Hacker News new | past | comments | ask | show | jobs | submit login
Why Investigators Fear the Two Boeing 737s Crashed for Similar Reasons (nytimes.com)
149 points by ejstronge on Mar 14, 2019 | hide | past | web | favorite | 112 comments

From pilot forums..

Actually Boeing sent out a memo to the pilots of the MAX after Lion Air crash explaining what may have caused it and what to do. Here is basically what is happening. Upon climb out the pilot will engage the autopilot with it in NAV mode. At this time flaps are still extended for climb out and the aircraft maintains the climb as indicated on the FMS to the assigned altitude. Once the flaps are fully retracted the MCAS engages. At this point the MCAS looks at several things but one being the Vertical Pitch Indicator to either adjust the trim up or down as needed for climb or possibly level flight if needed at this point. The Vertical Pitch Indicator is reading on some MAX 8's 10-20 degrees higher than what the aircraft actually is pitched at. This engages the MCAS to lower the Trim to level the aircraft as needed. Well at this point the plane over pitches nose down because the Vertical Pitch Indicator is saying the plane is pitched up, which it is not. The pilot then will use the Trim on the yoke to trim up also pull back on the yoke and both the MCAS and the pilot are fighting to keep the plane level. Once the pilot stops with the trim adjustment the MCAS goes to full trim down due to adjust for the 10-20 degree increase in the Vertical PItch Indicator it was getting, which is incorrect. At this point the aircraft is sadly most likely in downward angle and unable to recover from it. Boeing did say in a memo what could be done to help stop this during flight. The pilot's can manually override all autopilot functions including MCAS, either first or at that point put the Flaps extended one notch to disengage the MCAS.

Furthermore, it seems that the MCAS acts by moving the entire horizontal stabiliser... whereas the yoke moves the elevators only. This means the MCAS will always "win" in a metaphorical tug-of-war between the two controls.

However, the incident of B737 diving into the ground has been observed before, eg: https://en.wikipedia.org/wiki/Flydubai_Flight_981 and https://en.wikipedia.org/wiki/Tatarstan_Airlines_Flight_363 , on aircraft not equipped with MCAS. In those cases a possibility of a mechanical malfunction of elevator controls has been raised (but not confirmed) as a possible cause of the crashes.

The manufacturer is telling pilots to extend flaps to disengage MCAS? How is there not a tactile illuminated button in all those 3,000 knobs and switches that the pilot can simply flip to disengage that system? This is just getting weirder and weirder.

That seems quite wrong - maybe got garbled somewhere. Boeing’s official instructions to to all operators after the Lion air crash was to follow existing stabilizer runaway procedures by flipping two switches in the center of the aircraft to disable electronic control of the stabilizer.

(You can see a video of this procedure from a few years ago: https://www.youtube.com/watch?v=3pPRuFHR1co&t=154)

Thanks, that makes much more sense. Maybe they figure it's faster in an emergency to simply pop out the flaps one notch instead of the pilot attempting to find a button which they rarely press.

Although, truth be told, I imagine every 737 pilot is keenly aware of that button's location now...

I would have thought that was true after the first crash. I'm honestly baffled how it could happen again. If I was aware of this issue from the Lion Air crash, wouldn't every pilot in the world know about it?

I mean, if a plane crashes, doesn't every pilot in the world inquire about what caused it? I wouldn't think that would escape a single pilot. Or do I have it wrong? Do they not perceive crashes the same way the general public does?

> If I was aware of this issue from the Lion Air crash, wouldn't every pilot in the world know about it?

Knowing that some new mechanism exists that can mess your flight doesn’t necessarily help much:

The truth is that the plane even with that new device turned off behaves differently than the plane for which you trained!

The cause is Boeing’s lie that it’s “just more efficient old plane” when in fact it is a new plane with different and worse flying behavior. The device was implement to hide that fact.

So even if you as a pilot know about it, until you diagnose it and turn it off the plane can already be in the almost unrecoverable state. But once you turn the device off you‘re supposed to save the plane for which you weren‘t trained and in the alteady hard-to-save state!

> wouldn't every pilot in the world know about it?

Perhaps so, but with alarms blaring and potentially misleading data on the flight displays and the aircraft bucking around near the terrain at 400 knots... Things might be difficult to put into context.

Investigations tend to keep things hush till they have final conclusions, and everyone involved is scared of lawsuits so doesn't want to talk about problems or workarounds on-record. Talking about a workaround or distributing a fix before it's vetted by all the right people and procedures could also be legally risky if the workaround causes other issues.

That tends to mean a lot of issues re-occur multiple times before they're resolved.

Hush? It's an airplane. Shouldn't the same models be grounded?

Of course, it didn't help if Boeing blamed the pilots (if it was true?). This would imply to other pilots that everything is ok with the planes.

Even a plane with a bunch of faults is much safer than a typical journey by car.

Ground a few planes, and the extra car journeys caused by people choosing to drive due to disruption to their travel plans easily causes more deaths.

Humans are particularly bad at handling unlikely events (such as the risk of a plane crash), and in those cases, their 'gut reaction' frequently is the wrong one.

Cars have a lot of issues, but those issues are more forgiving than a similar issue in a plane.

If my cars engine has issues, I can pull over. If a plane's engine has issues, there's a good chance its going to kill everyone on board

If I had the choice between a 'typical journey by car' vs a 'plane with a bunch of faults' I'd definitely prefer the typical journey by car.

A 'plane with a bunch of faults' could range from a non-issue to an imminent crash, a typical journey by car tends to do what it is supposed to do, it is the 'exceptional journeys by car' that are the problematic ones.

Air travel is safer statistically because most planes do not have 'a bunch of faults', when you start out with faulty planes the stats will rapidly turn against you.

The FAA did put out a notice to pilots last year, I don't know if that kind of information reached Ethiopia Air though

> Following the [lion air] crash, the Federal Aviation Administration issued an emergency directive warning that faulty AOA data could trigger a dive, and requiring pilots be made aware of a "runaway stabiliser" checklist.

This trope is getting old (though you might not have intentionally insinuated it) - Ethiopian Airways is not some backwater airline: it flies into the US and governed by FAA rules - so yes, that information reached Ethiopian Airways as they have a direct line of communication to the FAA.

Sorry, it wasn't a knock on Ethiopian Airways - they're a Star Alliance member, I couldn't imagine their safety standards are different from any top-tier airline.

I meant it more as in I don't know how these things propagate internationally. I couldn't find a mention if the European EASA also put out a notice, or if the FAA one gets distributed worldwide (directly to pilots? airlines? via local regulators?), or how it works. You say they get notices from the FAA since they fly to the US, does that mean e.g. all airlines that fly to Japan also get Japan air safety notices? The Japan regulator wouldn't put out their own notice on the 737?

I'm honestly trying to understand how the notices get categorized, prioritized, "deduplicated" and distributed to everyone who needs to know without overwhelming pilots with unrelated stuff.

The FAA deals with airlines, which figure out how to disseminate the information internally.

> I'm honestly trying to understand how the notices get categorized, prioritized, "deduplicated" and distributed to everyone who needs to know without overwhelming pilots with unrelated stuff.

Any international organization that has to abide by laws/regulations in different jurisdictions has to be competent at this. There will be bureaucrats and infrastructure to deal with the bureaucracy: country reps, lawyers, former pilots, endless mailing lists, etc.

Tons of reasons this can happen. Best analogue I have is to patched security vulnerabilities in software still existing in some systems because a sysadmin isn't aware.

You mean something like STAB TRIM CUTOUT located prominently in the center console under the throttle?




Huh, the pitch indicator is known to be off by a huge margin on some planes and the problem is not fixed?

The problem is fixed by extending the flap one notch to turn off the MCAS. no big deal.

Flap deployment speed limit on the 737 Max is 230kts. ETH302 was already at least 50% beyond that.

So now you have an aeroplane with its flaps ripped off, perhaps asymmetrically for additional control challenges. Next step?

That's not a fix it's a workaround. The pitch being off is the problem. Turning off mcas solves it but it does no4 fix mcas

I’ve been trying to understand how the vertical pitch measurement is supposed to be suitable for use as part of the MCAS decision process at all. Seems to me wind shear and other atmospheric effects are going to make an airspeed based measurement of angle of attack _always_ subject to condition dependent measurement error. The fact that the pilot is assumed to have a more navigationally useful estimate for angle of attack (in that they are able to know when the angle of attack measurement from the sensor is wrong) makes me think there is something fundamentally flawed about the way this quantity is being estimated by the existing sensor — the modality of measurement might not be an appropriate way to control this flight process if atmospheric effects can bias the measurement in a way that can be calibrated by pilot knowledge but with these corrections not reintroduced into the flight process ...

Planes can have multiple AoA sensors, and often do. An indicator which makes the pilot aware of divergent AoA sensor readings is a paid upgrade.

Though I also have to agree, as soon as this sensor's input became crucial to determining operational boundary conditions, the system needed some rather stringent hardening. It's a bit like leaving your auto-scaling load balancing infrastructure exposed and getting upset at your AWS bill the first time someone DDoS's it to inflate your hosting footprint.

Yes, like if your car overaccelerates what you have to do is just turn the volume one notch up.

This is more like car over-accelerates, open the boot door.

I'm sure no offence was intended, but I'd be quite upset if I heard an engineer say a fix is "no big deal" if I were grieving a lost family member.

Pretty sure that was sarcasm. Boeing is basically saying, there is an automated system that causes planes to crash, and to turn it off you need to perform some completely unintuitive sequence of actions. These planes should not be allowed to fly.

This is the most deadpan comment I've ever read.

I thought MCAS was only active when autopilot is disengaged?

No time to act as discribed at 1100 ft AGL (see crash site terrain hight) and another one: NO MCAS if flaps, if autopilot, etc...any co-relation between MCAS and radio altitude and/or GPWS ? So MCAS can act even below 1000 ft AGL ? well done boeing !

All correct ! but NO time to execute all this good advice at 1100 ft AGL !! look at crash site terrain hight ! jeez

Based on what I'm reading, it sounds like a massive liability is on their hands.

You're being downvoted, however that is correct.

Norwegian Air was the first out of the gate with compensation demands on Boeing.[1] They've recently been in a perilous financial condition, so I'm sure this situation is hurting them more than some.[2]

Boeing or their insurer/s will eat a large payout to victims and I'd expect a bunch of airlines to receive some manner of compensation for the mass groundings.

They can afford it. $9.5 billion in cash and $11.8 billion in operating income the last four quarters.

It's really the reputation damage that hurts a lot more than the direct liability. That damage can last a decade or more and spur long-term erosion (lost sales also obviously hit the service business, which Boeing has worked to expand). They've seen their market value go from ~$80 billion to a recent high of ~$250 billion over just 30 months. That puts them among the elite global corporate giants in valuation. That's based on the way they've been outperforming Airbus the last few years and the improvements in their business that took profitability from $5b to $10b+ per year. Especially with China's state-made 737 competitor inbound, now is a very bad time to be screwing up in that segment. China would love nothing more than to use an opportunity like this to replace Boeing and launch their competing Comac C919 into the market (set to go on sale soon).[3] China can effectively throw a state switch and force all domestic airlines to buy the C919 when it becomes available; the mess with the 737 just makes it that much easier to justify.

[1] https://www.marketwatch.com/story/norwegian-air-expects-comp...

[2] https://simpleflying.com/norwegian-air-facing-collapse/

[3] https://www.foxbusiness.com/features/chinas-boeing-737-max-c...

The C919 has been almost ready for a real long time now. Comac can't even get their license built DC-9 up and running. As of last year I think there were five flying a total average of around one flight segment daily. The C919 is an entirely new design intended to have more domestic Chinese content — fat chance it'll be in production real soon now, even less of a chance it becomes a serious contender.

Meanwhile the Russians, who have significantly more experience with airliners, released the SSJ-100 to fairly positive reviews. Unfortunately the after sales support was trash and airlines struggled to keep it flying. Selling a competitive airplane is far more than just the plane itself. There's no way Airbus or Boeing see Comac as a threat.

The real threat was Bombardier. Their C-series was an extremely competitive offering and Bombardier has the support infrastructure in place to sell it. Both Swiss and airBaltic seem to love it. Of course, Airbus bought a majority stake in the program.

There's always a massive lawsuit when it comes to airplanes. Any crash, ever, they're gonna have to pay out. So that's the safest bet you can make.

Ok, so disclaimer: not a pilot. At all. But a comment someone made here in relation to the A380 being retired sprang to mind.

The question came up of why not drop from 4 engines to 2 (on the A380)? Someone answered that as soon as you change the engines you have to redesign the wing. As soon as you do that you have to redesign the fuselage. At that point you may as well just design a brand new plane.

So, my understanding of the 737 Max design is essentially that Boeing recycled the aging 737 design by putting more powerful engines on it. To make this work, they had to move the engines (and maybe the wings) forward? Is that right?

So the MCAS system seems, at least to me, designed to compensate for an engine, wing and frame configuration that was never designed for.

Add to that that Boeing _seems_ to have cut corners here and given what I think is called a common rating to the 737 Max (with existing 737 models) and not really educating pilots on what MCAS is, why its necessary, when it might kick in and how to override that if the plane is behaving incorrectly.

So, complete non-pilot and non-aeronautical engineer that I am, that seems... bad.

What's more, Boeing's reaction to the two crashes seems woefully tone deaf. Two crashes of new aircraft within 6 months of each other in perfect weather... caution would seem to dictate to ground them. I mean they grounded early 787s when they found an issue causing batteries to catch fire. What's different now and then? That issue I don't believe led to any total loss of the aircraft and those board, not even once let alone twice.

Add to that that the AOA sensor (from what I've read) doesn't have any redundancy and it just seems like a bad time is inevitable.

Could Boeing really have cut corners in this way? If so, that seems woefully shortsighted.

We're not going to solve this problem in a HN forum. Planes have a ton of hours put on them before a passenger ever steps onboard. This is a heavily regulated industry; and a lot of trust is placed in Boeing and the FAA to only build and certify safe planes. People make mistakes and that is why so many people are involved using best practices in the development of complex systems such as these.

The FAA burned all the trust and goodwill it gathered in the last 20-30 years -- depending who you ask -- within an extremely short time frame: on March 12 between 7am and 10am Pacific or so. It was pretty impressive. It was crystal clear the EASA is offering a face saving opportuntity to them by using the member states to do airspace bans but not suspending the planes of their respective airlines. Then, a few hours after that, when the FAA decided to continue the wave the "I am with stupid" flag they did the unthinkable and declared the 737 MAX not airworthy contradicting the FAA. This has never happened before. At this point the trust in FAA is gone and it is seen as a political and corporate puppet.

It was reported a day or two ago that Trump had a phone call with the CEO of Boeing, with the result being continued operation. Then, Trump makes the announcement today about the grounding, seemingly to try to take some credit for it.

It’s insane. There must be some despondent folks at the FAA right now. Years of trust squandered.

Is it possible the US have grounded them because Canada banned them in their airspace?

If the FAA didn't cave in immediately from the EASA issuing "do not operate the airplane" then it certainly wouldn't from Transport Canada. These two do not quite carry the same weight, you know? 150M passengers enplaned and deplaned in Canada in 2017 as per https://www150.statcan.gc.ca/t1/tbl1/en/tv.action?pid=231002... and https://ec.europa.eu/eurostat/statistics-explained/index.php... says this number was 1043 million in the EU. According to https://www.bts.gov/newsroom/2017-traffic-data-us-airlines-a... even the USA is below that nunber with 965M.

Thank you for a detailed answer.

> There must be some despondent folks at the FAA right now. Years of trust squandered.

Can you imagine the real workers, now some idiot Trump named to be deputy director and currently is the acting FAA director wiped out so much of their work... oh dear.

Even if they had been 100% sure the plane is sound, just to keep the larger status quo they should've acted in accordance with the EASA.

The Space Shuttle had a tonne of hours put on it before we learned of the O-ring problem catastrophically.

Everyone was trying their hardest and did their best, everything was built to spec but there were organisational and communication problems. The message that the dreaded O-rings weren't good for a cold weather launch was not heard or understood. No malice was involved but, when you have a schedule and deadlines with lots of large organisation hubris these things happen.

In only the broadest strokes is the current Boeing situation analogous to the sad fate of the Space Shuttle, however, just because a lot of testing has gone on and just because there is a lot of regulation does not mean that the layman appreciation of the situation is wrong. The layman understanding can be big picture rather than bogged down by intricacy of details.

These calls always seem easy in the aftermath, when you have the benefit of hindsight. In both of these cases, billions of dollars are on the line, but the truth is, there is always some risk. There is, as we've all heard, a greater risk of us dying in our car than any type of air catastrophe. Yet we still drive. How many flights have these planes taken? I can almost guarantee that Boeing did not build this sensor; the sensor has a spec sheet, its operational capability was verified in a laboratory and the FAA approved it. If the problem was not recoverable, they would have grounded the planes after the first accident; it appeared to be a training issue. Anyway, we can only make decisions with the information we have; it was the satellite data that suggested there was a link. It was then that the direction was clear and the call was made; this system is too dangerous as it is.

Making decisions based on fear is a fools errand. It is how you spend trillions of dollars waging wars combating terrorism when the risk of a terror related death is the US is lower than getting struck by lightening.

But the choice here is not "fly or drive", the choice here is "fly 737 MAX or fly another plane". I expect very few fliers have the budget and time to change flights, and those who can, will.

No they knew of O-ring burns before the catastrophe. They measured it, and determined that it was a tolerable amount, because it was less then 10% penetration. Problem was the root cause of the burns was not adequately determined. When you don't know the o-rings are twisting due to internal stress and that those stresses have a dependence on temperature, you miss that catastrophic O-ring failure is a risk at near freezing temperatures.

The second space shuttle crash was the same problem again. Falling chunks of ice and foam. Since nothing major had happened previously, it was assumed everything would be fine.

These failure and problems were the warning signs that something was wrong. There was no design spec for acceptable o-ring burn through or a leading edge heat shield damage threshold spec.

I bet we will find that Boeing and the FAA talked themselves into believing that

> The Vertical Pitch Indicator is reading on some MAX 8's 10-20 degrees higher than what the aircraft actually is pitched at.

Seems like not everything was built to spec

I'm really getting tired of hearing "Boeing cut corners" every time this comes up. Boeing is more a customer-driven company than most other industries. They would want to design a new plane, they in-fact were proposing a redesign of the 737 using composites and new aerodynamics, but their customers, who drive everything they do, wanted more of the same immensely popular 737, but more efficient. So they had to take the 737 and make the changes to it to get the desired efficiency. The airlines also wanted minimized difference training for the MAX for pilots already flying the NG, so Boeing worked to minimize large changes. Somewhere in this pile of things they decided, and the FAA approved, MCAS being omitted from the training. That has since been largely rectified by the airlines.

I also have noticed that everyone seems to want to pile on Boeing, as if all these other factors don't exist and Boeing is the one employing, training and supervising the pilots and maintenance people.

While this is supposed to be a site for hackers, I would somewhat expect that an engineering problem, which this largely is, would be dealt with like an engineering problem. But the commentary on this issue seems to largely be driven by emotional responses. Noone is considering the potential and now actual impact of grounding aircraft. Grounding aircraft may actually cost more lives than letting them fly, because of the higher mortality rate for other modes of transportation.

In the case of the Ethiopian crash we know very little about what actually happened, what we do know not only looks similar to the Lion Air crash, but also looks similar to the last Ethiopian crash, which did not involve a 737 MAX aircraft at all.

Many people also criticize Boeing's reaction without realizing that there is a lot of money riding on exactly what they say and do. No they will not ground a while aircraft type just because two crashed, especially when there is no official data to go on for the second crash. They will be very reluctant to admit fault, since every time they admit something is wrong customers demand money and people sue.

The way it happened with the FAA making the call its less on Boeing to compensate airlines for the inactivity of their aircraft. Which is not what Boeing wants, but is less shitty.

I couldn't agree more. I don't know if any of the commenters have built systems where lives or livelihoods are on the line; involving decisions where billions of dollars and many lives are at stake; nobody sets out to build unsafe systems or cut corners. The development process for aircraft is built with redundencies and multiple levels of validation; everything is tested and certified by internal and external experts. Software engineers fly on these planes monitoring their own software while in flight. I wonder how many people have trusted their own life with their code?

My bosses DID set out to build unsafe systems, to save money, and ordered me not to fix unsafe software for that reason; even after explicit warnings. Which crashed that nonprofit. It's common. There are many motivations including agent-principle problems, narcissism, and more. "Nobody sets out to make an excessively safe product that won't turn a profit" would be more accurate.

> by putting more powerful engines on it

I think the focus was on the engines being more efficient, not more powerful. See for example the wiki page [1]

"In summer of 2011, the objective was to match the A320neo 15% fuel burn advantage, but the initial reduction was 10–12%; it was later enhanced to 14.5%: the fan was widened from 61 inches to 69.4 inches by raising the nose gear and placing the engine higher and forward, the split winglet added 1–1.5%, a relofted tail cone 1% more and electronically controlling the bleed air system improves efficiency."


Completely agree with your analysis, and not a pilot or aeronautics-related engineer either.

So if we're both right, responsibilities should be cleared. IMO the person/team that didn't prepare for sensor redundancy should be even criminally liable, and the authorities that were calling for not grounding the plane yet, should be dismissed, for complete disrespect to cautionary safety.

>NTSB final accident reports, with findings as to probable cause and recommendations intended to prevent future accidents, are statutorily prohibited from coming into evidence at trial; however, NTSB factual reports are generally admissible at trial. NTSB investigations typically last for a year or longer.

Just a quick Google there. Do note that the intent behind most NTSB reports being inadmissible in a trial is to secure cooperation from all parties in order to get an accurate representation of the problem that caused the disaster in the first place.

That's not to say lawsuits or criminal charges won't happen, but those proceedings will not be able to use NTSB investigatory results to make their case. They will have to independently prove their case through other means.

NTSB is not in the blame and punish game, and any civil case against Boeing will have a he'll of a time. Never mind that a criminal case will still be required to be levied against an individual at the company, and would have to prove criminal intent.

There will not likely be a terribly satisfying legal remedy to this. Taking VW and the financial industry as a starting point for successful pushback against industry, I'm not expecting a terribly satisfying legal conclusion on this one. There will probably be a review of corporate culture, maybe some extra scrutiny for Boeing for a while, some token legislation, and maybe a charity fund for victims.

Least that's what my magic 8 ball expects.

The idea that we should parade the heads of the people responsible for this around on a pike is a very emotional reaction. Of course, these will be the people who will be less likely to make any mistakes with this sort of system again, since hundreds of millions have been burned and lives lost. They will be the ones who are skeptical of datasheets, who will request more thorough testing, will spend extra time making sure they don't make the same mistake again.

Of course, we could just ground every plane when another plane of that type crashes, and be "100% certain" of no repeats.

> The idea that we should parade the heads of the people responsible for this around on a pike is a very emotional reaction.

No, it is what you normally find in laws. Doctors who perform poorly at their job are liable. Engineers who disregard basic safety measures established for decades (such as sensor redundancy) should be too (or the managers that approved the cost cut to not include redundancy).

It is common to "recycle" an existing airframe by upgrading it in various ways. In fact the customer pressure on Boeing to create the 737 Max (referenced in a sibling comment) was in part driven by the success of Airbus doing the same thing to create the A320neo. The 737 Max is the 9th or 10th variant that Boeing has created of the 737. The FAA has a process for certifying these variants.

The concept of a variant of an existing plane model should not be the focus of concern IMO.

It's a new plane. They kept the shape of the very popular 737, to the extent they could, purely for marketing purposes, even though this meant creating the first dynamically unstable ("wants to crash") civilian airliner in history, to keep that shape. (Modern fighter jets are supposed to be dynamically unstable, post Genghis John, for better maneuverability.)

Did anybody notice this 737 Max-7 test on Tuesday? Quite an interesting altitude and speed pattern --


When I replayed the flight on this map I noticed that the airplane lost a little bit altitude when it made a turn. And some turns are very narrow. It is normal for an aircraft to lose a altitude when turning.

Could you explain to a layman like me the interesting parts?

Check the 'detailed track log' - you have to click to show it on mobile - for the altitude & speed graphs.

A normal flight (https://flightaware.com/live/flight/QFA804) goes up smoothly. This one very much does not, and the patterns shown bear a resemblance to a layman like me to the published flight patterns of the Lion Air flight.

Note the very steep climb on this flight also. Who knows what that means.

People who know way more than I do, please step in!

The apparent difference in steepness arises from different scaling in both altitude ("typical flight" reaches almost 12 km, while "test flight" reaches almost 6 km) and time (the 5 minute intervals are twice as large in the "typical flight" graph).

The actual time to reach 6 km was around 6.5 minutes in the "typical flight" and around 5 minutes in the "test flight", as can be seen by scaling both graphs to the same proportions. (I've done it quickly with Paint here: https://i.imgur.com/IZyYoCl.png)

Is that 1.5 minute difference really that high? Honest question, I really don't know anything about this topic.

EDIT: Sorry, this should have been a reply to @rkagerer's comment.

If a test aircraft, it is probably very light, Boeing Field is probably under noise abatement procedures and they might just be optimizing for minimal flight time to get to test altitude. Would have to compare to normal test/aircraft checkout flights and not airline tracks.

But no doubt they would be flight testing the reputed MCAS changes at this time so it’s a reasonable guess.

I would assume they were testing flight overide logic... which takes some balls...

Yeah, I can hardly imagine being the crazy/gutsy person actually doing a test of failure modes on a plane in the air. I mean, what exactly are your options if it goes wrong? Maybe you can parachute, but seems like a long shot given altitude and the plane itself (just aim it at an uninhabited area before you jump?)

I'm just... gonna go back to my office job, I think.

The type of error that this appears to be is solved if you know what it is, you can just turn the MCAS system off and hand-trim the plane no problem. So testing the failure at a high altitude is of a relatively low risk, you've specifically briefed and likely practiced in a simulator exactly how to turn these systems off. The pilot not flying the plane almost certainly has their hands right on the switches and or fuses.

Note, though, that as soon as you do that, you're flying an airplane with flight characteristics that are unlike any other aircraft you've ever flown (unlike any ever built before); an unstable aircraft with extreme stall characteristics you've never encountered - not a normal airliner. And you're close to the ground already.

Serious question:

WHY has not Boeing turned every single aircraft of theirs into a drone-capable device?

They should have a person, like we have with air traffic control, sitting monitoring all planes in flight, but that person is sitting in a pit on the ground.

They have to monitor X number of craft - and in the event of any anomolous data.... be ready to take over control of the machine in the air.

Imagine if all pilots had a life sensor on them, and should they fail - control of the craft auto switches to ground control.

We can already kill millions of people with dron pilots in nevada. Why not make drone pilots fly airliners.

Heck, in a few years - you just need a few technicians in the cock pit to be sure there are not physical issues - and let the ground pilots do all the work - but the air pilots can take over if the ground connection is lost...

Here are two obvious reasons:

1. Lack of low-latency communication. The Ethiopian Airlines crash flight occurred when flightradar24 couldn’t see the plane. Two-way communication is even harder than just receiving transponder signals.

2. How, exactly, is a random Boeing employee going to help? They’re thrown into a situation in which sensors are faulty and they can’t see out the windshield.

What makes you believe that a ground pilot will be more competent than the existing air pilots?

It is very rare for crashes to occur because the pilots both become incapacitated. I do remember a crash that occurred because the pilots allowed the cabin to become depressurized and failed to diagnose the problem. Eventually everyone, including the pilots, fell unconscious and the plane circled on autopilot until it ran out of fuel.

Arguably this should be possible for cars as well, as a transitional stage towards full self-driving capability.

With planes you could argue that the additional complexity of implementing something like this would be a safety liability in itself. It's not 100% clear what problem it would solve. On the other hand, plenty of people would pay to not have to do the driving themselves.

I freaking love this idea with cars.

Pay disabled people to GTA 5 you home....

(aside: japanese cafes are employing disabled-driven robots....)

Dont be so hard on yourself, office jobs crash and burn at a much higher rate.

I suppose if the hazard pay is enough to set my family up for life...

There is literally no amount of money that compensates for an absent parent. YMMV, of course, and the calculus is different if you're single or something.

That's like saying money can't buy happiness. Technically it's true but money can buy you things that make the difference imperceptible.

Perhaps the steep climb has to do with how the incident airports were setup?

“Data released by the Sweden-based service suggested the aircraft had climbed almost 1,000 feet after taking off from Addis Ababa, a hot and high-altitude airport whose thinner air requires extra effort from an aircraft's engines.” https://af.reuters.com/article/worldNews/idAFKBN1QR07Z

Extra effort from the engines? That sounds a little silly, like they just need to be motivated to work a little harder. Take-off power is take-off power.

Reduced air density just means the engines don't produce as much power and the wings don't generate as much lift; a double-whammy of lower available engine thrust and greater speed required for the same takeoff roll.

> Take-off power is take-off power.

I don't think it is anymore - not with noise abatement rules. You can take off in X meters with 100% throttle, or 2X with 80% - but the runway is 4X meters, so you takeoff with reduced throttle to make things quieter for the neighbors.

I don't know whether there are really regulations for that but as a pilot (which I am not) I would not want this as this would lower the decision velocity (I think it's v2) after which the start can't be aborted.

It has some interesting dips that seem to be testing some form of sudden dip in altitude which could be seen as related to the system under scrutiny, a safety system designed to prevent stalls which could in some cases "send the plane into a fatal descent if the altitude and angle information being fed into the computer system is incorrect"

(not that any of this confirms that the MAX has issues or this is what they were actually testing)

Illustrating the answers from snek and jen729w to help clarify: https://i.imgur.com/p8uEwMh.png

Out of all the previous Canadian Minister's of Transport we've had over the years, Garneau's opinion is one I'd actually pay attention to. He flew on the space shuttle three times, was in space for 30 hours, and has a doctorate in Electrical Engineering. Not that he's an expert in this, but if there's any Canadian politician in the past few years who would be able to question their public servants and make sure the data supported the decision, it's likely him.

Canada's two largest carriers, Air Canada and Westjet, have a combined 37 Max 8s in their fleets, so I'm sure this wasn't a position Garneau took likely.


This guy handed me my degree when I graduated, and I share your respect for his accomplishments. But I do expect he took some PR flak, considering just two days ago he spoke fairly strongly against a ban (stating he would board a 737 MAX "without hesitation") even as other countries including the EU implemented one, and despite the Air Canada Pilots Association calling on him to "to take proactive action to ensure the safety of the Canadian travelling public".

Not disputing your point, or even criticising his actions (I don't know enough about the situation to do so, aside from having a stake by being booked on an affected flight). Just filling in a bit more of the story.

With all due respect to Garneau, I don't think the statement that he wouldn't hesitate to get on a max 8 doesn't really mean much considering he willingly flew 3 times on the most deadliest aircraft of the modern era.

I would respect anyone who changes their opinion after seeing new facts.

I don’t understand this ridiculous burden we place on our politicians. Why can’t they change their opinion on things when new facts come to light? I am not saying changing principles but opinions.

Like someone asked “I see new facts and change my opinion. What do you do?”

Well, calling the shuttle an aircraft is a bit of a stretch. It was more like a slightly aerodynamic rock.

I appreciate the insight, and I didn't hear those comments from two days ago. That is, indeed, unfortunate. I would have expected him to at least hedge himself a bit. I haven't been involved in the aviation industry in over 20 years now, but I've always been a huge advocate for how safe flying is. Many times I've said similar things to family and friends when discussing some one-off plane crash. And I still fly very frequently.

But two brand new planes falling out of the sky for no obvious apparent reason? I'd have expected him to be a bit more cautious in his response.

I think he did hedge yesterday. Something like not grounding now, but "all options are on the table".

There's nothing new here - it's rehashing the same "these are similar so may be the same cause" as every other article on the topic.

Not sure why you're getting downvoted. You are exactly right, this article contains zero new information. It's just another opportunity for HN "experts" to pontificate about airplanes using mostly incorrect armchair analysis.

> It's just another opportunity for HN "experts" to pontificate about [X] using mostly incorrect armchair analysis.

Isn't this just classic Gell-Mann amnesia? All HN threads are like this. (Probably all threads everywhere on the internet are like this.)

The next time you read a thread on [to skim the current front page] Facebook data, consensus protocols, Google hardware, facial recognition, death metal, or the British coastline...before reading the comments remind yourself

"this is just another opportunity for HN 'experts' to pontificate about [X] using mostly incorrect armchair analysis"

Yes, absolutely! I actually had the same thought yesterday and I walked away from HN for a break, grumbling to myself about how true the old joke is about computer programmers having so much hubris that they think they are smart enough to solve every problem in the world ;-). See also Soylent.

Yesterday it was just an overload because we must have had a half dozen 737MAX threads in the previous 24 hours. And the quality of the commentary was going down, not up, as time went on.

Totally agree with your advice to keep in mind that this applies to all topics. A little part of me keeps whispering that I should avoid looking at the comments altogether. I rationalize that on HN I am getting good quality commentary and the first few contrary arguments will help me see the article/news/whatever in an objective light, but then I totally get sucked in.

> Isn't this just classic Gell-Mann amnesia?

My favorite-thing about the Gell-Mann amnesia effect is that the coiner of the term himself, in the very speech where he coined the term, fell victim to that effect.

To Michael Crichton's great credit, in the 2002 talk where he defines the term, he gives us an extremely specific, testable hypothesis we can use to evaluate his idea.

He says that the "effect" is as follows: it is not worth your time to read the newspaper and all newspaper stories which contain any kind of "speculation" or "predicting" have no predictive value.

The example he chooses is the 2002 United States steel tariffs. Most mainstream newspapers he reads have quoted experts whose consensus is that the then-contemporary 2002 tariffs will negatively affect GDP. His hypothesis is as follows: we will find in retrospect that these predictions were useless and US GDP will not clearly be damaged by the 2002 tariffs.

Was he right? Well... Years later, there is consensus among experts that US GDP was in fact negatively affected by 2002 steel tariffs (links to sources at https://en.wikipedia.org/wiki/2002_United_States_steel_tarif... ).

Driven to attack the idea of trusting any experts who predict the future, Crichton's key example was readily disproven.

I have often re-read Crichton's talk and have found myself wanting to project extra layers of meaning on top of his ideas; I find myself arguing on his behalf to try to change what he is saying to make it "more true". The words he uses are carefully chosen and seem like they add up to an interesting idea.

It is true that people seem to believe information without factual basis when it's presented in the "news" format; it's also true that people seem to believe opinions they already agree with when presented to them in any format; and it is certainly true that people asked to make predictions with incomplete information sometimes guess wrong, yet their "best guess" predictions are sometimes presented without robust explanations of the uncertainty and what it means.

All that is true. But it doesn't change the fact that the "Gell-Mann Amnesia Effect" was introduced by saying that the untrustworthy lamestream media needs to stop yammering already about those steel tariffs hurting GDP, which there is no way they could predict. (It appears the media was right on that one.) Worse than relying on anecdata, though, his talk does not propose a robust way of actually measuring whether media predictions are accurate -- he doesn't even try to measure that, let alone present results.

(I have giggled at this before - https://news.ycombinator.com/item?id=18005236 )

"He says that the "effect" is as follows: it is not worth your time to read the newspaper and all newspaper stories which contain any kind of "speculation" or "predicting" have no predictive value."

To me, "no predictive value" is consistent with sometimes being right, possibly by chance. It means being wrong enough that you can't assume either positive or negative correlation with the truth. It's easy to hide statements which are true and should be obvious in a sea of confusing nonsense.

I haven't read all the breaking news coverage, but my perception was that previous articles only said that there were similar characteristics between the crashes, such as happening shortly after takeoff. But prior articles lacked more specific detail, whereas this one shows what the publicly available data (as of today, AFAIK, which apparently caused Canada to ground the MAX planes) actually is -- e.g. the recorded vertical speed of Egyptian Air flight.

Obviously the black box data will be more granular, but this "preview" is pretty useful. I remember some initial articles referring to witnesses who claimed seeing [0] "smoke and debris", which would imply a different situation than the Lion Air flight.

[0] https://globalnews.ca/news/5042552/ethiopian-airlines-crash-...

AFAICS, the "new satellite data" is just the usual ADS-B data that led everyone else to the same conclusion, but picked up by a global network of satellites[1]. (Sites like FlightAware24 are instead based on data contributed by a global network of volunteers listening to ADS-B broadcasts in their local area.)

1. https://aireon.com/resources/overview-materials/its-just-ads...

"The oscillation of roughly 15 to 20 seconds is a telltale sign that suggests the M.C.A.S. system may have been involved, said R. John Hansman Jr., a professor of aeronautics and astronautics at the Massachusetts Institute of Technology.

"Planes like the Boeing 737 oscillate naturally, he said, because of turbulence and and other effects. But those swings have different time spans: either between five and eight seconds, or a minute or longer. The variations in the intermediate range of 15 or so seconds have no other obvious explanation, he said."

Both planes porpoised with periods of~20 seconds during their flight.

Assuming this is a phugoid oscillation?

Here’s the FAA Emergency Order of Prohibition:


On March 13, 2019, the investigation of the ET302 crash developed new information from the wreckage concerning the aircraft's configuration just after takeoff that, taken together with newly refined data from satellite-based tracking of the aircraft's flight path, indicates some similarities between the ET302 and JT610 accidents that warrant further investigation of the possibility of a shared cause for the two incidents that needs to be better understood and addressed. Accordingly, the Acting Administrator is ordering all Boeing 737 MAX airplanes to be grounded pending further investigation.

I’m curious about “the wreckage concerning the aircraft's configuration just after takeoff“ part as that’s the first I’ve heard about aircraft configuration having been recovered from the crash site. Maybe something about the horizontal stabilizer positon when the crash occurred?

Afaik the horizontal stabilizer is deflected by a worm gear. Even with a totally destroyed fuselage I think it would be possible to recover parts of the gear which allow judging its position at time of impact. As well it is in the aft part of the plane where also the flight recorders are placed as they have there the best chance of survival.

Besides the engines moved forward those are 2x higher bypass engines and thus point of thrust is moved even more forward resulting in even more of pitching torque - basically the plane seems to be like a Hayabusa bike lifting the front wheel up at the good throttle.

Similar reasons would be better news. Different reasons could mean they are hunting two fatal bugs.

In this list of 737 accidents, correct me if I'm wrong, seems that the crashes for similar reasons (shortly after takeoff) are 3 , not 2. There is other one (May 18, 2018 ) involving a 737-100/200 that seems strictly related to the other two by time and dynamic: https://en.wikipedia.org/wiki/List_of_accidents_and_incident...

In that case it was a really old aircraft that stalled in the way that MCAS is designed to prevent.

One thing to note is that this flight is very similar to the previous Ethiopian crash.

How did Boeing QA sign off this anomaly https://en.wikipedia.org/wiki/2019_Boeing_737_MAX_groundings...

Tried to read on Android Chrome, cannot click away cookie policy. Am I the only one? Miss safari reader mode...

Android Firefox has a solid reader mode

And ublock/umatrix, which are also helpful for making things usable.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact