Hacker News Reader

Comment by Miraste | Hacker News Reader

Two of those are manufacturing mistakes, and it seems likely the door one is as well. Not that that helps the passengers, but they're not systemic design flaws.

That anti-icing system is deranged, though. They effectively installed a timed detonator on the engines and want a safety exemption for it.

View on HN

Ah you’re right. They’re not design flaws, just manufacturing issues, so sign me up for the first flight when the max flies again…

People falling out the sky because the wing falls off because someone forgot to bolt it on aren’t going to care if it’s a design issue or a manufacturing issue. Boeing is doing both, so the blame lies with them either way.

The statement pointing out that it’s a manufacturing error (I’m guessing) was not intended to be solution to the problem. It pointed out that those problems have less to do the certification of the design and more to do with a manufacturing defect. If the manufacturer created parts that were up to specifications, these things would have not been a problem. This is a very important distinction because a design flaw is a much bigger deal for this aircraft type than that poorly manufactured components.

Boing is still at fault, yes, but we should exercise restraint in becoming too reductionistic on complicated engineering problems.

> It pointed out that those problems have less to do the certification of the design and more to do with a manufacturing defect.

You might really enjoy a book called the The Design of Everyday Things by Don Norman. He covers your comment here in detail.

One of his key insights is that "human error" are far too often weasel words that prematurely end a conversation (or investigation) into root cause. He goes on to detail a whole tree of "human error" so we can speak about mistakes, lapses, and errors that humans make.

The meat of his point is that some types of human error are very hard to design out of your system, but _many_ types of human error can and should be expected by the designer (or engineer), and appropriately handled.

In this instance, if a human (or perhaps a few) can make a single error when affixing the door plug to the aircraft, like improperly torquing a bolt. And that simple error risks a catastrophic loss of the airframe, then you probably have a _design_ issue and not a "manufacturing issue caused by human error".

Why exactly should we "exercise restraint in becoming too reductionistic on complicated engineering problems"?

When we reduce complicated problems down to inaccurate trivial ones by stripping out important details and nuance, we end up with a caricature of the original - one that easily devolves into a to strawman argument to serve someone’s point. This new representation masquerading as the original can carry the same weight as the one it was based off of.

This is spreadsheet brain thinking – likely the same MDD dorks used to justify cutting corners.

There's no nuance when people are dying. None whatsoever.

If someone can't agree to that sort of black-and-white thinking, probably they should be working in an industry where innocent lives aren't dependent upon sound decision-making.

I absolutely agree. I meant it as a comparison to issues like the infamous MCAS, which was wrong on purpose on all 737 Max 8s everywhere.

There isn't a change in outcome between the flaws, but I think the difference between a mistake and a known issue that was left in while the company tried to change regulations to allow it, all for a tiny cut to the BOM, is worth noting.

Question is, why were these things manufactured wrong. It's well possible that Boeing's engineering documents are poor or misleading, triggering human error during manufacturing.

This is of course pure speculation and it might equally well be some single manufacturer pressuring ("optimizing") their employees (or even machines) past the point of reliability.

Either way I'm not gonna fault anyone for refusing to fly on a 737 MAX. At some point you gotta make a call and shift your assumption from "isolated engineering/manufacturing mishap" to "corporate screwed the entire product top to bottom".

> Spirit [AeroSystems] is responsible for the entire fuselage, including the cockpit, in all Boeing jets, and the entire fuselage for the 737 MAX models, according to the Seattle Times.

https://www.msn.com/en-us/travel/news/boeing-737-max-loses-e...

So Boeing didn’t actually manufacture the plane. But they are still responsible for ensuring its manufactured correctly.

A manufacturing mistake is, in all likelihood, another type of systemic fault. Why would you think only one aircraft would suffer from it?

Back in 2014[1] Al Jazeera (the international edition) had a pretty good in-depth report of issues with the manufacturing line for the 787.

There are known issues regarding quality assurance at Boeing for a decade now, they keep going down the drain. The MBAs from McDonnell-Douglas won, and properly tarnished Boeing's image...

[1] https://www.aljazeera.com/program/investigations/2014/7/20/t...

The troubling thing to me here over and above these issues is if they (Boeing) think some of these things are ok to the point that they ask for an exemption, what else is there that would fail rigorous safety checks but has been deemed ok by management, and has not come to light yet? We may never know until it’s too late.

A manufacturing mistake which makes it into service is a failure of QA and testing systems design. (At least above a certain threshold which varies depending on the industry etc. etc.)

Depends. You can’t do e.g. crystallographical analysis of metals inside an engine if the engine is already put together. You have to trust your vendor that they built the engine correctly and checked the materials themselves. Or at most send auditors to the vendor.

> You can’t do e.g. crystallographical analysis of metals inside an engine if the engine is already put together.

QA is not forced to only deal with assembled components. They can be embeded before the assembly process to QA the parts. For example they could peform crystallographical analysis on a subset of blades which are ready for assembly.

They can also take apart a certain percentage of randomly selected engines, perform crystallographical analysis on a subset of parts and then re-assemble the engines. Many options.

Titanium fan disks, for example are all required to be inspected not just when installed but also at regular engine maintenance intervals. The inspection requires essentially complete disassembly of the fan (so it is required during particular engine maintenance events) followed by the application of penetrating dye and inspection at a narrow granularity.

This kind of maintenance and inspection actually can be required and performed, it just costs more time and money. If we want to be all market capitalism about this we could require tests necessary to ensure safety and let engineers, business people, and executives make decisions that price in the cost of dangerous and risky designs that require constant and invasive inspection and maintenance.

The only real difference with today would be regulators having a spine and/or more than pro forma power to enforce their decisions.

> but they're not systemic design flaws.

Hum, yes, instead Boeing seems to have systemic manufacturing flaws. Do we have a reason to believe those are contained into the Max line?

Isn’t the manufacturing process and quality assurance part of the design in “products” like these? It is in car manufacturing so i assume it should, I don’t believe is sort of an “artisan” production line.

The procedures are part of the design certification. And while some vary from one design to another, many of them do not.

Separating “manufacturing mistakes” only makes sense if someone else is responsible for manufacturing. As far as I know, Boeing is responsible for both the design and the manufacturing of these aircraft so the difference is purely informational, but in terms of criticizing Boeing mostly irrelevant.

Spirit AeroSystems makes the fuselage for the 737 Max. [1][2]

It's even in the article: "Spirit AeroSystems, which makes the fuselages for the planes, referred CNBC to Boeing when asked about the incident"

[1] https://en.wikipedia.org/wiki/Spirit_AeroSystems

[2] https://www.spiritaero.com/

For all intents and purposes, Spirit is part of Boeing.

Spirit was Boeing Wichita until 2005, and today Boeing represents 85% of sales.

Boeing decided to outsource to Spirit, Boeing is responsible to attest the quality of what Spirit is delivering for a Boeing product.

If Samsung/Apple would outsource their batteries to Megabattery Company LLC and those batteries started to randomly explode we would all be blaming Samsung/Apple for not doing proper QC. I hope we all hold Boeing to a much higher level of scrutiny than cellphone manufacturers.

At some point it has to be simple liability for the final seller, ignoring all subcontractors.

Otherwise there’s too many shellgames you can play.

So not only the design quality is flawed but also the manufacturing process is botched up. That’s reassuring, I guess, because two kind of flaws cancel each other out

Sure they do. At this rate the planes will stop working well enough to take off, eliminating accidents forever.

This looks like manufacture cost cutting - an issue no amount of good design/engineering can fix.

Boeing is no longer engineering focused. It’s a numbers business pumping out planes as fast and as cheap as they can get away with.

design for manufacturing is a part of engineering

Reply on news.ycombinator.com