Hacker News Reader

Comment by tiahura | Hacker News Reader

An MBA designed MCAS?

An MBA designed the door plug installation process?

Maybe it’s the Boeing engineers that need some scrutiny?

I think MCAS was one of those organizational failures. I could see myself implementing something like that. I imagine the conversation would go a little like this:

"We need to replicate the pitch performance of the 737-800. These engines pitch the plane up when accelerating and it messes up go-arounds."

"What if we have the computer just pitch down to compensate? We know the angle of attack and the engine thrust output."

Meanwhile in another conference room:

"United is telling us that Airbus is giving them an A321 for 10% less than us. What can we cut?"

"No regulation requires us to have 3 angle of attack sensors. We could get by with 1 and stick it on the Minimum Equipment List if it fails."

"That sounds great."

Then the 1 angle of attack sensor fails the system that was designed for 3. Was it the engineer who agreed to hack 737-800 handling emulation onto the 737 MAX that failed here? Was it the engineer that agreed that the AoA sensors didn't need to be redundant, perhaps before MCAS was even invented?

These complex failures are rarely the result of one individual failing. Everyone did their job; the business saved United 10% (I actually have no idea who their launch customer was), the engineer saved thousands of pilots from being pulled off the line for re-training. But combined, it was a tragedy. Ultimately, it's the organization and its processes that failed, not an individual. As a manager, you own the organization and the processes.

Minor correction about the Max8 that makes it worse.

Turns out the Max8 does have two angle of attack (AoA) sensors.

But the MCAS system only read from AoA1, and never validated it against AoA2. It was recieving info from AoA2, but never used it!

Makes sense. Two sensors are about as useful as 2-node Consul cluster.

What do you do with 2 sensors?

You can do something with 3 or more, but 2 are completely useless.

Why would it be useless? Boeing's update for the Max8 MCAS (to make the FAA happy), was to make the system validate the two values.

If there is a mismatch, the system outputs an error and MCAS will disable itself.

In fact, that is literally one of the requirements set by the FAA to begin flying the Max8 again

> FAA Airworthiness Directive approved design changes for each MAX aircraft, requiring input from two AoA sensors for MCAS activation, elimination of the system's ability to repeatedly activate, and allowing pilots to override the system if necessary.

You can signal an immediate, specific warning if they disagree.

Turn on the AOA sensor error idiot light

Stick pushers aren't anything new, and the sad irony is that except for the two tragedies, MCAS worked so well that in the thousands of flight hours the Max had, no pilot had even noticed it.

It absolutely was the fault of the engineers who failed to realize what would happen with an AOA fault. I'd love to be corrected, I don't believe a single email from the engineers to management telling them would would happen was ever dug up. And, there's never been an engineer whistleblower saying "I warned them."

Has a single PE lost their license over the Max?

> that in the thousands of flight hours the Max had, no pilot had even noticed it.

This is false. At the time of the incident (>1month) The FFA and NTSB found 23 reports of uncommanded pitch down attitudes during takeoff.

Quick Google gave me this.

https://www.sbs.com.au/news/article/it-pitched-nose-down-us-...

Normally I would agree if an engineer directly knew what they were doing was bad and didn't say no but in the case of MCAS: pilots had no idea it was there. Yes, it also wasn't properly redundant and that is bad but had the pilots known it had existed in the first place and understood it was operating off a single data source they would have acted accordingly and a lot of people would still be alive. Door plug I can't speak to.

Hiding MCAS to avoid recertification was well above the paygrade of an engineer at a bench poking a circuit.

Also, if you want a short version of how this stuff goes very wrong even when engineers do say something, the challenger disaster[1] movie is great. Chase it with the NASA documents[2] for the long version.

[1]: https://www.youtube.com/watch?v=bvv2-7iOD_8

[2]: https://www.nasa.gov/history/rogersrep/genindex.htm

Edit: typo

Engineering failures don’t just happen in some contextless organizational vacuum

Engineers can't fail, it's always management's fault.

Engineering needs to be embedded in a system that can handle the fact that engineers fail and still produce reliable and effective outputs. The primary purpose of management is making sure such as system exists and functions. Managers are the engineers of the entire meta-system, basically.

If management don't want to take responsibility, they could take a pay cut and give the money to the engineers.

Unironically, yes? Management is where the buck stops.

Yes, in fact the Captain is always responsible for literally everything that happens on the ship.

> Engineers can't fail, it's always management's fault.

or, what can engineers possibly know about engineering: https://www.airliners.net/forum/viewtopic.php?t=213075

Engineers can fail, and when they don't QA may, but it's management that chose to believe otherwise

Sounds like you suggest that an MBA have no influence on the production and output of the organization, the performance, not at all, they just stand there watching helplessly while others do something some to them uncontrolable way.

Can they all be fired and the wealth they leach redistributed then please? :)

The three dimensions of every project are always:

Cost

Quality

Time

MBA push for cost, then time (to market), above quality.

Cost and time can't really be fudged. There are multiple ways to deflect and obfuscate issues with quality, so of course that comes last.

Curious who believe makes so-called "executive" decisions in an organization...

If you believe it should be engineers, we're fully in agreement, but then why do we have managers...

Maybe the QA engineers were fired because they were doing their job but the MBAs optimised for the wrong metric?

Ding ding ding.

They don't put Boeing engineers in management or executive positions these days.

Heck, they fired the last CEO over the 737 crashes but he was literally only on the job for 3 years and had no involvement with the 737 design. The problem? He was an engineer and he was already stirring up shit inside. So they threw him under the bus and replaced with him another useless goon that spouts PR bullshit and has no technical experience other than milking companies dry.

Read about the Space Shuttle Disasters. There are several good books on them.

Reply on news.ycombinator.com