News for hackers. Stuff that matters.
Re:Canada grounds Boeing 737 Max 8 by mcv on Wednesday March 13, 2019 @12:40PM (#19380473)
From what I understand about the issue so far, the systemic flaw is not in the aircraft, but in the training of the pilots. Or the lack thereof. Boeing wanted the 737 MAX to be considered just another 737 that didn't require any new training, but there are apparently significant changes compared to other 737 models that leave pilots feeling ill-prepared.
Re:Canada grounds Boeing 737 Max 8 by jostmey on Wednesday March 13, 2019 @12:50PM (#19380610)
My understanding is that the Boeing 737 max jet is unstable. Instead of redesigning a new plane from the ground up, Boeing placed the larger engines forward and upward on the wings. Boeing then tried to issue a "software patch" on a bad hardware design. Boeing came up with an automated system to seize control of the aircraft in the event that the flight became unstable and neglected to inform pilots of this new automated system. Although not confirmed, it appears that both crashes may be the result of pilots wrestling with the plane's automated system, which they knew little about and may have been faulty
Re:Canada grounds Boeing 737 Max 8 by cmurf on Wednesday March 13, 2019 @01:32PM (#19381177)
That is not accurate. There is a sufficiently different pitching moment in high angle of attack attitude that results in the engine nacelle producing lift. That means the flight characteristic of the MAX is sufficiently different to otherwise warrant its own type rating separate from previous 737 generation type rating. Boeing and FAA agreed to a software abstraction to obviate the need for a separate type rating for pilots.

The plane is not unstable. A pilot who is trained for the actual flight characteristics absolutely can handle the plane. The problem is they weren't required to get that training, because 737's still all share the same type rating as if they all have substantially the same flight characteristics. Which is generally true except one particular attitude, high angle of attack.

It's an open question to me (I'm no one, but I am a pilot and a flight instructor in a previous life) exactly how different stall avoidance could be with MCAS disabled (which is indirectly done by setting stabilizer trim to cutoff) once it has already induced a dive by having fought with it a little bit too long. If it's a normal procedure to disable the very software abstraction that means you don't need a type rating for this airplane, but the stall avoidance technique is substantially different from other 737s - that's very concerning to me. Not that the plane is unstable or unsafe. But that there's a disconnect in pilot knowledge and actual airplane flight characteristics with this abstraction turned off.

Re:Canada grounds Boeing 737 Max 8 by tzs on Wednesday March 13, 2019 @01:26PM (#19381085)
Unstability is not necessarily considered to be bad hardware design. Stability comes at a cost, such as less maneuverability or lower efficiency.

Before fly by wire, it was always worth going for stability because pilots couldn't fly unstable aircraft. Fly by wire enabled making flyable unstable aircraft, and so trading off some inherent stability for more maneuverability (military) or higher efficiency (civilian) became an accepted design option.

Re:Canada grounds Boeing 737 Max 8 by lisper on Wednesday March 13, 2019 @12:59PM (#19380741)
Private pilot here. That is exactly right, based on everything that I've read and heard from other pilots, some of whom fly the Max 8.
Re:Canada grounds Boeing 737 Max 8 by djsumdog on Wednesday March 13, 2019 @01:32PM (#19381166)
Other articles posted said there is a way to turn off the auto-correction, but it's not intuitive. On earlier 737s which had this, it came on when engaging automatic ascents/descents and was disabled when a pilot disabled those systems.

In the case of the Max8, it's always on and there is additional training needed to react quickly when this happens. I think all current pilots who are watching this situation should know what to do at this point, but the needed the awareness brought by the situation. In the two crashes so far, it seems like the pilots had years of commercial experience.

Re:Canada grounds Boeing 737 Max 8 by lisper on Wednesday March 13, 2019 @01:43PM (#19381315)
Knowing what to do and actually doing it under stress are two very different things. The aircraft that I fly, a Cirrus SR22, also has an electric trim and so there is the potential for something to go wrong and the trim to go rogue. When that happens, dealing with it is a "simple" matter of pulling the right circuit breaker. But the circuit breakers are on a panel located near your right shin. They are not easy to see nor to reach, and if you don't have the location of the trim breaker memorized, the odds of being able to find it and pull it while the plane is trying to kill you are rather low. Happily, there has never been a runaway trim on an SR22 (AFAIK) so this doesn't keep me up at night. But I can certainly understand how even a pilot that was prepared for this eventuality might fail to deal with it properly while under duress.
Re:Canada grounds Boeing 737 Max 8 by planteen on Wednesday March 13, 2019 @01:10PM (#19380891)
You are discounting the possibility of software controlling an aerodynamically unstable airplane safely. This has been happening for decades (see X-29, B-2, etc)
Re:Canada grounds Boeing 737 Max 8 by localhost on Wednesday March 13, 2019 @01:44PM (#19381319)
Sure, but those pilots knew they were flying that kind of aerodynamically unstable airplane ahead of time. These pilots apparently were thinking "this is just a 737".
Re:Canada grounds Boeing 737 Max 8 by mcv on Wednesday March 13, 2019 @05:25PM (#19384011)
Exactly. The plane might well be perfectly safe in properly trained hands, but it seems rather obvious by now that it doesn't fly the same as a regular 737.
Re:Canada grounds Boeing 737 Max 8 by jacquesm on Wednesday March 13, 2019 @01:44PM (#19381316)
Those airplanes were designed from the ground up with that control system in mind (the F16 is one of those). It's not something that you would normally retrofit on an otherwise unchanged design, the original 737 was never meant to be flown in that mode and a software patch to make it work that way sounds like a dangerous plan.

On another note, I don't think the 737 Max is inherently unstable, so I'm not even sure this line of reasoning applies at all or even in part.

Re:Canada grounds Boeing 737 Max 8 by sangnoir on Wednesday March 13, 2019 @02:06PM (#19381583)
Did any of those aircraft quietly oppose the pilots authority?
Re:Canada grounds Boeing 737 Max 8 by mcv on Wednesday March 13, 2019 @05:27PM (#19384027)
I remember reading 30 years ago that the F-16 has 3 computers that vote on how to respond to user input.
Re:Canada grounds Boeing 737 Max 8 by AtariBaby on Sunday March 17, 2019 @08:09PM (#19417668)
I just had to say wow, this is the most interesting thing I’ve read in a while
Re:Canada grounds Boeing 737 Max 8 by ishan1121 on Wednesday March 13, 2019 @01:11PM (#19380896)
This is the conclusion I too could come up with. But whatever, I think these 2 737 crashes have pushed fully automatic flights a bit further.
Re:Canada grounds Boeing 737 Max 8 by shanth on Wednesday March 13, 2019 @01:11PM (#19380897)
If this turns out to be a software bug of dangling pointer... oh imagine that poor developer has to take all the karma of lost lives...
Re:Canada grounds Boeing 737 Max 8 by jostmey on Wednesday March 13, 2019 @01:20PM (#19381014)
Yes, it's possible that an engineer made a mistake--Boeing will certainly blame everyone else including the pilots and engineers. But the bigger picture here is that the plane might be unstable, it used a "software patch" as a band-aid, and it neglected to inform pilots about the new automated systems.

I suspect the blame lies with Boeing's management and with the FAA for being complicit with Boeing.

Re:Canada grounds Boeing 737 Max 8 by drewg123 on Wednesday March 13, 2019 @12:59PM (#19380745)
Reminds me of a company I used to work for where when we updated our PCIe NICs, we would never, ever change the PCIe device or sub-device ID. This was to avoid a triggering a re-certification from various companies (MS WHQL, Vmware ESX testing, etc). The first thing the driver did was to talk to the firmware and ask it for a set of supported features..
Re:Canada grounds Boeing 737 Max 8 by elliekelly on Wednesday March 13, 2019 @12:50PM (#19380613)
My understanding is that it's a bit of both. This Slate article[1] described it as both a design flaw and an effort to paper over that design flaw by tweaking the autopilot system.

>Boeing swapped out the engines for new models, which, together with airframe tweaks, promised a 20 percent increase in fuel efficiency. In order to accommodate the engine’s larger diameter, Boeing engineers had to move the point where the plane attaches to the wing. This, in turn, affected the way the plane handled. Most alarmingly, it left the plane with a tendency to pitch up, which could result in a dangerous aerodynamic stall. To prevent this, Boeing added a new autopilot system that would pitch the nose down if it looked like it was getting too high.

Because of the changes to the autopilot system it functions differently than pilots expect it to during an emergency situation on this particular plane.

[1] https://slate.com/technology/2019/03/ethiopian-air-crash-whe...

Re:Canada grounds Boeing 737 Max 8 by msbarnett on Wednesday March 13, 2019 @12:57PM (#19380719)
Yeah, my understanding is essentially, the changes made the Max 8 not behave aerodynamically like other 737s. This would normally mean that pilots would need to requalify on it as a new plane type, in order to legally fly one.

But that's more expensive for Boeing's customers, so to make it more attractive, they decided to "paper over" the differences in behaviour with software. Fly it like any other 737, and if it starts to stall in conditions where other 737s wouldn't, the software corrects things!

Except, except, like all abstractions, this one's leaky -- and they've only shifted the difference in behaviour from "stalls at angles that would surprise a 737 pilot" to "noses-down in ways that would surprise a 737 pilot" along with "can't be muscle-overridden the way the autopilot on other 737s can be".

Re:Canada grounds Boeing 737 Max 8 by cmurf on Wednesday March 13, 2019 @02:23PM (#19381788)
It's not the autotrim that you can't fight. The plane can be trimmed (elevator surface) in such a way that pilot yoke input (stabilizer surface) cannot fight against. That is common to all 737's, and all pilots know about it. It is not that the autotrim feature can't be fought.

What's new in the MAX is the autotrim doesn't terminate when the pilot gives back pressure on the yoke if MCAS senses high angle of attack, it will do an uncommanded nose down (every I think 10s). This is to compensate for the MAX engine nacelle which provides a meaningful amount of lift to cause a pitch up moment in high angles of attack, unlike previous 737's.

I wonder whether the stall behavior between 737's is slightly dissimilar or very dissimilar without MCAS, between 737NG and 737 MAX? And then quantifying that difference. I also seriously wonder if it's even slightly more likely for a pilot to induce a stall as a result of not knowing about the natural uncorrected stall behavior of the new design compared to previous designs. MCAS is not my concern, it's how the plane flies without it that I wonder about.

Also, WSJ is reporting that the MCAS software update mandated by the FAA in an airworthiness directive was delayed due to the nearly five week government shutdown. I'm already imagining the tweets...

Re:Canada grounds Boeing 737 Max 8 by amyjess on Wednesday March 13, 2019 @01:33PM (#19381184)
> To prevent this, Boeing added a new autopilot system that would pitch the nose down if it looked like it was getting too high.

The biggest problem here is that the new autopilot system relies on a pair of Angle of Attack sensors. Those sensors are prone to damage, and if the sensor is damaged, the erroneous information it reports can engage the new autopilot system. One of the reasons why Southwest thinks it's immune to this problem, by the way, is because they paid extra for a system that displays an alert when the Angle of Attack sensors disagree with each other, so pilots are aware of what's wrong and can take action to stop it. Most other carriers, including Ethiopian and Lion Air, didn't.

Also, other autopilot systems can be disengaged by pulling back on the stick. This one can't: you have to manually disengage the system.

Pilots were never informed of any of this because Boeing wanted to maintain the fiction that the 737 MAX is no different than previous versions of the 737.

Re:Canada grounds Boeing 737 Max 8 by nabla9 on Wednesday March 13, 2019 @03:32PM (#19382702)
It is a systematic flaw.

Boeing announced that it had been working on a flight control software upgrade that includes updates to the MCAS flight control law, pilot displays, operation manuals and crew training. It's coming in few weeks and expected to be made mandatory by April by an FAA airworthiness directive.

Re:Canada grounds Boeing 737 Max 8 by carapace on Wednesday March 13, 2019 @12:53PM (#19380654)
What you are describing is a systemic flaw in the aircraft.
Re:Canada grounds Boeing 737 Max 8 by DoofusOfDeath on Wednesday March 13, 2019 @12:45PM (#19380532)
Perhaps I haven't followed the news closely enough, but wasn't the Ethiopian plane also spewing excess smoke at the time of the crash?

If so, wouldn't that suggest a failure mode other than the one that's viewed as a training issue?

Re:Canada grounds Boeing 737 Max 8 by msbarnett on Wednesday March 13, 2019 @01:02PM (#19380778)
Eyewitness reports of plane crashes nearly always involve fires in locations no two eye-witnesses agree on, smoke, wings falling off, etc. These are rarely born out by subsequent investigations.

See, eg, " Reliability of Eyewitness Reports to a Major Aviation Accident" https://commons.erau.edu/cgi/viewcontent.cgi?article=1040&co....

A plane crash is a very fast, horrifying tragedy to witness. It shouldn't be surprising that peoples' imperfect memories afterwards are completely scrambled

Re:Canada grounds Boeing 737 Max 8 by ceejayoz on Wednesday March 13, 2019 @01:35PM (#19381213)
Yup. Even professional pilots get it all wrong:

> almost all the eyewitnesses, including experienced pilots, gave grossly inaccurate accounts when compared to the film record

Re:Canada grounds Boeing 737 Max 8 by xefer on Wednesday March 13, 2019 @01:01PM (#19380761)
There were eye witness reports to that effect, but the proof would be in the flight recorders.

https://www.reuters.com/article/us-ethiopia-airplane-witness...