I work under the presumption that when the Boeing CEO admits fault that quickly and unreservedly he knows they'll cop it and either does the right thing or tries to divert the public eye from more serious shortcomings.
Either way it is a damage limitation excercise one would undertake only if there is no chance to pin it on the pilots.
Either way it is a damage limitation excercise one would undertake only if there is no chance to pin it on the pilots.
Last edited:
I was allowed to drive the (manual steering) Miata today. It's work to turn the steering wheel. If I want to increase the rate of turn, it is harder work.
We also have a cushy power-steering minivan. Turning the wheel is negligible work, and a harder turn is not harder work.
I used to drive a 1979 ThunderBird. The wheel turned easy, but reaction was s-l-o-w. "Begin turn!" "taking up slop in box" "taking up slop in arm" "taking up slop in rods" "taking up slop in joints" "flexing tire sidewalls against large mass/inertia of Burd nose"....
We had a mail-truck with quick pickup and slow steering. Turning at a driveway, you had to work the steering *before* you stomped it. Lost a mailbox or two that way.
I "can" manage all these different steering responses (mostly). But aircraft are 3-D problems and things sometimes get more complicated. If you are experienced in one plane's recent models, it is very nice if the new version "has the same feel", you don't have to re-learn muscle sense.
We could make the Miata steer more like the minivan. Indeed today's power-steer MX-5 is halfway between, and could maybe be made more minivanny. While we can't make the BlunderBurd steer like the tiny Miata, the 'Burd was really a 1966 Galaxie with 13 years of fat on it. We could now ask for some automatic anticipation to take-up steering slop and inertia and make the '79 steer more-like the '66. (But I suspect Ford's goals were to make the 'Burd *different* from the snappy-if-cheap Fairmont.)
(I assume that, by "having the airplane feel like an airplane and react to control inputs like an airplane", you don't mean making a B52 or a F104 "feel like" a Piper Cub or DC-3.)
Correct. It’s not about if the control force is exactly the same, it’s about the increase in control force required to increase the rate at which the airplane is changing. Greatly simplified, if you want to pitch up more than you are right now you have to pull harder. If the airplane doesnt require more pull (or push, whatever...) to increase the rate of change, that’s bad as the airplane wont react in a way that is logical to the pilot’s muscle memory and cognitive understanding of how an airplane is supposed to react.
The car equivalent would be strong self centering and non linear force vs steerage.
Earlier power steering cars got this all wrong with too much boost and little to none self centering.
This also translates to no road feel, the aero equivalent is stick feedback as to what the control surfaces/flight characteristic are doing.
Dan.
Earlier power steering cars got this all wrong with too much boost and little to none self centering.
This also translates to no road feel, the aero equivalent is stick feedback as to what the control surfaces/flight characteristic are doing.
Dan.
I remember reading that the day before the crash (I think it was the Lion Air one) a similar situation occurred on the same plane that would crash the next day. Only this time there was an extra pilot in the cockpit, a guy who was off-duty and just sat with his buddies. He was aware of the issue and got himself involved and successfully recovered from the situation.
So here was at least one pilot who knew what he was doing, but the regular crew wasn't, apparently.
Jan
... that was an unfortunate typo!!!
There are now two open questions:
1) The precise chain of events so it is possible to find a fix for this particular plane;
2) What process or policy made it possible for this chain of event to occur.
If and this seems to be the case it started with the AoA sensor failing why was that not designed to fail safe?
Why was only one sensor used by the system and a seemingly essential warning only a cost option?
How was this certified to be airworthy?
How come that a similar MCAS exists in another Boeing plane (the KC 46 tanker which is not based on the 737) that does seem inherently safer as it uses two sensors, is overridden by stick input and doesn't switch itself back on?
1) The precise chain of events so it is possible to find a fix for this particular plane;
2) What process or policy made it possible for this chain of event to occur.
If and this seems to be the case it started with the AoA sensor failing why was that not designed to fail safe?
Why was only one sensor used by the system and a seemingly essential warning only a cost option?
How was this certified to be airworthy?
How come that a similar MCAS exists in another Boeing plane (the KC 46 tanker which is not based on the 737) that does seem inherently safer as it uses two sensors, is overridden by stick input and doesn't switch itself back on?
The question here is why the concept of failsafes hasn't saved the aircraft? The concept is simple. It's a bit like when being told ignorance of the law is no defence. Here also it is the prime requirement. I see an ideal use of AI for evaluation. Ask it to do endless simulations to flag up the dangers.
I understand one of the greater dangers of self driving cars is when they cause a dangerous situation then insist on handing back control to the driver. This sounds like the same problem?
My ex just bought a nearly new Ford . It's 1000cc turbocharged. The lady who owned it was having a baby and needed a larger car ( ?!? ) I was dubious. The car is great and even sounds like a larger engine. Then the oil warning light comes on. The dipstick says fine, there is a known problem with the car. I had to say in no way drive it. It was just a sensor. Unfortunately next time it happens I suspect she will ignor it and that time it will be real. The Morris Minor could work at 1/4 bar oil presure. 4 bar being usual. The advice was if the oil light went out on the move all was well. My similar engine did the same. It was the sensor. There was no big end thump. Imagination thought maybe there was? A mechanic said change the sensor and then ask your questions. It was a one minute job.
I understand one of the greater dangers of self driving cars is when they cause a dangerous situation then insist on handing back control to the driver. This sounds like the same problem?
My ex just bought a nearly new Ford . It's 1000cc turbocharged. The lady who owned it was having a baby and needed a larger car ( ?!? ) I was dubious. The car is great and even sounds like a larger engine. Then the oil warning light comes on. The dipstick says fine, there is a known problem with the car. I had to say in no way drive it. It was just a sensor. Unfortunately next time it happens I suspect she will ignor it and that time it will be real. The Morris Minor could work at 1/4 bar oil presure. 4 bar being usual. The advice was if the oil light went out on the move all was well. My similar engine did the same. It was the sensor. There was no big end thump. Imagination thought maybe there was? A mechanic said change the sensor and then ask your questions. It was a one minute job.
Ok, can you help me to understand. If the weight distribution requires constant elevator pressure, then they should be redesigned to reinstate the centre (permanent trim) rather than increase the range of travel?
I'm trying to believe this, but there would have to be no negative side effects.
List of accidents and incidents involving the Boeing 737 - Wikipedia
Seems the whole 737 familly has been unlucky?
757 767 777 all the best of the best.
Seems the whole 737 familly has been unlucky?
757 767 777 all the best of the best.
Angle of attack sensors are just dumb sensors like many other sensors on an aircraft. It is the flight systems - not the sensors - that are supposed to be designed as redundant and fail-safe.
MCAS should have / could have been easily overridden with a simple flick of a switch but the point is, it wasn't. Twice. And this has seemingly cost hundreds of lives. But the question is what chain of events, what missed opportunities, led us to where we are today.
Was the system design the only point of failure? Of course not. There are always many holes in the swiss cheese that have to line up* - but a simple error checking process maybe could have prevented this - and was obviously a failure mode that Boeing had considered, seeing they were willing to charge extra for that particular 'safety feature'.
Forgetting the system software design for a moment, there are so many factors to be examined: the crew training (e-learning) - did they understand it? Did they physically do it? Did their language proficiency meet the level of the training?; the emergency checklists (which were subject to an Emergency Airworthiness Directive after the Lion Air Accident); the correct identification of the fault (the previous Lion Air 'near miss'); the experience or inexperience of the operators (apparently ~200fh total for the copilot); was there a Crew Resource Management problem?; were the faulty sensors all from the same manufacturer / batch / tech?; was the fault in the system's handling of data rather than acquisition by the sensor?;... and on and on.
As 6L6 and others have said ad-nauseum, until the final investigation report comes out, answers [1] and [2] can only be speculative because there are just so many parts of the puzzle that need to be sorted through.
* Professor James Reason's Swiss cheese model of accident causation is a good beginners reference.
Last edited:
For me this in one of the most interesting posts here. Thanks for the info. I knew there had to be 1000s of hours logged on the 737Max without MCAS causing crashes. That makes the 2 recent errors very rare. Fatal, but rare. What happened in the two fatal incidences that did not happen 41,000 other times?
Since the preliminary report we know that the Ethiopian pilots followed the correct prescribed procedure repeatedly.
The 737 MCAS switches between the two sensors before each flight but it uses the input of only one so while the fail safe option physically exists it is not used.
So in the Ethiopian case the one not used showed a steady and correct angle of 15deg when it mattered while the one that was used by MCAS on that flight varied rapidly between 12 and 70degrees.
Some suggested that the faulty sensor was hit by a foreign object like a bird but that also has been ruled out in the preliminary report.
- Status
- Not open for further replies.