Investigators including Boeing are going to have to determine whether or not this crew is properly trained and of course, Boeing is going to have to focus on whether or not they have a software or technical issue.
Crashes are always the nexus of a chain of unfortunately and unlucky events. I have no reason to suspect this crash is any different. The only thing I'm confident of is that no matter what the problem is, Boeing will make changes and pilots will receive additional training, and perhaps be required to do simulator time.
Once the 737-MAX is ungrounded I will have no qualms getting inside of one.
Yeah, but it doesn't take Nostradamus to make that prediction. If Boeing doesn't fix their issues, they won't sell planes.
So, say we blame the pilot for incorrectly deactivating the faulted system within a time-frame that allows for recovery; which I think is a silly thing to do : at what level of added complexity and extended abnormal checklists do we transfer the blame from the pilot being unable to follow a checklist thoroughly and quickly enough, to the airframe manufacturer for producing a product with so many potential failure vectors that were added primarily as a cost-savings or for-profit desire?
The 737 MAX is unstable inherently for the sake of mileage, to reduce the overall cost per ticket, and increase corporate profit. I understand the importance of cost/benefit analysis, but at what point are corporations liable for teetering towards the profit side at the sake of safety?
This is blatantly false. The MCAS system isn't running a PID control loop or anything like that. It's applying stupidly simple logic to its inputs and giving equally simple outputs.
At worst you could claim the aircraft is less stable than the original 737. But more specifically I'd say the 737 Max exhibits undesirable control behaviours at the approach to the stall. The MCAS system was required to make this behaviour certifiable.
But the new Boeing 737 Max planes exactly at that dangerous moment behave differently to the pilot's input than what the pilot was trained for, because these new planes have different geometry. They are new, but being sold as the "same old." Selling point: "no pilot's retraining needed."
Boeing of course knew that planes behave differently, but actively tried to hide that, as much as not even reporting and documenting to the pilots that there's a new device built-in ("MCAS") to "move the controls itself" differently than what the pilot would do based on this training.
Now, if the MCAS misbehaves, the pilot is supposed to recognize that the "moved controls" are undesired, manage to turn the MCAS off fast enough (even using a circuit breaker! according to the Boeing's explanation after the first crash), and then again rescue the plane which behaves differently than the plane for which he's trained to have built-in reflexes!"
The pilots are supposed to be trained in the simulators to be prepared for the behavior in extreme situations, not to have a new plane that behaves by-design exactly not as they are trained.
The way they were trained, when some undesired movement occur, their reflex reaction corrects the problem. Not so in this case. Their trained reflexes didn't help. Instead, the faulty MCAS continues. That's why even the circuit breaker step is mentioned. But even after turning the MCAS off, the plane still behaves differently because it is actually of different geometry.
All that can happen at the moment the plane is not high enough to be safe to do enough maneuvering instead of hitting the ground.
EDIT: responding to the answer under this post:
> you seem to be assigning a lot of weight to a very minor system
"I" seem? At this very moment there is a world-wide belief that what you name "very minor" issues lead to death of 350 people in only 5 months time-span, to the point of grounding all the MAX planes.
Also relevant: https://news.ycombinator.com/item?id=19398267
[1] They are flying airliners for a living, not small planes.
The AoA sensor, previously largely irrelevant in civilian aviation as posted by an earlier poster, became a deadly concern seeing as a malfunctionimg sensor pumping inaccurate data to a downstream control unit with seemingly no way to validate or reject incoming bad data would respond just as happily to a reading 20 degrees off of where it should be. That this happened in a manner completely foreign to someone thoroughly experienced with a 737 just exacerbates the risk. Especially when said response is no longer overriden in the traditional response in old airframes. This is a usability regression, and should have been explicitly documented.
You seem very convinced that MCAS is some trivial system when it is very clear physically and legally speaking that that +/- 2.5 degree stabilizer movement at the right time is what the air-worthiness certification is dependent on.
>on re-cert as 737, and failure to train
Just because the damn thing flies doesn't mean basic automation system architecture, design, and ethical principles stop applying. You do not make a decision for an operator/user and not communicate the importance to them ahead of time. Doing so fundamentally changes the nature of the device.
To put it another way:
>"If you change the working parts, you make a different machine."-The Protomen, Father of Death
Or if you prefer someone a bit more established here's Edward W. Demmings view on it:
>"Every system is perfectly designed to get the results it gets."
From which comes the corollary:
>"Given two systems, if one produces results irreproducible in the other, then the systems are not the same."
>On Software that has the potential to kill people
Has no one learned from the lessons of THERAC-25? How many more need to die before "hide interlocks in software and skimp on training" stops claiming lives?
I fight the attitude that leads to these types of poorly thought through decisions every day in far less life threatening systems. Seeing it happen in such a high stakes industry just makes it all the more painful to have to endure.
>on appeals to occupation/authority
You don't need a pilot's license to connect the dots. You just need time, the right skill set and exposure to engineering in multiple contexts, and enough exposure to human social dynamics to realize that the MCAS augmentation is exactly the type of innocuous looking change to have slipped through the cracks. When you're in the trenches developing highly complex systems, you are in a highly faith based environment in the sense that while you are working from empirical measurements and simulations you have to have faith that all the relevant questions have been asked and answered. Engineers frequently discount the impetus of Sales/Business pressure, then turn around and don't question the deadlines those commitments made elsewhere, and accept the consequences thereof (I.e. questions not asked/answered due to time constraints).
As I said before. I'm waiting for more data. Even if MCAS isn't involved, the above mentioned issues are gross failures of system implementation that need to be remedied or otherwise addressed.
I'm not arguing it's one and done, but even if it turns out leprechauns swarmed the plane en masse disassembling it in flight, enough information has come to light that anyone can see there has been some serious ball dropping going down; ball dropping so serious that uninitiated customers are speaking out in discomfort instead of simply "leaving it to the eggheads".
I am very vocal on issues like this. I have been told I have a knack for the highly technical subjects, the inquisitivenes to run down things I don't know, and a bullish propensity to follow the facts wherever they lead. I see it as my responsibility to ask and find answers to the questions others don't even know how to ask, and to present them as best I can so that others may understand and come to their own judgements. Hell, I even learn something every now and again.
If that is your definition of non-triviality then the ash tray in the toilet is also a non-trivial system since the certification is dependent upon it being there.
MCAS is only there to compensate for one part of the flight envelope where normal operations shouldn't be taking place.
Point being, without that system the aircraft would still fly. In fact Lion air wouldn't have happened without it.
Of course there could be a fault (beyond the nonsense of it not being triple redundant). Or it could be something unrelated to MCAS which is my own pet theory (partly based on the reports of pitch excursions with the autopilot engaged where MCAS is irrelevant).
It's a minor system. There is no firm evidence yet that it was responsible for this crash other than some ADS-B data which only really shows that the aircraft was having control difficulties.
Pilots are trained not to get near the stall. Unless something has gone seriously wrong an Airliner will never go near that region of flight in normal service. Most airliners have stall behaviour which is likely to enter a spin or otherwise difficult to recover from, that is why most are fitted with stick pushers which prevent a stall ever occurring.
There are lots of other systems which can control trim on an airliner (the Mach trim system, the auto autopilot etc). Each of these have failure modes that can lead to a runaway trim situation.
The MCAS system is trivial. It may have a flaw which led to both of these accidents but it isn't the world is falling and Boeing tried to patch it with different software.
The 737 MAX is actually proof that even in the modern day, a basic grasp of philosophical principles is relevant.
Namely, being able to identify when a redesign of a plane becomes substantially different enough to no longer be the plane it was based on anymore.
The MAX blew past that line as soon as the system layout changed such that the pilot's mental models were no longer capable of accurately predicting the behavior of the plane.