Hacker News Reader

Comment by madengr | Hacker News Reader

This will probably be the safest plane to fly on, considering the amount of scrutiny the FAA is now applying.

Would restarting Chernobyl make it the safest NPP considering the amount of scrutiny that it would be put through?

Maybe we should look into re-introducing Windows Vista, considering the scrutiny that it would require it must be the safest and most stable OS ever.

But Windows 7 was just Windows Vista with some relatively minor changes and it was fine.

The problem with the 737 MAX is that Boeing tried to make a plane with a completely different airframe and added a bunch of software to try to make it fly exactly like second generation 737s so that pilots didn't need to retrain. But the abstraction was leaky and it turns out that when you have a different airframe, you have to retrain your pilots.

That said, Boeing did some janky stuff to get this nonsense certified by the FAA and the whole design probably needs to be recertified, which will take years.

Windows Me was probably a better example but not everyone will remember it.

Anyhow, the point is that scrutinizing a bad architecture doesn't make it any better. Scrutinizing a chicken won't make it an eagle.

I worked at a help desk when Windows ME came out. As I recall, the most salient difference I found between it and XP was that they removed a bunch of tools that I had come to rely on for diagnosis, troubleshooting, and just generally understanding what was going on under the hood. I guess it was ostensibly better, but I never did feel as comfortable working with an ME box.

With that in mind, yeah, Windows ME sounds like a pretty great analogy.

But ME came out before Windows XP..

Yeah, brain fart. I was thinking 98.

> The problem with the 737 MAX is that Boeing tried to make a plane with a completely different airframe

You've got it backwards. The problem was that they used the same airframe with larger engines that had to be mounted differently due to ground clearance issues.

No, that reactor had an inherent instability that could not be fixed.

The MCAS has an issue with redundant AOA sensors; that specifically needs to be fixed. The MCAS caused problems on other flights, but the pilots knew to disable it by flipping the circuit breaker. The pilots on the crashed planes did not know to disable it. That is a training issue.

Fix the redundant AOA and provide training to disable the MCAS, and it will be fine.

>No, that reactor had an inherent instability that could not be fixed.

Interesting maybe could have been fixed with software. I am thinking of Reactor Characteristics Augmentation System, or RCAS.

Also, Chernobyl exploded because they pressed the halt button at a very bad time. This also sounds like something that can be fixed in software.

> No, that reactor had an inherent instability that could not be fixed.

There are many exactly same reactors as was in Chernobyl fixed and running still:

> Most of the flaws in the design of RBMK-1000 reactors were corrected after the Chernobyl accident and a dozen reactors have since been operating without any serious incidents for over twenty years.

Even Chernobyl itself had 3 remaining reactors working for many years after the disaster, last decommissioned in 2000.

The positive void coefficient problem[1] was intractable (without putting the whole reactor vessel inside a pressurized tank), but it was the combination of that and the control rod design that caused the explosion. The control rods were modified trivially. The reactor is basically safe, though you wouldn't repeat the design choices.

[1] An IMHO stupidly obfuscated way of saying that the cooling water around the reactor was an unpressurized pool that would boil if it got hot. Low-density bubbles of steam absorb fewer neutrons than liquid water, so when it boils the core gets hotter, not cooler.

Fun fact:

> Most of the flaws in the design of RBMK-1000 reactors were corrected after the Chernobyl accident and a dozen reactors have since been operating without any serious incidents for over twenty years. [99]

Including 3 remaining reactors of Chernobyl Power Plant itself.

[99] - https://en.wikipedia.org/wiki/RBMK

Little known fact: Chernobyl was still a functional power plant after the meltdown and the remaining 3 reactors in the building were operated until 91, 96, and 2000.

Ten plants of same design are still operational today.

MAX is Defective by Design. Leave the aerodynamically unstable designs for the hotshot fighter planes; everyone else just wants to get from A to B without cratering into the ground inbetween.

It's not aerodynamically unstable. It's aerodynamically different than the 737-800 but Boeing decided instead of new pilot training and possibly re-certification, they would address the difference in it's flight characteristics by automatically adjusting the elevator trim when the aircraft is in a specific off-nominal situation such that the 737 MAX 8 feels and behaves just like the 737-800 (MCAS). That's not a bad engineering design decision IMO, but the follow-on decisions (or mistakes) made by Boeing (one AoA truth source instead of two, no extra training, zero indications, basically don't tell anyone about it) is the problem. Once those problems are fixed, no 737 MAX 8 will ever fail due to MCAS again. In fact it's going to be one of the - if not the - most scrutinized (and thereby safer) aircraft ever made.

I exaggerate slightly for [black] humor, but the point is hanging those huge engines so far forward puts its center of thrust uncomfortably ahead of its center of gravity, making it naturally want to pitch nose-up in flight.

All the rest—dodgy sensors, crappy software kludge, no pilot retraining—is hacks laid on top of that initial flaw just to pretend it isn’t there. A lot of us here build complex systems for a living; if that isn’t a “smell”, what is?

(Aerodynamically unstable design is a thing, btw, e.g. Typhoon. But that’s Military, which can afford fly-by-wire software that works.)

A stable design might’ve lifted the airframe further from the ground to give those huge engines the clearance they need under the wing, keeping them to the COG and thus ensuring the plane continues to fly naturally level as per its original design. (Any air geeks want to check me on that?)

But obviously that would’ve cost more, and now Boeing’s run by short-term profit chasers, not adult engineers, the attitude is “What can we get away with?”, not “What do we need to build?” That alone’s reason not to trust their bloody product any more. Not least cos what else might they be tempted to cut corners on?

I mean, a shitty iPhone will screw up your day; a shitty Boeing will screw up the rest of your life.

This is why Corporate Manslaughter laws should include hanging the Board by its heels.

Is it unstable? I thought the problem was that they didn't want to require retraining so they built in a flight control system that made it seem like the previous generation plane from a flight dynamics pov. That system would only take inputs from a single sensor on any given flight. If that sensor had issues then the flight control system would misbehave. If they had gone the retraining route and not included MCAS, then the plane would be fine. At least that was my understanding.

I think most aircraft are unstable or have quirks outside their approved flight envelope. For passenger aircraft you want it to be docile and predictable inside that. The MAX far as I've gleaned naturally wants to pitch up at high angles of attack. And that the problem is more severe than they expected when they initially designed the thing.

The MCAS might have been okay if the behavior matched initial expectations.

Are you sure ? I though it wanted to pitch up when engines where at full throttle, not because of angles of attack.

I think what I've read is the engine nacelles create lift at higher angles of attack. Because the nacelles are larger and located more forward than in older models. That's the source of the extra lift. Might also be the engines also contribute.

On the other hand I'm just a bystander in all this.

Also consider the position of the thrust point relative to the plane’s center of gravity. Being positioned below and ahead of the CoG, the engines naturally try to push the plane around it, i.e. nose-up.

Mind you, my physics ed stopped after high school (though even that much is enough to do a double-take at MAX). This looks like a much better explanation:

https://www.quora.com/Where-should-the-center-of-mass-lift-a...

It pitches up with throttle, due to the position of the engines. Angle of attack sensors feed MCAS, so it can detect the pitching up and adjust the elevator trim to counteract it.

I suspect they’d need to teach those pilots to add a lot of trim. Putting those big engines well ahead of the CoG makes them throw its nose up; it’s basic Newtonian levers.

Not that any plane flies perfectly level at all speeds, which is why they need trim in the first place. But presumably there’s a “sweet spot” which the original 737 was designed to nail in normal flight, which subsequent revisions have moved further and further away from.

At some point you really need an adult in the room to stand up and declare “no more”. But by accounts McDonnell-Douglas already showed all the grown-ups the door. That’s the nice thing about responsibility: just disseminate it enough, and no-one’s to blame.

Reply on news.ycombinator.com