Boeing Crashes Highlight the High Costs of Cheap Government
nymag.com
nymag.com
What if a bigger, fatter, better funded FAA is still in cahoots with Boeing?
That doesn't even pass the smell test.
This is classic motivated reasoning. The authors know that the US Airline safety profile over that time is stellar. Knowing they can’t base their claim on the actual statistical safety record, thy resort to using a comparison with vague words such as ‘suggests’ and ‘inordinate’ and emotionally charged phrases: such as ‘corporations to gain inordinate influence’ and ‘many people died.’
They are hoping that your emotional brain will overlook the lack of argument of real quantitative evidence and jump on the emotional bits of the argument.
This is similar to how anti-immigration groups will seize on a high profile crime committed by an immigrant and use that to say that immigration should be stopped/cut back without discussing if immigrants are actually more likely to commit violent crimes than natives.
Yeah, well, great. But aircraft development timespans are on the order of ten to fifteen, sometimes twenty years—how many new models (or significant variants) were shipped in that window?
IOW, any "ten-year-old track record" in transport category aviation isn't really saying much due to the timescales involved.
Another good example of that is the defective Bay Bridge replacement between Treasure Island and Oakland. That bridge is going to fail during the next earthquake because of design flaws that the oversight agency didn't have the manpower to spot. CALTRANS after three decades of Republican lead budget cuts doesn't have the ability to safely oversee projects of that size anymore.
CALTRANS after three decades of Republican lead budget cuts doesn't have the ability to safely oversee...
The new Bay Bridge debacle falls squarely on the Browns.
The FAA also has a lengthy history of certifying planes on the basis of "equivalent level of safety" despite having characteristics that violated airworthyness regulations.
See the 727, 707, and several other Boeing offerings that were rejected by the ARB in the UK. D.P. Davies is the primary source. His interviews are archived by the Royal Aeronautics Society.
No underhanded motivation here. Just wanted to set the record straight.
Other countries aren't fairing so well. Is this a failing of the U.S. Government's FAA? No.
I have a good friend who was a Naval aviator and did crash investigations for the Navy, his ordered list for causes of aviation crashes. The top 3 causes he believes are responsible for 99.9% of all crashes.
1. Pilot error due to fatigue (Which the U.S. has made strides to drastically reduce) 2. Pilot error not due to fatigue 3. Pilot lack of training 4. Plane failure (hardware, software, mechanical, etc)
What specific difference training did U.S. pilots receive that foreign pilots did not? How are pilots in other countries improperly trained? What is the evidence either 737 MAX crash had anything to do with training? If better training can mitigate flawed design, should that absolve any portion of manufacturer liability for any discovered flawed design?
In a proper investigation, all bias must be removed, and that includes statistical likelihood of prior causes of crashes.
Well, they all have far more than 200 hours of flight time, for one. Even the copilots.
> What is the evidence either 737 MAX crash had anything to do with training?
The fact that the problem was displayed to the pilots (trim severely out of norm) and they didn't follow the correction procedures properly. Though that might not be "training" as much as "competence".
> If better training can mitigate flawed design, should that absolve any portion of manufacturer liability for any discovered flawed design?
Really depends on the flaw.
You assume without evidence the problem manifests like runaway trim. And then proceed to assume without evidence the incompetency of two pilots. And yet you refuse to assume the incompetency of the automation, which didn't mere fail, it countermanded with lethal force the 100% correct control inputs of the pilots. The idea they have to assert their sanity by flipping two switches, is about as ridiculous as automation not having to assert its sanity when it, by design, trusts a single point of failure one upping that by trusting it while it's failing and one upping that to extreme by taking action on it without any regard to other input sources including the pilots themselves. It's beyond obscene to me.
I have personally trained private pilots with competency at identifying failed gauges, knowing which one to disregard, and estimating the missing values by inference from the remaining gauges. And this is in sub-200 hour pilots. The skill required for the instrument rating exceeds this particular automation in the AOA failure case by quite a lot. And the skill required for commercial and ATP ratings is even greater than that by quite a lot. Meanwhile MCAS upset from judgement of a single sensor induced such an unusual attitude it became unrecoverable.
So the assumption of incompetency by pilots, while cutting automation all kinds of slack is to me 180 degrees backwards.
> You assume without evidence the problem manifests like runaway trim. And then proceed to assume without evidence the incompetency of two pilots. And yet you refuse to assume the incompetency of the automation, which didn't mere fail, it countermanded with lethal force the 100% correct control inputs of the pilots. The idea they have to assert their sanity by flipping two switches, is about as ridiculous as automation not having to assert its sanity when it, by design, trusts a single point of failure one upping that by trusting it while it's failing and one upping that to extreme by taking action on it without any regard to other input sources including the pilots themselves. It's beyond obscene to me.
I didn't say that it manifests like runaway trim. I said that it manifests as "severely outside the norm trim", which it most certainly does. Unless you're saying that full nose down trim is normal? As for "asserting incompetence without evidence", they're certainly less competent than pilot #3 on the Lion Air flight the night before. (since, you know, that guy realized what was going on and how to resolve it)
As for the automation being incompetent, yep you're right! It is! But airplane crashes don't happen for one reason - not commercial flights at least. They happen when a sequence of failures occurs, just like here. Boeing failed to consider the worst case behavior of MCAS, Lion Air maintenance fucked up, and the Lion Air pilots weren't able to diagnose the failure that morning. (The failure pipeline of the Ethiopian crash is still to be determined) As for why Boeing failed there... It's at least partly because the automatic trim controls weren't considered a safety critical part, since they had multiply redundant backups. It's just that the backups were "turn off electronic trim" and "grab the trim wheel and hold it in place", both of which require pilot action. (Allowing stuff like this to get grandfathered in is arguably a hole in the FAA's systems)
> I have personally trained private pilots with competency at identifying failed gauges, knowing which one to disregard, and estimating the missing values by inference from the remaining gauges. And this is in sub-200 hour pilots. The skill required for the instrument rating exceeds this particular automation in the AOA failure case by quite a lot. And the skill required for commercial and ATP ratings is even greater than that by quite a lot. Meanwhile MCAS upset from judgement of a single sensor induced such an unusual attitude it became unrecoverable.
Well then why was the issue so difficult for the pilots to diagnose here? There's a gauge for trim, and I'd hope it would be looked at when trying to figure out why you have to keep pulling back on the yoke so hard.
Trained Behaviour:
- In case of continuous trim
- perform yoke jerk to disable the automatic trim (per previous 737 models)
- turn off electronic trim
- grab the trim wheel and hold it in place
---------------
MCAS Behaviour:
- On nose down event, pilots pitch the flight up, adjust trim
- Which disables MCAS for 5 seconds
- No continuous trim event manifests
---------------
From what I've read on some pilot forums, Boeing made a few errors:
- Removed "Yoke Jerk"
- The 5 second delay removed the continuous trim conditions
- Didn't sufficient train pilots on this condition i.e. intermittent automatic trim
- Pilots did not see the behavior that they were trained for in the few seconds between each nose down event
I'll agree with #2-4, but #1 was the entire point of MCAS - to reduce the response of the aircraft to pulling back on the yoke in high AOA situations. Disabling MCAS when pulling back on the yoke would thus make the aircraft less safe, and cause it to (rightly) fail certification.
How do you know it manifests as trim at all? How do you know it's not an elevator problem? But you're certain, gotcha.
>they're certainly less competent than pilot #3
The pilot who wasn't flying, because he couldn't fly, was in a position to watch the scene unfold and happened to be the only person with both the trim wheel and yoke toggle in his field of view? Are you aware most airlines with four crew for international flights require all of them in the cockpit for takeoff and landing expressly to aid in CRM? If two high hour pilots are all that's needed, why require four? Sometimes in an emergency a different perspective might be useful? But again you're certain.
>Boeing failed to consider the worst case behavior of MCAS
Not in evidence.
>Lion Air maintenance fucked up
Not in evidence.
>Lion Air pilots weren't able to diagnose the failure that morning
We don't know how far along they got. The first step in commanded pitch, roll, yaw, or power changes is reaction based on training, not diagnoses. The plane must be stabilized to buy time to conduct a diagnosis. There's circumstantial evidence of startle factor which can be a considerable inhibitor to diagnosing anything.
>Well then why was the issue so difficult for the pilots to diagnose here?
I have no certainty based on the available evidence. That's the point. What we know is, they were expressly not trained on any differences between the prior model and the one they were flying, let alone the existence of MCAS or MCAS upset. And many U.S. based pilots who fly MAX and found out years after the fact for the first time of its existence and function were pissed that it took an accident to learn about it.
Why would U.S. pilots be pissed if this "certainly" manifests in such a familiar manner that there'd be no difficulty, no delay, and no worry whatsoever, and that U.S. pilots would have easily recovered from the identical scenario?
I was under the impression that MCAS only controlled trim, and that the Ethiopian crash was found with full nose down trim. If it's an elevator problem, we're all chasing shadows.
> The pilot who wasn't flying, because he couldn't fly, was in a position to watch the scene unfold and happened to be the only person with both the trim wheel and yoke toggle in his field of view? Are you aware most airlines with four crew for international flights require all of them in the cockpit for takeoff and landing expressly to aid in CRM? If two high hour pilots are all that's needed, why require four? Sometimes in an emergency a different perspective might be useful? But again you're certain.
Yeah, that pilot. The guy who I would not be surprised to find out that he's not Lion Air. (not confident either way) He had some advantages, but either way he did what several other crews have apparently failed to. As for airlines with four crew requiring all crew in the cockpit for takeoff and landing, well why wouldn't you require all the crew to be there, even if it's an extremely marginal safety improvement? Hell, it's probably worth it just to ensure that you have all four on board at departure.
> Not in evidence. (x2)
I think if Boeing had considered MCAS driving trim to maximum they would have implemented it differently. I'd be happy to hear your opinion on that. And Lion Air maintenance seems to have failed to replace/fix the AOA sensor - not sure why you're saying they didn't fuck up. (It may have been a policy fuckup, but still in the general case of "if the sensor had been repaired this crash would not have happened)
> We don't know how far along they got. The first step in commanded pitch, roll, yaw, or power changes is reaction based on training, not diagnoses. The plane must be stabilized to buy time to conduct a diagnosis. There's circumstantial evidence of startle factor which can be a considerable inhibitor to diagnosing anything.
They had 7 minutes from reporting flight control problems to the crash.
> I have no certainty based on the available evidence. That's the point. What we know is, they were expressly not trained on any differences between the prior model and the one they were flying, let alone the existence of MCAS or MCAS upset. And many U.S. based pilots who fly MAX and found out years after the fact for the first time of its existence and function were pissed that it took an accident to learn about it.
> Why would U.S. pilots be pissed if this "certainly" manifests in such a familiar manner that there'd be no difficulty, no delay, and no worry whatsoever, and that U.S. pilots would have easily recovered from the identical scenario?
There's never any certainty in whether a pilot will successfully recover from an incident. I still think recovery is possible, and from the reactions of American and Southwest pilots I would say probable in an aircraft flown today. I do not know if it would have been probable before the Lion Air crash. I also expect US maintenance to be better. Between those two factors I'd expect a factor of 100+ safety improvement, putting accident frequency back around the normal level.
You said "it manifests" as in situ, and in situ the pilots knew nothing about MCAS. You are continuously injecting present day understanding into the past and then impugning the pilots for not having that same knowledge. It's perverse aside from being anachronistic.
>Lion Air maintenance seems to have failed to replace/fix the AOA sensor
This is more fact free commentary. Why bother? I'm not sure how you've read the KNKT preliminary report on Lion Air 610 when you make such a baseless assertion. It explicitly says it was replaced and tested per manufacturer recommendation.
>They had 7 minutes from reporting flight control problems to the crash.
And what's your point? That in your non-pilot experience it is inconceivable that a plane couldn't be stabilized in 7 minutes and the pilots never arrived at effective troubleshooting state of mind?
Maybe instead of speculating about things you know nothing about and injecting the present as a way to judge the competency of dead people, you'd be better off reading the entire 223 page Air France 447 final report? You can speculate all you want as you read how all of your preconceptions and assumptions are wrong and why.
> You said "it manifests" as in situ, and in situ the pilots knew nothing about MCAS. You are continuously injecting present day understanding into the past and then impugning the pilots for not having that same knowledge. It's perverse aside from being anachronistic.
I meant this as a "this is what the issue was" sort of way, and thought that you were suggesting that the Lion Air or Ethiopian crashes had been caused by elevator failures.
> This is more fact free commentary. Why bother? I'm not sure how you've read the KNKT preliminary report on Lion Air 610 when you make such a baseless assertion. It explicitly says it was replaced and tested per manufacturer recommendation.
It was more from some things I had remembered reading a few months back (after that report came out) suggesting that the replacement part was not a known-good-from-manufacturer part and that Lion Air had a history of improper/insufficient maintenance. If that was wrong, I apologize.
(Also, interesting report - no one ever really talks about the other sensor failures)
> And what's your point? That in your non-pilot experience it is inconceivable that a plane couldn't be stabilized in 7 minutes and the pilots never arrived at effective troubleshooting state of mind?
The fact that the crash took 7 minutes was intended as a counterfactual to your assertion that the pilots did not have time to diagnose anything. Since my impression of CRM in an incident was "one person flies the plane, one person diagnoses the problem, communicates with ATC, and checks procedures". Again, if this is wrong, I apologize.
> Maybe instead of speculating about things you know nothing about and injecting the present as a way to judge the competency of dead people, you'd be better off reading the entire 223 page Air France 447 final report? You can speculate all you want as you read how all of your preconceptions and assumptions are wrong and why.
This is rather insulting, but ok. I'll admit I've never read that entire report, only summaries. Might be worth a night at some point. But I'm not sure that that situation could be entirely blamed on the manufacturer either... Yes crews fail, yes that has to be prepared for by the manufacturer. At the same time, there's a reason that accident investigations attempt to avoid assigning blame and instead try to identify improvements - and at least part of that is that accidents happen when all sorts of issues happen at the same time.
I'll concede they should have communicated better at the hand off, but let's not pretend this was a textbook case of trim runaway. They were "behind the plane" since MCAS wasn't documented in any training material.