The thing is, in each if these lessons, it's not the engineers who have anything to learn, but the people making business decisions. Every one of these stories has an engineer saying, "I think this is a bad idea," only to be overridden by someone making business decisions.
There is more value from these lessons being be taught in MBAs, and I don't think they are.
No amount of engineering safety can overcome unsafe business leadership.
At least in the space shuttle cases, the people overriding the concerns were themselves engineers. The idea that managers, not engineers, cause failures is simplistic.
Take, for example, the Thiokol engineer who told Roger Boisjoly's wife that if the Challenger inquiry led to Thiokol's shutdown, he would leave his kids on Boisjoly's doorstep - was that sound engineering reasoning?
I'd like to share AvE's video[0] here of a tear down of an AoA sensor. It's really quite interesting. More info on the AoA sensor's role in the Wiki entry[1].
TLDR/TLDW: AoA sensor measures "angle of attack." AoA sensors are physically over-engineered, but the electrical components may be questionable. If an AoA sensor fails, it may cause a stall. The 737 Max is equipped with two sensors, however, effectively utilized only one at a time because it cost extra for the "AOA-disagree" feature.
If I'm not mistaken, the Ethiopian case had a 'Anti-Ice warning' from the AOA sensor, why did MCAS consider the data of a sensor that potentially had an issue is anybody's guess.
> effectively utilized only one at a time because it cost extra for the "AOA-disagree" feature.
Yeah, props to Boeing who put a basic security feature as an "extra cost" option. But hey, who needs to know their AoA sensors are failing, especially when the plane is commanded by that, right?
Has anyone tried building aircraft AoA sensors other than vane type sensors?
For example, when I was designing a home weather station for myself, I wondered if the anemometer and wind vane could be replaced with ultrasonic sensors and emitters. If you take four ultrasonic emitter/ultrasonic sensor pairs and arrange them in a square, and measure the time of flight of four pulses, one from each corner emitter to the diagonally opposite sensor, you can determine the component of the wind vector along that diagonal. Combine the components from the diagonals, and you get the wind vector.
A bit of research turned up that this was not an original idea, and it does work. There are commercial anemometers and wind vanes (are they still called wind vanes when they don't have a vane?) for terrestrial use that work this way. There have also been many home made ones, such as this one [1].
Offhand, I don't see why such a thing would not work for an airplane.
Off the top of my head, there are at least two more potential ways of measuring AoA:
1. Similar to your ultrasonic lattice, but using a cone of hot wire anemometers, so that each wire is set at an offset angle to the 'zero' direction. As the AoA changes relative to the cone, the resistance on the different hot wires changes (some wires become more perpendicular to the freestream, thus cooler; others will align closer to the freestream with less opportunity to cool). By measuring the resistance differences with a bridge, it's possible to derive, angle of attack, sideslip and airspeed.
2. An elliptical body with various pressure ports around the circumference, plus a pitot port. The angle of the flow can be measured by the difference in pressure between the circumference ports (and dividing by the pitot pressure).
[1] is a description of an available device using two pitots on different axes, together with a static-pressure reference:
"The AOA tube is constructed with two tiny machined holes to create differential pressure sources. One hole is bore-sighted at the front of the AOA tube along the longitudinal axis, while the second, located at the bottom of the AOA probe, is pointed downward at a 45-degree angle to act as a reference source.
"Garmin’s normalized AOA includes an additional static-air-pressure source to allow the air-data computer to calculate airspeed and air density, important in calculating precise AOA. The addition of the static-pressure reference input makes the AOA indication accurate regardless of the airplane’s weight, speed or configuration."
The conclusion, "nobody did anything wrong, in a strictly technical sense" is too generous to Boeing and the FAA, given that MCAS did not have the redundancy that its classification as 'hazardous' required.
It would probably take either an inquiry like those into the space shuttle accidents, or a whistleblower, to know whether anyone was raising concerns. I can believe it is possible that everyone was convinced by the "in the worst case, it is no worse than ordinary trim runaway" argument (which does not excuse the failure to meet the redundancy requirement), but then the Ethiopian Airlines crash, and its crew's inability to manually re-trim, might raise the question of whether trim runaway on any 737 model is more of a risk than had been supposed (to be fair, trim runaway has not caused 737 crashes in the past, AFAIK.)
I think that it was criminally negligent by Boeing to not pull the plane after the Lion Air crash.
They were the only ones, who really knew the inside details and they were the only ones to know how to (possibly) fix it before a lot more people had to day.
Instead of being completely fortright and transparent they published an apparently completely worthless checklist. And all in the name of greed and corporate profit.
That they still smear the pilots and throw them under the bus despite all the facts that came to light leaves me speechless.
Really? My takeaway from the article was to never, ever get into a 737 MAX plane, regardless of what hack they manage get approved. And that their quality control has gone completely secondary to profits.
I don't see why Boeing would pay to have this published.
The technical details seem correct. (I have seen 0.8 degrees elsewhere as the original amount of correction MCAS could apply per activation, not the 0.6 here. Doesn't matter much.)
The conclusion is most charitably described as fanciful. There are strict rules about an automated system being able to apply more "authority" than the pilot can counter, which every engineer and manager involved knew well.
There are strict rules about a single sensor failure that can make the plane unflyable, which all knew well.
This is not a case of simple negligence to check documentation. This is a case of deliberate coverup of known risk factors to dodge rules specifically intended for airframe changes of this kind.
It was the process of dodging these rules which kept the MCAS design from being reviewed competently.
I just saw this [1] (though it was published a week ago), supporting everything you are saying, particularly about a two-sensor version of MCAS being rejected because it might require simulator time. Information showing culpability is leaking out.
"Although the FAA is responsible for the safety of any airplane manufactured in the United States, it delegates much of the certification to the manufacturers themselves.
It has to in order to get anything certified at all, says Jon Ostrower, editor-in-chief of The Air Current and a former aviation reporter for The Wall Street Journal. Boeing already has the people and the expertise, it pays better, and it isn’t susceptible to government shutdowns. The FAA, meanwhile, says it would need 10,000 more employees and an additional $1.8 billion of taxpayer money each year to bring certification entirely in-house."
I wonder if therein lies the solution. Allocation of some of the budget there instead of military perhaps? Not to mention the policy paralysis leading up to stupid government shutdowns putting people at risk.
Mandatory spending: $2.841 trillion (this includes the following)
Social Security: $1.102 trillion
Medicare: $679 billion
Medicaid: $418 billion
The military budget isn't a whipping boy to just cut into every time you need money for something else. It's a priority for a lot of people in the US, whether you agree with them or not. If something like increasing the FAA budget is actually important, the metric to go by should be, "am I willing to cut social security and healthcare to pay for this?" Because that's what the government is actually spending money on, and that's where it would have to come from.
The US military budget is consistently called out because the United States spends more on it than all other countries in the entire world, combined. Those you list are less called out because the they're supposed to be directly funded by line item taxes. There's an explicit social security and medicare tax.
Serving the people does not have to cost service to the people. Serving the people does not have to be a zero sum game, when we could instead take from the excesses of violence against others to serve the people.
While the US does spend a lot, it does not spend more than "all other countries in the entire world, combined." Total reported military spending is $1.8 trillion, with the understanding that there are some countries (China) that are very likely massively under reporting their military expenditures.
Looking at the raw numbers is ingenuous as well. As a percentage of GDP or government spending, the US doesn't even break the top 15 in the world on military spending.
Also, the US spends more on healthcare and social security than any other country. Why are you giving those careful white glove treatment?
> Also, the US spends more on healthcare and social security than any other country. Why are you giving those careful white glove treatment?
The US is one of the most populous countries with any sort of widespread social welfare program. Those programs scale in cost with population size. There's no reason for military spending to also scale in the same way with population size. You aren't comparing the same things, so stop pretending you are.
You've somewhat missed the point I was trying to make. Let me try to walk you through it in in a different way:
* The original argument is that the US military spending is abnormally high, when considered as a raw dollar amount.
However:
* The US military spending is not actually abnormally high when considered as a percentage of GDP or Government budget.
* There's hypocrisy in only considering the military in raw dollar amounts, but considering health care and social security as a percentage or per capita expense.
* If you also consider social security and healthcare by their raw dollar amounts, you also end up with absurd levels of spending.
* You should be pinning spending for health care, social security, and military to the same per capita or GDP metrics, and only then be objecting to them if they're abnormally high.
* It's unethical to pick and choose what spending you personally object to, and then cherry-pick the method you judge that specific spending by to support your personal decision.
You can't just choose to cut only the parts of the US budget you personally dislike. That's not how a budgeting in a democracy works. If you want a budget increase in one area, then that money has to come from all other parts of the budget. If 60% of the government spending is on social security and healthcare, then that suggests 60% of the spending in increasing the FAA budget is going to be coming from social security and health care.
Why not charge those that want to certify something? While this may create bad incentives ("make'em fail so we get a second fee"), I have no doubt this can be designed around with some thought.
I think the problem is what to do with these people if there is nothing to certify. Or how to ramp up with when suddenly there is a lot to certify. Companies know what they are planning so they adjust verification headcount early. A government agency won't have that information so capacity planning will be very difficult.
I think the punishment to Boeing should be that HQ has to move back to the industrial park in Everett from Chicago. And the C-suite executives must work from offices inside the main assembly building for 20 years. Force them to reconnect with the products they make every day.
They should force Boeing to have actual engineers on the top management, including the entire board of directors and everyone with a "chief something executive" title. Business majors and MBAs with no engineering backgrounds have no place in the top management of an aircraft manufacturer.
What's the Vegas line on when the plane will be flying again? Boeing says their software update is almost ready. However, I find it difficult to believe that it should fly until after investigations of the 2 crashes are complete and the results from them are used to re-examine the design. That would take a significant amount of time.
Then there's the political dimension. Let's say there are Congressional hearings ongoing that are ripping apart the FAA to figure out if they delegated too much and why they missed stuff. Who would want to fly on a plane undergoing that kind of scrutiny?
There's also the possibility that the FAA would unground the plane but other countries do not. Usually, much of the rest of the world just trusts the FAA, this is likely to not be the case here. Even if it was ungrounded in the US, who would want to fly on a plane if its not approved to fly in Europe?
I don’t think so.
2 total losses over 350 aircrafts in 5 month.
In a year, rounding per defect, you have 1% of possibility to lose a 737 max.
I don’t know the exact numbers but I seriously doubt that 1% of all the people that drive on a car are dead every year.
Garrison is too cute by far to wade directly into the controversy--he knows that he will be crucified by all of you if he does. But he is a truly great writer who has forgotten more about aviation than almost anyone else alive will ever know.
Garrison is very, very good at obliquely expressing the thing that you all won't even consider without stopping by Home Depot for a tiki torch and a pitchfork.
The problem with the 737 Max, put quite simply, is us. Our attention spans have been whittled away by Twitter and Netflix, but they weren't that great to start with. We have learned to love our post-facts world; whatever feels "right" must be true, the technical details be damned. Real expertise, gained through years of experience and learning, is devalued. Throw a tall man in a good-looking uniform in the cockpit, give us decently low ticket prices, and everything will be alright.
The problem was envisioned before the advent of technical automation in aviation: humans are bad at monitoring machines. It has only gotten worse as machines have gotten better at flying planes: manual flying skills have degraded, in some cases to the point of non-existence. No technical fix from Boeing can ever solve this problem.
In order to solve it, we have to accept that something in a technical system will always break, and humans must be competent to take over when it does. Boeing could build their entire plane out of AOA sensors, and still an automated trim system could run away. There is simply no alternative, at this point, to having a human competent in manual flying skills ready to take over.
In order to solve it, we have to admit that we have been conned by a widespread conspiracy to exploit deflationary forces to keep us happy and sort-of-productive while enriching the exploiters and increasing the disparity of wealth. They make things cheaper while wages stagnate, and we cheer. Right up until a 98-hour "pilot" drills 300 of us into the ground at 25Gs because he has absolutely no idea how to manually fly a plane.
Hard things are hard. They require time and experience to master. Facts are real and objective.
The Internet has enabled every ignorant asshole with an opinion and zero actual information to shout as loud as everyone else, and it has forced journalists to churn out sensational crap that panders to the ego of the crowd.
Our system is bad. We are bad. The warnings are everywhere and we're ignoring them to watch Game of Thrones and drive our stupid cars to our pointless jobs that only make things worse.
Saint Exupery's long dead. He's not going to leap into action and save us, we have to do it ourselves.
46 comments
[ 2.7 ms ] story [ 98.1 ms ] threadThere is more value from these lessons being be taught in MBAs, and I don't think they are.
No amount of engineering safety can overcome unsafe business leadership.
Take, for example, the Thiokol engineer who told Roger Boisjoly's wife that if the Challenger inquiry led to Thiokol's shutdown, he would leave his kids on Boisjoly's doorstep - was that sound engineering reasoning?
You also forgot Galloping Girdie. The Tacoma Narrows bridge.
That and the rudder reversal problems Boeing had back in the 90's.
[0] https://www.youtube.com/watch?v=NhZ0D-JRtz0
[1] https://en.wikipedia.org/wiki/Boeing_737_MAX#Maneuvering_Cha...
TLDR/TLDW: AoA sensor measures "angle of attack." AoA sensors are physically over-engineered, but the electrical components may be questionable. If an AoA sensor fails, it may cause a stall. The 737 Max is equipped with two sensors, however, effectively utilized only one at a time because it cost extra for the "AOA-disagree" feature.
> effectively utilized only one at a time because it cost extra for the "AOA-disagree" feature.
Yeah, props to Boeing who put a basic security feature as an "extra cost" option. But hey, who needs to know their AoA sensors are failing, especially when the plane is commanded by that, right?
For example, when I was designing a home weather station for myself, I wondered if the anemometer and wind vane could be replaced with ultrasonic sensors and emitters. If you take four ultrasonic emitter/ultrasonic sensor pairs and arrange them in a square, and measure the time of flight of four pulses, one from each corner emitter to the diagonally opposite sensor, you can determine the component of the wind vector along that diagonal. Combine the components from the diagonals, and you get the wind vector.
A bit of research turned up that this was not an original idea, and it does work. There are commercial anemometers and wind vanes (are they still called wind vanes when they don't have a vane?) for terrestrial use that work this way. There have also been many home made ones, such as this one [1].
Offhand, I don't see why such a thing would not work for an airplane.
[1] http://embedded-lab.com/blog/making-an-ultrasonic-anemometer...
Its possible failure modes, too, seem a lot more complicated than for a physical vane.
"The AOA tube is constructed with two tiny machined holes to create differential pressure sources. One hole is bore-sighted at the front of the AOA tube along the longitudinal axis, while the second, located at the bottom of the AOA probe, is pointed downward at a 45-degree angle to act as a reference source.
"Garmin’s normalized AOA includes an additional static-air-pressure source to allow the air-data computer to calculate airspeed and air density, important in calculating precise AOA. The addition of the static-pressure reference input makes the AOA indication accurate regardless of the airplane’s weight, speed or configuration."
[1] https://www.flyingmag.com/how-it-works-angle-attack-indicato...
It would probably take either an inquiry like those into the space shuttle accidents, or a whistleblower, to know whether anyone was raising concerns. I can believe it is possible that everyone was convinced by the "in the worst case, it is no worse than ordinary trim runaway" argument (which does not excuse the failure to meet the redundancy requirement), but then the Ethiopian Airlines crash, and its crew's inability to manually re-trim, might raise the question of whether trim runaway on any 737 model is more of a risk than had been supposed (to be fair, trim runaway has not caused 737 crashes in the past, AFAIK.)
They were the only ones, who really knew the inside details and they were the only ones to know how to (possibly) fix it before a lot more people had to day.
Instead of being completely fortright and transparent they published an apparently completely worthless checklist. And all in the name of greed and corporate profit.
That they still smear the pilots and throw them under the bus despite all the facts that came to light leaves me speechless.
I don't see why Boeing would pay to have this published.
The conclusion is most charitably described as fanciful. There are strict rules about an automated system being able to apply more "authority" than the pilot can counter, which every engineer and manager involved knew well.
There are strict rules about a single sensor failure that can make the plane unflyable, which all knew well.
This is not a case of simple negligence to check documentation. This is a case of deliberate coverup of known risk factors to dodge rules specifically intended for airframe changes of this kind.
It was the process of dodging these rules which kept the MCAS design from being reviewed competently.
[1] https://www.seattletimes.com/business/boeing-aerospace/a-lac...
It has to in order to get anything certified at all, says Jon Ostrower, editor-in-chief of The Air Current and a former aviation reporter for The Wall Street Journal. Boeing already has the people and the expertise, it pays better, and it isn’t susceptible to government shutdowns. The FAA, meanwhile, says it would need 10,000 more employees and an additional $1.8 billion of taxpayer money each year to bring certification entirely in-house."
I wonder if therein lies the solution. Allocation of some of the budget there instead of military perhaps? Not to mention the policy paralysis leading up to stupid government shutdowns putting people at risk.
Department of Defense : $574 billion
Department of Veteran Affairs :$78.9 billion
Department of Education : $68.2 billion
Department of Health and Human Services :$65.1 billion
Environmental Protection Agency: $5.7 billion
Does this not seem off to you?
Mandatory spending: $2.841 trillion (this includes the following)
Social Security: $1.102 trillion
Medicare: $679 billion
Medicaid: $418 billion
The military budget isn't a whipping boy to just cut into every time you need money for something else. It's a priority for a lot of people in the US, whether you agree with them or not. If something like increasing the FAA budget is actually important, the metric to go by should be, "am I willing to cut social security and healthcare to pay for this?" Because that's what the government is actually spending money on, and that's where it would have to come from.
Serving the people does not have to cost service to the people. Serving the people does not have to be a zero sum game, when we could instead take from the excesses of violence against others to serve the people.
Looking at the raw numbers is ingenuous as well. As a percentage of GDP or government spending, the US doesn't even break the top 15 in the world on military spending.
Also, the US spends more on healthcare and social security than any other country. Why are you giving those careful white glove treatment?
The US is one of the most populous countries with any sort of widespread social welfare program. Those programs scale in cost with population size. There's no reason for military spending to also scale in the same way with population size. You aren't comparing the same things, so stop pretending you are.
* The original argument is that the US military spending is abnormally high, when considered as a raw dollar amount.
However:
* The US military spending is not actually abnormally high when considered as a percentage of GDP or Government budget.
* There's hypocrisy in only considering the military in raw dollar amounts, but considering health care and social security as a percentage or per capita expense.
* If you also consider social security and healthcare by their raw dollar amounts, you also end up with absurd levels of spending.
* You should be pinning spending for health care, social security, and military to the same per capita or GDP metrics, and only then be objecting to them if they're abnormally high.
* It's unethical to pick and choose what spending you personally object to, and then cherry-pick the method you judge that specific spending by to support your personal decision.
It would also make it easier for new competitors to enter the market since they would need to build/create one less structure.
And it could make flying safer since the regulator would see that Company A does X, and Company B should too.
Safe practices shouldn’t be a company secret, they should be viral.
Then there's the political dimension. Let's say there are Congressional hearings ongoing that are ripping apart the FAA to figure out if they delegated too much and why they missed stuff. Who would want to fly on a plane undergoing that kind of scrutiny?
There's also the possibility that the FAA would unground the plane but other countries do not. Usually, much of the rest of the world just trusts the FAA, this is likely to not be the case here. Even if it was ungrounded in the US, who would want to fly on a plane if its not approved to fly in Europe?
Garrison is too cute by far to wade directly into the controversy--he knows that he will be crucified by all of you if he does. But he is a truly great writer who has forgotten more about aviation than almost anyone else alive will ever know.
Garrison is very, very good at obliquely expressing the thing that you all won't even consider without stopping by Home Depot for a tiki torch and a pitchfork.
The problem with the 737 Max, put quite simply, is us. Our attention spans have been whittled away by Twitter and Netflix, but they weren't that great to start with. We have learned to love our post-facts world; whatever feels "right" must be true, the technical details be damned. Real expertise, gained through years of experience and learning, is devalued. Throw a tall man in a good-looking uniform in the cockpit, give us decently low ticket prices, and everything will be alright.
The problem was envisioned before the advent of technical automation in aviation: humans are bad at monitoring machines. It has only gotten worse as machines have gotten better at flying planes: manual flying skills have degraded, in some cases to the point of non-existence. No technical fix from Boeing can ever solve this problem.
In order to solve it, we have to accept that something in a technical system will always break, and humans must be competent to take over when it does. Boeing could build their entire plane out of AOA sensors, and still an automated trim system could run away. There is simply no alternative, at this point, to having a human competent in manual flying skills ready to take over.
In order to solve it, we have to admit that we have been conned by a widespread conspiracy to exploit deflationary forces to keep us happy and sort-of-productive while enriching the exploiters and increasing the disparity of wealth. They make things cheaper while wages stagnate, and we cheer. Right up until a 98-hour "pilot" drills 300 of us into the ground at 25Gs because he has absolutely no idea how to manually fly a plane.
Hard things are hard. They require time and experience to master. Facts are real and objective.
The Internet has enabled every ignorant asshole with an opinion and zero actual information to shout as loud as everyone else, and it has forced journalists to churn out sensational crap that panders to the ego of the crowd.
Our system is bad. We are bad. The warnings are everywhere and we're ignoring them to watch Game of Thrones and drive our stupid cars to our pointless jobs that only make things worse.
Saint Exupery's long dead. He's not going to leap into action and save us, we have to do it ourselves.