually moving.
I preferred hand flying the Boeing (B-767)over the Airbus series, but hand flying an Airbus 319 is fun too.
I always encouraged my First Officers to disconnect the autopilot, and even Auto Thrust if they wished..to hand fly.
Simulator sessions always required extensive hand flying, often on a single engine.
2
Why? Airbus pilots are entirely familiar with not having a yoke. AF447 crashed because of inadequately award pilots more than anything else.
1
It is a nagging back of the mind tug that after the first accident it came out that even the pilots were unaware of the new technology and its ramifications. It does not seem reasonable that the manufacturer who thoroughly tested these airplanes and the FAA had no inkling there was a problem. Would delay have caused financial difficulty or worse a perception that the plane was not yet fully airworthy? If there was ever a time for a thorough investigation of who knew what and when, it is now.
4
I note that both flights are in relatively underdeveloped countries. Local ground crews, perhaps undertrained, are totally in charge of distributing the load. Excess load in the rear compartment pulls the tail down. This, in combination with the instability caused by the more powerful engines, may have overwhelmed the automatic correction system. I do not think that there is a way to check the weight distribution after the crash.
8
The author writes, “Many modern airliners are capable of taking off, flying and landing without any human assistance.” This statement is patently false. I am an airline pilot who has flown both Boeing and Airbus aircraft and there is no aircraft that I have ever flown or know of that can take off with the auto pilot on. All takeoffs are manually flown by the pilot. Though, it is true that the autopilot may come on in the first 30 seconds or so after takeoff and remain on for the rest of the flight, including the landing. Even so, an incredible amount of human assistance and input into the automation is required to manage the aircraft throughout the flight and safely to landing, even when everything is working perfectly. I don’t think I’m just nitpicking semantics, as this misunderstanding of modern airline cockpits speaks to the credibility of the author.
125
@Andrew J. Chisenhall
Thank you for this information. As a non-pilot, I'm curious whether commercial jets like this can land on auto pilot? I've wondered that since September 11, 2001 -- if the passengers on Flight 93 had succeeded in taking control of the plane back from the terrorists who had presumably killed the crew, would they have been able to land it with autopilot help?
6
@Nikki.... Yes, they can land on autopilot PROVIDED the aircraft is properly configured for the landing. A job done by the pilots....managing airspeed as flaps are deployed and deploying the landing gear at the appropriate time. No autopilot does all that.
Whether or not someone on a hijacked airplane could configure the aircraft is dependent on the skill set of the pilot that happens to regain control of the aircraft. I could do it because I was an Airline Pilot for 35 years and flew large jets. A fairly experienced general aviation pilot could probably be coached on how to configure the airplane while on autopilot ( assuming radio contact with air traffic control ),
but I would not be especially confident that a passenger at random would be successful.
It is not as simple as you may think. A properly trained pilot just makes it look easy.
16
@BBH
Thank you very much for your reply. Most of the general public (like me) have no idea how much is done by the pilots and how much is done automatically by the plane. What we see in movies can be wildly inaccurate depending on plot convenience. I still have fond memories of going in the cockpit of a military C-5 Galaxy transport when I was a teenager (my mom was in the army at the time) and being shown how the radar worked. I have great appreciation for the skilled men and women who make flying as safe as it is.
4
The series of Dehavilland Comet crashes in the 50's introduced the industry (and world) to the concepts of metal fatigue..and how the fuselage could just disintegrate after repeated expansion/contraction.
Unfortunately, Boeing is learning something new here. There must be a system or mechanical flaw which, when the exact confluence of events occur..causes the plan to think it is stalling and thus try to dive.
It is also interesting that both crashes occurred around the same latitude of the earth..
Is it possible that there is a equator/heat component with the systems??
3
like so many other facets of our lives, there is an over reliance on the perceived accuracy and efficiency of technology. technology can do a lot of things, but just like artificial intelligence, real intelligence has no substitute, and judgement in the skies requires well trained professionals, with experience, training, and practice. this "software" is described like a footnote in a manual - yet it's impact is way beyond a footnote.
2
Under the circumstances, he second crash might justify corporate manslaughter charges and huge insurance claims.
Boeing's solution, to the first crash, to which the FAA agreed was to allow the Max 8 to create an emergency situation, which the pilots were required (by a memo) to sort out.
NO, the proper solution is not to create the emergency in the first place.
Both Boeing and the FAA fully understood the causes of the first crash within a few days of getting the black boxes.
Why did it take the grounding of the type by the whole world to get them to take the only sensible action?
52
@PaulDirac
"Why did it take the grounding of the type by the whole world to get them to take the only sensible action?" you ask? Answer: Greed
6
It is hard to prove a negative. Impossible, in fact. But with reasoned, evidence based investigation and analysis to address any problem found, and a fix engineered, we should have confidence in the aircraft again. We have done this with many aircraft and initial flaws to great effect, resulting in the safest mode of transportation in the world.
The bigger hazard is the insertion of politics into the engineering process. Nothing good comes from that.
3
@Jon
You say "It is hard to prove a negative. Impossible, in fact."
That's simply untrue. The following negative statement: "The earth is not round", is an example.
2
I fail to understand why Boeing used some kind of mechanical detector to feed the planes attitude to the M.C.A.S system, instead of gyroscopes. Planes have used gyroscopes for almost all kinds of navigation and instrument information for many years.
The most basic things a new pilot used to learn was Needle, Ball, Airspeed, has that been discontinued? Most modern planes have two gyros, if they are out of sync the pilot should know what to look for immediately. As was pointed out, at least two incidences occurred and the pilots disengaged the system and took manual control.
This should have triggered an AD, Airworthiness Directive, and all planes should have been inspected immediately. As has been noted, the certification process has been handed over to the manufacturer instead of independent FAA inspectors. another example of Republican privatization of government responsibilities, this time resulting in the deaths of several hundred people.
You are now seeing at least one of the results of St. Ronnie's downsizing the government, not only the FAA but several regulating agencies like the EPA, and FDA where fatalities are not so obvious.
71
@David Underwood
The "some kind of mechanical detector" you mention is the 737's Angle of Attack sensor. It measures the angle the wings make versus the oncoming air stream. An angle of attack that's too high will trigger an aerodynamic stall. To prevent that from happening the plane's guidance software will push the nose down, hence reducing the angle of attack. If the AoA sensor is sending faulty signals this can result in mayhem.
There is no gyroscope in the world that can perceive this angle, hence the existence of AoA sensors. Virtually all commercial planes flying today have such a sensor.
35
@David Underwood
I think the issue here is that the orientation of the plane is not the same as angle of attack. A plane is tipped backward while climbing, but the angle of attack with the oncoming airstream is still small. But it the plane tips back while still moving parallel to the group, the angle of attack becomes large and the plane stalls.
6
@David Underwood
AOA (Angle of Attack) is related to attiude, but may diverge due to the winds or convection currents.
The plane's wings "feel" a composite of the attitude and air flow.
Gyroscopes show you the net attitude, which can be far enough from the AOA to render them unreliable for this function.
5
According to reporting, the larger engines had to be moved forward to provide necessary ground clearance. This caused the aircraft to be inherently unstable. Boeing's solution was to install an automated system to provide the needed correction.
Another solution would have been to install taller landing gear which would have been much more expensive than the automation. I would venture to guess, and this is just a guess, that the instability was discovered after the design of the landing gear was set. The cheapest quickest route was then to fall back on automation and software.
Another factor here is that these planes are designed in a computer instead of the old way of building prototypes and testing them in the air as the design progresses. This is much more expensive and slower. That means unit number one is the prototype.
So the issue I raise is the policy of relying on automation, not just to fly the plane, but to create it in the first place without traditional prototype testing.
75
@Bruce Rozenblit It's my understanding that the new engine placement doesn't render the Max unstable per se; more precisely, it makes the airplane fly differently and makes it a bit less forgiving. The issue is that Boeing programmed the flight control software to make the Max feel the same as the previous models, so that a pilot could fly it under the same type rating (a training classification) and the airlines wouldn't have to pay thousands of dollars per pilot for added training. The problem with this ploy is that it's only effective if the flight control system is working properly.
The lessened stability was easy to predict from simple aerodynamic calculations. The reason the landing gear is short, as I understand it, is that Boeing reused the same basic landing gear from the previous models to keep costs down. Since 737 gear retracts inwards (easily visible in the lead photo), it can only be so tall before the wheels run into each other.
19
Since they fitted larger engines to an existing fuselage, they most certainly did "discover" the instability after the gear design was set. But they have superbly capable aeronautical engineers who would have known from the outset that the revised engine position would lead to instabilities in certain operating envelopes. It isn't something they chanced upon in testing.
No, the blame here lies in the business decisions, not the engineering ones. It was a business decision to overcome the instability with software instead of redesign the landing gear. It was a business decision to sell the software system as not requiring training.
Finally, your comment about designing in a computer as being slower and more expensive than physical prototyping is incorrect. It is completely opposite to reality.
33
@C Walton
Thanks for your input, but that's my point exactly. If the aircraft had the proper height landing gear, the conditions that precipitated this condition would never have occured.
Whether the condition is described as instability or lessoned stability, the result is the same. The thing fell out of the sky. I liken this situation to car that doesn't track in a straight line. The driver can easily keep the car going straight, but should not have to. The reduced stability creates a condition for something to go wrong, and it did. Over 300 people are now dead. I'm not an aerospace engineer, just an electrical type. But it seems to me, the way to make an aircraft as safe as possible is to make it as inherently stable as possible. That obviously did not happen which is demonstrated by the need for automated correction system.
20
Several local witnesses interviewed by Reuters reported smoke and debris trailing from the aircraft before impact. Have these been followed up? Also, several articles have referred to the loss of radar tracking prior to the time of impact. Would that have been a transponder failure? Why? This indicates possible physical damage or fire on the aircraft prior to impact. The Lion Air investigation indicated the plane entered the water intact, engines running, and no indication of onboard fire. These would seem unrelated to the autopilot operation unless it caused such extreme maneuvers as to damage the airframe.
7
@Gordon Sell
I'm sure all leads are being pursued. However, eyewitness accounts can be notoriously unreliable -- as an example, the crash of TWA flight 800 off the coast of Long Island. Witnesses claimed to have seen a "streak of light" they believed might have been a missile heading toward the plane. However, the cause was eventually determined to be an electrical spark in the center wing fuel tank. (The wikipedia page for TWA flight 800 includes all the details, including how the eyewitness accounts are explained, if you are interested in further info about that crash.) So I'm sure the witness accounts will be investigated, but will not be used to rule any other possibilities out at this time.
3
@Nikki Living in Brooklyn at the time, I remember well the shock and controversy's surrounding TWA 800, and I'm well aware of the potential inconsistencies of eyewitnesses. However that was at night, the Ethiopian reports are daytime, closer, and more consistent. Judge for yourself: https://bit.ly/2ENO1kF If the NYT blocks the link, just google "Ethiopian plane smoked and shuddered"
Flight automation systems are great ... as long as they work properly and as expected. The key takeaway is that pilots need to practice, in simulators, how to respond to the unexpected and the unanticipated. As the saying goes "stuff happens" and practicing what to do when it does so that decisions come from memory and are not made in the heat of the crisis is crucial to proper response when systems do not do what was intended.
Many incidents attributed to "pilot error" (e.g., AF447) are more properly attributed to "pilot ignorance" due to lack of proper practice and training.
3
Any automated system - and every aspect of modern life depends on them - presents the hazard of bad design. It's pretty clear from the evidence we have that the MCAS system in the 737 Max might be flawed from many points of view. It was perhaps an inappropriate application of automation at the root of its inception, because it may have been intended to create a digital bandaid over an analog problem with the physical design of the aircraft. Or it could have a flawed input design (lack of sufficiently redundant data for such a crucial system, e.g., 2 pitot tubes instead of 3 or more), flawed implementation (too much correction over too long a period of time), and flawed operator interface, including a lack of training. None of these potential flaws makes automation the enemy. Bad design and bad intentions (what I would call sales interference in engineering decisions) are the enemy.
10
@Pete
'A digital bandaid over an analog problem' - That's it exactly. The 737 max needs to go back to the drawing board (or CAD program, rather) and be reworked so that the engines and airframe are in harmony. Basically what they have now is a not-great design, and they're trying to teach the computer (and pilots) to fly that questionable design. Boeing will of course deny that until the cows come home.
The flight envelope of modern jet airliners should be getting bigger, not smaller. There's no reason to be building hard-to-fly airplanes these days. https://en.wikipedia.org/wiki/Flight_envelope
6
@Pete
Good stuff here. A neat and concise summary of the situation as it exists.
1
Increasingly automated things like airplanes and surgery aren't always better. Knowledgeable people -- not luddites -- can see that new technology may cause malfunctions that claim human lives, as the new Boeing 737 MAX 8 planes have done. As robotic surgeries have done, too. The terror of airplane crashes, and pilots unable to deal with crisis in the cockpit, due to new technology, is unacceptable.
The crews all over the world of the new Boeing MAX 8 jets were not trained specifically for the computer tech changes wrought by the manufacturer. We await Boeing's explanation of why these 2 new planes crashed in Indonesia and Ethiopia. The NTSB's report on the similarities between the two crashes will shed light on the Boeing company's attention to their planes. More technologically advanced isn't always better.
3
If concerns about automation in the cockpit take hold, it could impact Airbus and the entire industry as much as Boeing. Airbus was a leader in this, and Boeing is following. The Air France Flight 447 crash in the Atlantic in 2009 was in substantial part an automation issue.
The subject was dodged in public, but active among pilots even then. Check the phrase "Scarebus" for those with strong feelings back then.
This subject is larger than Boeing, and larger than this one model of airplane.
I'm glad to see it examined, finally, since I feel the industry has gone too far in the direction of computerized controls of flight. It is vulnerable. It works great when it works, and kills when it doesn't.
7
..just like human pilots. Except that the automation's safety track record is far better.