Hacker News Reader

Comment by kjar | Hacker News Reader

This plane appears to be a (sorta) flying dumpster fire rushed to market against Airbus with tragic consequences. One design flaw after another is MAX failure.

View on HN

The industry is pretty homogeneous and somewhat of a revolving door between the few OEMs. What are the odds that the code on this plane is orders of magnitude worse (in terms of code quality) than every other similar bit of code in the sky?

Edit. I should have to say this but I'm not defending Boeing here. I'm saying it's not at al unforeseeable that the other planes, even ones by other manufacturers are just as bad. Clearly they thought nothing was abnormal about this one until it started falling out of the sky. I see this kind of like diesel-gate. If one of them is cheating they're probably all cheating.

Or conversely what are the odds that the others have similar undiscovered or undisclosed issues?

Fairly low, given that every other type in widespread service has been flying much longer than the MAX with a very good safety record

And Boeing do have form when it comes to software.

https://www.engadget.com/2015/05/01/boeing-787-dreamliner-so...

Upgrades to the 737 do have a particularly constrained design space, though (due to the oft-mentioned very low ground clearance).

Why can’t they change the ground clearance and remove the constraint?

The engines on the original 737 design were quite small by modern standards, and the wings were low as well. As such there aren't really a ton of places to put new, larger engines. They basically have to be placed further forward and higher up. This means the pitch moment created by throttling up/down is different on different versions of the 737, which is what required the MCAS device in the first place - they were trying to make sure the 737 MAX flew the same as previous models.

I think you’re just repeating the impact of the low ground clearance, rather than explaining why there still has to be low ground clearance.

You’ve said they can’t have larger engines because ground clearance is low... we already know that... but why must ground clearance be low?

Ah, you mean why did they originally do it? Sorry, I misunderstood. The original reason was that the 737 was designed that way so that it could be boarded with just stairs instead of a skybridge: https://simpleflying.com/737s-low-to-the-ground/

This was super helpful to small carriers at the time, but as time has gone on it's caused problems.

Because they would have to redesign the parts of the fuselage that holds the landing gear when they are retracted - as far as I know, there's not enough room to house taller landing gear.

Boeing did actually design extendable landing gear for the 737 MAX-10 model, which hadn't shipped yet. That aircraft is longer. To avoid tail strikes, the rear gear becomes longer as the aircraft is taking off.

I'm sure that costs more money. If it were on all 737 MAX aircraft, perhaps the engines could be lower. It'd be like a kneeling bus that gets lower when stopped to take on passengers.

Another option would be to make the engines move. They could be folded up or slid up-forward when driving around the airport. The FAA might have issues with this though. I don't know of any aircraft that does transformer stuff while zooming down the runway.

True, there are planes that compress the struts before retracting the gear for the gear to fit, so perhaps something like this could be designed for the 737 too? But that's increased complexity on a non-redundant part that you definitely don't want to fail, so it's not obvious how the trade comes out...

Not "perhaps". It actually was designed for the 737 MAX, which comes in 4 lengths. Only the long one gets the fancy gear.

Short video with gear model in action: https://www.youtube.com/watch?v=F4IGl4OizM4

https://www.geekwire.com/2018/boeing-737-max-10-landing-gear...

Scroll down for a nice diagram: http://www.b737.org.uk/737max10.htm

To be fair a similar system on the Dash 8 Q400 has a pretty nasty habit of folding up on landing.

You're talking about changes that would cause the plane to be a very different plane than the 1967 737, and would then need a new type rating. If you're going to do that, you might as well just throw out the 737 design altogether and make an all-new clean-sheet design. Boeing wanted to avoid that because that costs a lot more money and pilots would need retraining on the new type. They wanted to be able to call this piece of junk a "737" and claim that pilots didn't need any new training to fly it.

Because the airlines use hangars that are built for the current dimensions of the 737, not a bigger airplane.

And also you can't just redesign the landing gear without redesigning a substantial portion of the wings.

They could, but it would mean that a lot of the supporting infrastructure at airports would have to change, essentially negating much of the advantage from avoiding a new clean-sheet design.

As long as it remains sufficiently 737-shaped airlines can keep flying it to all the place they've been flying 737s for decades.

What supporting infrastructure would have to change? Jetways are compatible across a wide variety of planes. Yes, you can't use the same jetway on an A380 as a 737 (the A380 is just way too big), but you certainly can use the same jetway on a 737 as on an A320, and many other planes besides.

The reason they didn't want to do a clean sheet redesign of the 737 was to avoid pilot recertification, not because of ground infrastructure. Airports are already used to servicing dozens of different types of planes. Adding one more to the mix wouldn't materially change anything.

I can assure you you can use the same jetway on a 737 that you can use on an A380.

You sure? The A380 is so much larger (especially in width) that it can't fit into the narrower gate spacing typically seen for smaller planes. Sure, maybe one in isolation could service an A380, but ten in a row in typical airport spacing can't service ten A380s.

Separately, there's the issue that the A380 is so large that you need multiple jetways (at least one for each level) to efficiently load and unload everyone. Maybe you could slowly unload an A380 by debarking everyone off the lower level and forcing everyone on the upper level to walk down the plane's internal staircase, but this is extremely sub-optimal. In practice airlines don't do this; they use specialized gates set up for the A380's special needs.

That doesn't mean it can't use the same _jetway_ (we were discussing jetways, not gate spacing) you'd use on a 737 or that a 737 can't park at a stand wide enough to fit an A380.

They could, but my understanding is then it would count as a new airplane, with more training required for pilots to fly it. They were trying to pretend it was not a completely new airplane, so that pilots wouldn't need much new training to be certified to fly it.

It is turning out to have been a pretty bad decision, even from a mercenary point of view.

The constraint is the FAA would see it as a new aircraft and therefore have to certify a new airframe. The very in-depth review the FAA is currently doing on the MAX was the expensive (time and money) Boeing sought to avoid.

In hindsight, the in-depth reviews sound like a good idea in all circumstances. Certainly the answer is not to not do them!

Obviously this is not a great example, but with limited resources they will obviously need to divide them appropriately.

They can't do an in-depth forensic analysis of everything without unlimited money.

A longer landing gear to get the plane higher off the ground doesn't fit into the existing gear well in the wings. So you would have to redesign the wings. This required making sure that the loading on the main body stays the same. By now you are doing more than 50% the design work of a new airplane and you definitely can not keep the type rating. And if you need a new type rating anyway, you would redesign the cockpit, probably to be similar to the glass cockpits in the 777s.

It's worse than other such planes. For instance, the accident on Qantas Flight 72 is superficially similar - faulty data caused the Airbus A330's flight computer to ignore pilot input and abruptly pitch down, injuring a number of passengers - but it really wasn't. The Airbus engineers knew that angle-of-attack sensors fail, so unlike Boeing they designed their system to compare the input from multiple AOA sensors and ignore bogus data.

Unfortunately, that comparison algorithm had a flaw - it got confused when it received spikes of invalid data with a certain timing pattern and erroniously used the invalid data. That pattern shouldn't have been possible. No-one has been able to figure out any possible cause for it even in retrospect, and they certainly didn't anticipate it. However, the engineers designing the system did realise that the flight computers could have subtle bugs triggered by specific data timing - so not only did every flight computer have a monitoring channel running independently-written code checking its calculations, that monitoring channel was intentionally not synchronized with the main channel or any other flight computers. This meant every time one of the flight computers acted on bogus data and forcibly pitched down, the monitoring channel calculated values so different that the fault detection disabled its ability to do so within a few seconds.

The maximum allowed authority and the altitude at which the system was enabled were also much more carefully restricted than MCAS, so it couldn't take such erroneous actions in situtations where the pilot might be unable to recover. Combine the two safeguards, and something like Qantas Flight 72 with a few passenger injuries but no crash was pretty close to the worst-case scenario that could be caused by this weird and incredibly unlucky issue.

> That pattern shouldn't have been possible

Huh, this is new to me. I thought it was simply that two sensors failed, and their failed values outvoted the correct value in the quorum vote with the third sensor. What's the mystery about?

Section 5.3: https://www.atsb.gov.au/media/3532398/ao2008070.pdf#page225

A microprocessor failed and output corrupted data.

My instinct is that you're probably correct, and most planes are probably riddled with similar problems and would look bad if they came under the same scrutiny as the 737 Max is currently under. For example, I remember after the 787 was launched there was a few stories in the press about the batteries catching fire mid-flight, but it turned out the designers and engineers had accounted for that possibility and it wasn't really a big deal. They've improved the batteries, but it didn't actually hurt anybody.

However, the problem with the max seems to be that instead of designing for failure - assuming that critical systems might fail and the plane should recover gracefully from the failure of most internal systems - they seem to have designed for success, and made the assumption that no part of the plane will ever fail. And that's obviously not realistic, and we're seeing the consequences of it here.

> My instinct is that you're probably correct, and most planes are probably riddled with similar problems and would look bad if they came under the same scrutiny as the 737 Max is currently under.

The 737 MAX only looks bad because it's track record so far is that it's a death trap.

I remember after the 787 was launched there was a few stories in the press about the batteries catching fire mid-flight, but it turned out the designers and engineers had accounted for that possibility and it wasn't really a big deal.

It was a huge deal and the plane was grounded for three month due to those battery fires.

Originally the engineers did not account for battery fires and needed to refit a compartment to seal the batteries in case a fire occurs.

Luckily it was either on approach, or they could emergency land the plane when the problem occured.

A burning lithium battery in a flying plane is one of the worst imaginable scenarios you can encounter during a flight and is a fucking big deal.

Edited to add : https://en.wikipedia.org/wiki/Boeing_787_Dreamliner_battery_...

> My instinct is that you're probably correct, and most planes are probably riddled with similar problems and would look bad if they came under the same scrutiny as the 737 Max is currently under.

0.5% of all 737 MAXes which exist have killed everyone on board. No other mass market production airliner in recent history has that sort of record, AFAIK.

Wikipedia says 393 have been produced, so more like 0.25%.

2 out of 393 is roughly 0.5%

The rate of fatal crashes is more than an order of magnitude higher than anything introduced in the past 30 years.

The Concorde is the only thing that's worse. The A310 is closer than anything else at 1.35 fatal crashes per million miles vs the 737 MAX's 3.08- and that one was also designed to share a common type rating with an existing design.

http://www.airsafe.com/events/models/rate_mod.htm

Reply on news.ycombinator.com