737 MAX crashes “linked” by satellite track data, FAA says
arstechnica.com
arstechnica.com
Most immediately saw parallels with the Lion Air crash - but as always, we as laymen should refrain from coming to conclusions until after a thorough investigation has been completed.
The fleet has been grounded, and reports / data will slowly trickle out as more conclusions are made.
People say this, but why? It seems to go against human nature. People love to speculate based on limited information and/or understanding.
Concluding things off it is not.
We should resist that part of our nature. It is not rational.
What evidence is there that only acting rationally is "better"?
Speculation is fine, it's when someone is on Fox News as an Aviation "Expert" claiming that the plane was struck by an angry Zeus's lightning bolt, which many segments of the population will believe, that's when we have problems.
Being skeptical and keeping onself safe is fine, but things can go wrong if people are trying to falsely act like experts.
Every flight broadcasts this data kinda like GPS and all you need is an antenna to receive it. In fact the tracking is basically done via ground-based GPS(ADS-B) by measuring the time delays between different receivers in the vicinity. The problem is that there aren't many hobbyists in those parts of Africa so coverage is sparse. In America, its pretty spot on in most locations. A USB adapter can surprisingly pick up planes as far away as 400 miles.
https://www.rtl-sdr.com/buy-rtl-sdr-dvb-t-dongles/
E.g. you can get an SDR with built-in ADS-B specific low-noise amplifier and SAW filter also for $25. (This can't do general purpose stuff though, unlike the one linked by parent.) Or you can get their general-purpose SDR and their broadband LNA for $40 total.
No, position (likely received from plane GPS) is just openly transmitted in ADS-B frame.
At least, in areas where such sites have good coverage they may actually get closer to the ground data than the satellites are capable of picking up.
I don't mean to dump on doing a quality analysis, just that the quality of data is not as drastically different as you're suggesting.
Generally data is received every 5 seconds. In cases such as Lion Air, they downloaded all the granular data from nearby receivers and provided some analysis in their blog.
They also made available the raw data for the public to download:
https://www.flightradar24.com/blog/flightradar24-data-regard...
It sure is an impatient world we live in. Probably attributable to the Internet, our insatiable need for more information NOW. Everyone ripping on the FAA for not making a grounding decision until yesterday, even though the crash itself happened just four days ago.
Overall, I'm pretty happy with what I saw out of the FAA. It just seems like the FAA may have gotten the data last. Canada got it before we did. The EU got the blackbox first, so once they looked at the data, they banned the plane. China's response may have been preemptive, since there's no way they could have looked through the data at that speed.
Perhaps we should be pissed off at getting the data-last. But that's hardly the FAA's fault (maybe a State-department thing). In the future, Boeing needs to at least appear more neutral and less like a lobbyist when these events occur. There's significant distrust in our system these days, our regulators need to understand what it looks like when they're talking with the companies who make the plane before getting the black box...
Nope, the FAA had the data before the Canadians.
Literally every one would have been grounded had the FAA grounded the type after 2 accidents.
The first crash of the 777 occurred 15 years after it was launched with 1500 aircraft delivered, and even in that crash foul play is suspected.
The first crash (ignoring hijacking) of the 767 occurred almost a decade after it was launched, and the second was almost two decades after launch, and it's still not clear if it was suicide.
2008 British Airways Flight 38: 47 injured https://en.wikipedia.org/wiki/British_Airways_Flight_38
2013 Asiana Airlines Flight 214: 3 fatalities, 187 injuries https://en.wikipedia.org/wiki/Asiana_Airlines_Flight_214
2014 Malaysia Airlines Flight 370: disappeared https://en.wikipedia.org/wiki/Malaysia_Airlines_Flight_370
2014 Malaysia Airlines Flight 17: hit by anti aircraft missile https://en.wikipedia.org/wiki/Malaysia_Airlines_Flight_370
In fact it seems quite prudent for a regulator to do this.
It can be, though.
There is a risk inherent to air travel. The safest thing to do would be to ground all flights of all aircraft permanently.
That wouldn't be prudent, though, because we can mitigate the risk until the benefit of air travel outweighs the risk of dying during air travel.
The prudent thing to do is to carefully evaluate the risks, and make a determination when the risks become too great.
If a company releases a new version of their software and it crashes twice in similar ways (much more frequently than the previous version), that is also enough reason to rollback the release and investigate, especially if the software has any critical use cases.
Obviously commercial aviation has much higher standards than software vendors generally do. Airline tickets don’t generally come with 5,000 word EULAs with all-caps provisions like “THIS SOFTWARE IS NOT FIT FOR PURPOSE AND SHOULD NOT BE USED FOR ANY REASON AND YOU ASSUME ALL RESPONSIBILITY FOR ANY NEGATIVE OUTCOMES OF ANY KIND” etc.
Two crashes in close succession seems like an obvious abnormality. But because standards are so high, to commercial aviation providers and regulators, every single crash is treated as a significant abnormality. So the question is, why don’t they ground all models of a certain plane after a single crash? Once you understand that, the same logic applies to two crashes.
Clearly, a chance for a bug losing you 300$ is much less of a problem than a chance for a bug killing 300 people.
The second thing is that a software update may render previous knowledge useless.
Just because the old software ran successfully on 100k nodes, it might still be the case that 2 out of 1000 new versions fail, so then you have 0.2% failure rate. That doesn't mean you should take out all 100k nodes but it means you might want to stop the 1k nodes in accordance with the principles laid out above.
Finally, I am not even sure why this is a yes/no question. How about: You're allowed to use the plane but you cannot rely on a defense of "the FAA said it was ok" if something happens, so talk to your insurance before.
So this is both statistically suspicious, and points at a possible common cause for the crash, making it even more worrying that there's something seriously broken in this design.
Imagine the story would go like this: a girl heard the story about the boy who cried wolf. Uppon seeing a pack of wolves the girl doesn’t cry wolf because she wasn’t 100% sure that it wasn’t in fact a pack of stray dogs. Villagers die.
The original story teaches kids not to lie to gain attention, because nobody will believe you when you for once have a serious risk at hand. It has nothing to do with how we should treat real risks, where we have real indicators of increased risk.
Climate change is the ultimate example of such a risk. Better safe than sorry, because beeing sorry can easily become the end of the species here. Some airplane failing is arguably a tiny thing in comparison — but once you have been warned about a risk you can mitigate and you do nothing, you own it. Living is dangerous and everything has a remaining risk that stays dangerous even after any reasonable risk has been mitigated.
But this isn’t about totally ruling out all remaining risks, it is about ruling out risks that are easy to rule out.
If somebody told you the brakes of your bicycle are both broken, you would probably not lend it to someone until you convinced yourself of the condition of your brakes. You will do this despite the fact that cyling can be dangerous with a fully functional bicycle too.
[1]https://www.cnbc.com/2019/03/14/iridium-aireon-satellites-ga...
Who are pre-Internet people going to discuss it with? A handful of people at the dinner table or at work?
It's not about impatience, it's about opportunity.
https://news.ycombinator.com/item?id=19385980
"Why Investigators Fear the Two Boeing 737s Crashed for Similar Reasons (nytimes.com)" (18 hours ago)
The data already available is quite precise. The Ars article is showing less precise data than the NY Times.
It is identical data from ADS-B that we've had for days. People keep reiterating their "verified" line, but that had no basis when they said it, repeating it doesn't add to its legitimacy. The data today is the same data we had then.
It is essentially a way to ignore the data until it is politically expedient.
This is incorrect. Simply Googling '737 max executive order' turns up multiple hits. This doesn't mean there was an actual order, however.
[1] https://www.google.com/search?q=+executive+order+737+max
But of course, we should have a fully-functioning and fully-staffed FAA, too. But we don't, because of Trump intentionally hamstringing it, and because of antics like trying to get his personal pilot named head of FAA when he came into office.
It appears that Mr. Dunkin has decades of experience as a pilot and has no previous affiliation with the FAA.
I've got decades of non-management software dev experience, perhaps I should be in charge of Microsoft?
However, to be clear, he does have smaller scale management experience - most likely consistently, but at the very least during the Trump campaign.
Perhaps he would be wrong for the job, but certainly there are examples of both experienced and less experienced/inexperienced persons doing awful jobs at the helm of agencies, companies, and even countries.
I would also add that being in charge of such a massive organization often means you do not get to make off the cuff or unilateral decisions so easily as you would, say, at a 50 person organization. You do certainly get to steer, but there are many other hands on the wheel.
As for you being CEO of MS, Satya is doing a decent enough job + I know nothing of your background and you come off as under confident in your ability to do the job. So I'll go with no :p
Claiming he has no management experience is simply false.
Even someone who is not in that kind of position, but a major airline pilot, has significant management experience simply by virtue of the position. They're not like fighter pilots alone and unafraid up there flying around for fun.
Don't think I agree with this. The president shouldn't have to be involved. But if the FAA is acting in a bone headed manner, the president, as the head of the executive, is the check & balance that can get involved and fix the problem. Likewise the president is the one ultimately responsible to the people for the FAA's actions.
We don't know what extent Trump influenced the decision, but he was certainly involved, he made the announcement after all [0].
[0] Announcement https://www.youtube.com/watch?v=fs3NlxY5-iU
To the extent that you think the FAA is actually serving as an extension of the will of the president you're right that I misused it, but I don't think that describes reality.
It's a bit of a flaw in human nature that easily-pictured problems scare us much more than abstract/distant problems.
Is it silly to press for a moral value of striving to make our morality align with numeric measure of significance rather than sensational emotions?
The two fatal crashes were caused by an interplay of hardware, software design, and human factors, and studying things like this is fascinating to those of us who are interested in building systems so that catastrophes don’t happen.
MCAS was created as a way to compensate for instability caused by strapping large engines onto an aircraft not originally designed for it, as opposed to doing a redesign.
Iterating on a proven airframe has a lot of advantages for aircraft design and it is common in the industry.
It's also entertainment in the same sense as catastrophe movies.
Edit:
The basis for the “trolley problem” is precisely how you handle accounting for the numeracy of death versus the emotional involvement and personal culpability.
People who rationally perform cost benefit analysis about lives and well-being seem to have more psychopaths: see CEOs and other leaders. This isn’t necessarily bad, but it’s worth noting.
This number sounds highly questionable to say the least: https://www.cdc.gov/flu/about/burden/how-cdc-estimates.htm
With no way to compare it to a gold standard how could anyone know how accurate it is?
I havn't followed very closely, so take that with salt, but it's a starting place if anything.
We could stream this information via satellite uplink. Instead Radar and transponders is used to keep a close enough location to make recovery generally strait forward.
It’s really turning off the transponder while over the middle of the ocean that’s the issue for MH370.
Locating a crash within 5 miles makes recovery of the FDR strait forward. A location update every 30 seconds easily provides that and looks like a lot of data. But, to really reconstruct a crash you need a lot more data including control surfaces etc, which is where the FDR comes in.
https://www.cnbc.com/2019/03/14/iridium-aireon-satellites-ga...
https://www.bbc.com/news/world-europe-34793274
https://www.cbsnews.com/news/mh370-how-can-we-stop-losing-ai...
https://www.forbes.com/sites/rmark/2019/03/04/five-years-mal...
I think the black box was handled by British investigators:
https://www.telegraph.co.uk/news/worldnews/europe/ukraine/10...
[1] https://gulfnews.com/world/ethiopian-airlines-crash-germany-...
http://fortune.com/2019/03/13/ethiopian-air-crash-black-boxe... > but Ethiopian authorities would prefer to work with the U.K.’s Air Accidents Investigation Branch to ensure that U.S. experts won’t have undue influence in the probe of the American-made plane.
https://www.bloomberg.com/news/articles/2019-03-13/ethiopia-... > The choice of Europe over the U.S. National Transportation Safety Board for the analysis of the black boxes is a strategic decision for Ethiopian Airlines and the nation’s government, said Asrat Begashaw, the carrier’s public relations director.
With that said, AIAB has just as formidable a reputation as the FAA. If the truth is there, they'll find it.
Also, the F.A.A. has already been aware that there were issues with this plane (software fixes were being worked on), but the planes continued to fly. If it turns out that the information already known to them BEFORE the latest crash should have been sufficient to call for a grounding, I could see them attempting to cover their ass.
We are talking about governments here so the conspiracy theory that the person who opened the box substituted a different set of data is actually possible. Conspiracy theory territory for sure, so you really want to make sure that a fully trusted third party is involved. France seems as good as anyone to fit that bill.
If you were worried about conspiracies then multiple aviation authorities, or government departments, could send observers to watch the opening and imaging.
Secretary of Transportation Elaine Chao was Secretary of Labor under George W. Bush, Distinguished Fellow at The Heritage Foundation, and the wife of Senate majority leader Mitch McConnell.
Also, the altitude data seems to cut out at it's peak. Was the decent so rapid it couldn't be tracked?
The problem, he said, is that witnesses instinctively try to match events with their past experiences: ''How many plane crashes have you witnessed in real life? Probably none. But in the movies? A lot. In the movies, there's always smoke and there's always fire.''
https://www.nytimes.com/2002/06/23/weekinreview/ideas-trends...
See also “Reliability of Eyewitness Reports to a Major Aviation Accident”:
https://commons.erau.edu/cgi/viewcontent.cgi?article=1040&co...
They don't need to align. Eyewitness accounts are so bad, it's surprising they're allowed in court.
"Why Science Tells Us Not to Rely on Eyewitness Accounts - Eyewitness testimony is fickle and, all too often, shockingly inaccurate"
> Many researchers have created false memories in normal individuals; what is more, many of these subjects are certain that the memories are real.
> In one well-known study, Loftus and her colleague Jacqueline Pickrell gave subjects written accounts of four events, three of which they had actually experienced. The fourth story was fiction; it centered on the subject being lost in a mall or another public place when he or she was between four and six years old. A relative provided realistic details for the false story, such as a description of the mall at which the subject’s parents shopped.
> After reading each story, subjects were asked to write down what else they remembered about the incident or to indicate that they did not remember it at all. Remarkably about one third of the subjects reported partially or fully remembering the false event. In two follow-up interviews, 25 percent still claimed that they remembered the untrue story, a figure consistent with the findings of similar studies.
https://www.scientificamerican.com/article/do-the-eyes-have-...
[1] https://www.bostonglobe.com/metro/2015/05/23/eyewitness-acco...
[2] https://www.scientificamerican.com/article/do-the-eyes-have-...
> Although there were considerable discrepancies between different accounts, most witnesses to the accident had seen a "streak of light" that was unanimously described as ascending, moving to a point where a large fireball appeared, with several witnesses reporting that the fireball split in two as it descended toward the water.
> 258 were characterized as "streak of light" witnesses ("an object moving in the sky... variously described [as] a point of light, fireworks, a flare, a shooting star, or something similar.") The NTSB Witness Group concluded that the streak of light reported by witnesses might have been the actual airplane during some stage of its flight before the fireball developed
> In addition, 18 witnesses reported seeing a streak of light that originated at the surface, or the horizon, which did not "appear to be consistent with the airplane's calculated flightpath and other known aspects of the accident sequence."
> [T]he NTSB noted that based on their experience in previous investigations "witness reports are often inconsistent with the known facts or with other witnesses' reports of the same events."
That extra pilot, who was seated in the cockpit jumpseat, correctly diagnosed the problem and told the crew how to disable a malfunctioning flight-control system and save the plane, two people familiar with Indonesia’s investigation told Bloomberg.
Why wasn't this information passed on to all Lion Air pilots?
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.
This article is further missing the "new information from the wreckage concerning the aircraft's configuration" piece, which is in the NPR interview[2]: And that, coupled with some physical evidence we found at the crash site led us to believe that the similarities were too great not to consider that there was a common thread.
So it wasn't just the flight path, which indeed was sufficient for other countries, but the addition of physical evidence as well:
And when you have a common thread between two accidents, then the argument for grounding becomes necessary. Grounding becomes necessary, and so that's what we did. We didn't have that link until yesterday morning, yesterday afternoon about midday.
GREENE: But isn't this something that analysts and experts have been saying for days now, that these two crashes appeared similar?
ELWELL: Yeah. Many were saying it, but nobody had data to act on it. It was all conjecture. And in aviation, the FAA in the U.S. has always acted on data. We're a data-driven organization. We have the safety record we have today based on science, risk analysis and data.
1. https://www.faa.gov/news/updates/media/Emergency_Order.pdf
2. https://www.npr.org/2019/03/14/703298739/faa-acting-head-dan...
If it was a pair of Airbus aircraft (A320, A330, pick any you like) which suffered these two crashes, would the FAA have demanded the whole fleet grounded?
Even more interesting is the recent 767 Amazon cargo flight crash in Texas also did a node dive at around 8000 feet which somehow has escaped the news with this new crash [1]. The pilots decided on a path around weather, were at 11-12000 feet, descended 3000+ feet and then the plane essentially did a similar nose dive. Is this a data problem?
> As the plane passed through 12,000 feet at a ground speed of 290 knots (340 mph), the pilots indicated they preferred the westerly route option ATC had given them around the rain; air traffic control told them they would need to descend quickly to 3,000 feet to do so, and radar data reveals the Boeing turned to a heading of 270º as requested and descended through 8,500 feet. One minute later, the controller told the crew they would be past the bad weather in about 18 miles, and to expect a turn to the north. The crew responded "Sounds good" and "Okay," according to the NTSB, and the plane leveled out at 6,200 feet before rising 100 feet more.
> That, apparently, is when things went haywire. The aircraft began what the NTSB report described as "small vertical accelerations consistent with the airplane entering turbulence," according to the flight data recorders recovered from the accident scene. Seconds later, with the plane holding steading at 230 knots (265 mph), the engines went to full power, and the nose of the plane rose four degrees...then the aircraft pitched nose-down for the next 18 seconds, reaching a maximum pitch of –49º in response to the plane's elevator inputs.
The Ethiopia 737 Max has almost the same issue [2]
> ADS-B data recorded for ET302 by FlightRadar24 shows that the aircraft, after reaching an altitude of 8,025 feet above sea level, suddenly dipped, plunging 400 feet before recovering briefly. But the aircraft's vertical speed remained unstable, and a few minutes later it dove into the ground. For reference, the airport the flight took off from is at 7,631 feet above sea level—so the aircraft never reached more than 500 feet above the ground, not leaving much room for correction.
All of these planes had issues in the 6000-8500 feet range and then both suffered the same plunging fate after briefly recovering from a dip / nose stabilization issue.
Hopefully the ADS system is secure and not susceptible to infiltration/hacks causing the plane to react to incorrect data at a range that is not recoverable.
[1] http://www.thedrive.com/news/26933/amazon-boeing-767-cargo-p...
[2] https://arstechnica.com/information-technology/2019/03/faa-a...
For me, the 767 hit with some kind of strange turbulence then the engines went to max and caused a typical stall crash. You could google about the stall crash, such as : https://www.dw.com/en/why-do-airplanes-stall-and-why-is-it-s...
For both 737 max, their MCAS system thought they had a stall potential and triggered on nose dive while the airplanes were actually in the perfectly normal situation.
For 767, they do not have this autonomous system and the stall warning was not even triggered. So it was a purely manual operation. If they have a perfectly functioned MCAS like the 737 max has, they may have survived.
So a working MCAS could save plane from crash like 767 had. Just the MCAS on 737 max was not working correctly and MAY potentially have caused the tragedy.
It might not have anything to do with ascent/descent but the range / altitude and how systems react at that altitude. Maybe the systems are designed to be more reactive at this height because there is less chance to recover than if you are cruising at 30k feet.
What is most interesting to me is how the Amazon cargo plane 767 crash made almost zero news and has quickly been replaced in the news with the 737 Max 8/9 issues. Though all situations are very similar, straight nose dives that were sudden at the 6000-8500 feet range. There has to be something to that.
767 crashes and emergencies are EXTREMELY rare [1].
Most are related to human intervention such as terrorism, pilot error, fuel error and only a very small amount are mechanical errors. Of the 12 problems, only 5 were mechanical error.
[1] https://en.wikipedia.org/wiki/Category:Accidents_and_inciden...
The first 737 Max crashed in Indonesia, it did not have quite media coverage as well. The second one got much more attention because that is quite rare that two brand new flights crashed in such a short period. Everyone suspected it is caused by design or system flaws.
767 is a passenger and cargo plane.
I assume you mean the Amazon 767 was a cargo plane [1], in which case yes it got less attention as only the pilots/cargo crew were on board. However, a 767 crashing in the US seems like it would have gotten more focus.
Other 767s are passenger planes including two infamous ones American Airlines Flight 11 (Boeing 767-223ER) and United Airlines Flight 175 (Boeing 767-200) that were the two planes that crashed into the World Trade Center tower 1 & 2 on 9/11 [2][3].
For the most part these planes don't crash and a straight nose dive is a freak occurrence. Most of the other issues that weren't crashes with 767 were landing gear, short fuel related and terrorism.
It is very strange and eerie that this 767 from Amazon crashed almost similarly to the 737 Max 8/9 in Indonesia and Ethiopia recently, with such a long history of nothing like this.
[1] https://en.wikipedia.org/wiki/Atlas_Air_Flight_3591
[2] https://en.wikipedia.org/wiki/American_Airlines_Flight_11
[3] https://en.wikipedia.org/wiki/United_Airlines_Flight_175
Well, no, there doesn't have to be something to that. Random data often exhibits clusters.
However, it is unwise to assume that there isn't something to that. It just also is unwise to assume that there is. Keep an open mind and investigate.
The only thing I can think of why that would happen is a software error. Possibly the common occurrence of these planes where the nose is pitching up 4 degrees repeatedly to an almost turbulent state, then possibly the software on the plane thinks it is additively pitched up if there were 10-15 incorrect nose adjustments up of 4 degrees.
Could be a race condition, lock or some other issue where it only happens at that altitude because it has to adjust faster to prevent crashes and the lower you are the harder it is to recover.
My guess is the software acts more decisively at lower altitudes under 9-10k because that is closer to the ground and where some areas of the US have peaks in elevation to, so the aircraft must act quicker because it is less recoverable at that stage.
Another possibility is since it was a cargo plane there was shifting during the turbulent state but the plane was stabilized and moved up 100 feet before the -49 degree nose down state so that seems unlikely.
Pilot error seems unlikely as well as the pilots were skilled.
Mechanical failure seems rare as 767s just do not have nose down errors like this and almost no mechanical causes of emergencies or crashes.
There is always the possibility of sabotage as well, there are reasons groups might want to take down an Amazon owned plane.
Though most likely this is a software error or something exploited.
Also with respect to MCAS, it's not a factor when the autopilot is enabled, as the autopilot shouldn't permit approaching the edges of the normal envelope with such a high angle of attack that the MCAS routine would apply. I haven't read any report about whether the autopilot was enabled or not; let alone the exact sequence of modes enabled and disabled. So the autopilot itself can't be excluded as a suspect until there's evidence it was a non-factor.
What I recall reading is that the nose was already down somewhat before changing pitch 4 degrees THEN pitching sharply down in direct response to pilot input.
ADS-B data is being TRANSMITTED. The plane is not responding to ADS.
Possibly the software is more reactive when it is in ranges where it may be close to the ground or the highest elevation. Highest elevations in the US are around that range and a range where it is harder to recover so it may be more reactive. At 30k feet elevation/altitude (from sea level) or flight level (from ground) the software might not have to take as evasive moves.
https://www.flyingmag.com/everything-explained-all-about-alt...
https://en.wikipedia.org/wiki/Flight_level
Pilot vernacular often uses just `altitude` but which kind of altitude depends on context, but if there's ambiguity I say MSL (mean sea level) or AGL (above ground).
The previous commenter was thinking we were talking about height relative to ground level.
Altitude/elevation is height from sea level as specified.
I used flight level more as above ground level (AGL) but it also means not all the way to sea level and takes into account ground terrain.
The point being that all these planes (both 737s and the 767 Amazon cargo plane) all had similar issues between 6000-8500 feet in altitude (from seal level) but the ones that crashed on ascent took off from higher ground.
Moot point as it was sea level/altitude where all these plans had the same repeating/turbulent nose up 4 degree turbulence, then a direct 49 degree nose down state that led to the crashes, regardless of AGL.
Whether related or not the planes all experienced a nose dive in that range that was catastrophic.
It makes more sense with the 737 Max planes as they have center of gravity further back and software has to keep the nose down/straight, it makes zero sense for the 767 do similar movements unless it is common across software in that altitude range for some reason, possibly being more reactive at that range because of common elevations and it is harder to recover at that point.