I always wonder how and why witnesses are so unreliable. I know not everyone knows about aircraft, but how hard is it to fully understand and comprehend what you’re seeing? Just weird how so many people claim to see fire and smoke when they really didn’t. I know I know it would be a very intense situation to see an aircraft crash in front of you, but still.
Now I will say IF this witness is right and there was smoke, then that changes things a lot here.
We don't do it on purpose.
The reasons for witness unreliability is deeply rooted in our neurological makeup, and it affects our entire understanding of the world. Our brains seek patterns and will construct memories based on interpretation of data from our senses. In stressful situations such as seeing a crash, the bandwidth left for sense data decreases, and our brains seem to try to compensate by creating patterns from experience. In this instance, a person seeing a crash might be expecting a crashing aircraft to be on fire, and so his/her brain can very well tell him/her that there is a fire.
Interestingly, this kind of boas bias affects pilot training as well. There have been high profile crashes where pilots have shut off the wrong engine, despite instrument evidence telling them the other one had failed, because again, in stressful situations, our bandwidth seems to decrease and our brains try to fill in the blanks. This is why before we touch any guarded pushbuttons, e.g. the fire shutoffs for the engines, as pilot flying we must get confirmation from the pilot monitoring. It's not as if we can't see the big red blinky thing. It's that our brains are able to very powerfully tell us that what we are seeing is not real.
For the A320 it is integral in the FBW system and highly redundant including a fall back path on (sensor) fault.
For the 737MAX it is a stuck on wart relying on a single input, no fault check, no fall back, no documentation.
The way you talk in such glowing terms about the A320 and then in such clearly disparaging terms about the 737 leads me to think you have an inherent bias and your overall statement can’t be trusted as rational or objective.
It's not a matter of bias, it would take someone unfamiliar days to understand the sensor flow and computer processing paths, decision-making programming, error-detection and correction, as well as overall redundancy built into Airbus FBW architecture. It's probably one of the single largest investments in commercial aviation history (as a whole) and totally defines how Airbus products operate to the core. They developed this in the '80's and it's allowed all of their products to scale up seamlessly. Any type of system like MCAS is not only pointless due to speed and alpha prot programs, but any inherent instabilities or handling characteristics can be augmented in the basic FBW programming.
It's not a matter of one being better than the other, but if you're talking systems-integration and the overall "elegance" and simplicity of work-flow in the cockpit, one is absolutely better. This is exponentially more true when we're talking about updating products that are decades old.
I feel that this zeal to ensure backward compatibility for cost considerations is one of the problems. Regulators must take a call on what constitutes a new aircraft type. The A320neo PW issue is also an equally uncomfortable situation to have, frankly.
This is something that's become more and more frustrating for me lately. Regardless of what happened with this flight, it's annoying that a supplemental type certificate can be issued for an aircraft which clearly handles different enough to require an augmentation system. Likewise - if said system is so important, it's frustrating that it has no capacity to self-check its data before performing a physical action on the aircraft's controls. I'm not a conspiracy theorist in the slightest, but I've GOT to assume the FAA/EASA looked the other way on this one. I understand they claim it's just like speed trim, with two ways of disabling.. but, eh.
For this crash, I'm actually starting to wonder if it wasn't MCAS at all, but rather the reason MCAS exists (the nose-up / sketchy power-on stall characteristics). It's entirely possible that we're seeing a MAX that stalled (due to unreliable airspeed issues) in a manner that was nothing like the NG/Classic and MCAS wasn't enough and/or didn't work fast enough.
Just something to think about that I haven't heard mentioned yet. Kinda fits.
Yes this is correct, however the A320 has three sensors whereas the 737MAX has only two. If one starts to give erroneous data then the A320 has two others to fall back on whereas the 737MAX FMC has to decide which one to believe. I understand that the problem potentially here is that the system may 'fight back' on corrective pilot inputs on the 737MAX whereas the A320 will switch to 'alternate law' if it doesn't understand the data it is getting, effectively handing full control to the pilots.
Conjecture alert... Boeing, I am sure in good faith, has tried to retrofit automation within the physical limitations of an older design, but this has potentially created an Achilles Heal. That, compounded with a 'systems know best' approach to the logic is dangerous if true, in my opinion.
In summary it's a subtle but potentially significant difference in the coding of the systems.
A single pitot / ADR failure is a non-issue on an Airbus, in fact it should isolate it automatically and tell you. The plane can still perform a CATIII auto land actually. Even two failures isn't much of an issue from a systems point of view. Also such failures would NOT degrade FBW into ALT LAW, you should still retain all flight protections. In the newer 330/340 and all 380 you can lose ALL THREE pitot/ADR systems and still fly on the BUSS speed tape on PFD using extrapolated AoA data. Losing primary and stby pitot in a Boeing (minus 787) = big trouble. Losing even one can be very confusing and dangerous (ie. Birgenair 301 757 crash).
The additional (and fundamental) problem here is that, eventually (if we're talking sensor/computer failures) an Airbus (and AFAIK, every other Boeing) hands controls to you - with no actions being taken at any point. In the case of the MAX, sensor or data/computer failures can result in ACTION being taken. From an aviation systems engineering standpoint, that should never happen in my opinion - it shouldn't even be able to be certified. The issue here is that fixing it basically requires Boeing to develop an integrated computer system with redundant inputs that can check for data integrity - and poof, now look where we are.
Thank you for putting together this clear and informative post. Airbus FBW was designed from the ground up. Very different from what would seem the somewhat patchwork nature of some of the 737Max implementation.
Ok understood, but I guess what is throwing me off is what is stopping the pilots from literally just flying the aircraft? they have all the visual references in the world....its day light, its clear. Will the aircraft not respond to control column inputs?
In general, yes. The first rule is to fly the aircraft.
However, there are exceptions. In a FBW Airbus, if you're in Normal Law pulling back further will lead to a stall, the aircraft will not pitch up further.
Note that this is not a bad thing...
Not true. Fake news. Debunked up thread.
You're saying that flying for some reputedly dodgy airline is less safe than driving a car?
I think you mean Alternate law, you cannot stall like that in Normal mode.
Oops. My statement was slightly unclear, but you'll note I said "the aircraft will not pitch up further".
What I meant was that in Normal law, if pulling back further takes you beyond Alpha Max, the aircraft will inhibit further pitch up.
the captain starts flying solo and becomes an instructor
Good CRM should prevent that from happening.
CRM requires resources, what resources does a 200 hour pilot offer? Little experience beyond blind compliance and whatever was passed on in simulated emergencies. Can you give the hand flying of a damaged plane to the 200-hour pilot? Can you let him run QRH procedures without monitoring the actions taken while hand flying the plane. Would you trust a 200-hour pilot to land in CAT I mins without autopilot and the captain dead?
Granted, a 200-hour pilot doesn't have a lot of experience, but it depends on training. A number of major international carriers put 300-hour pilots in the right-hand seat, and militaries will happily send 300-hour pilots solo in a high performance jet. There's a difference between 300 hours tooling around in a bugsmasher vs airline targeted ab initio training.
Last edited by Starlionblue
on Mon Mar 11, 2019 11:27 pm, edited 3 times in total.
"There are no stupid questions, but there are a lot of inquisitive idiots." - John Ringo