3 ms·
Interesting you use the military as a reference point here for responding in critical situations. The military trains exhaustively so responses to critical sit
by linuxftw 7y ago
Interesting you use the military as a reference point here for responding in critical situations. The military trains exhaustively so responses to critical situations are reflexive. Fortunately, your father was a pilot, and not a grunt. When you're told to storm the beach, you storm the beach, doesn't matter if it means certain death for you, it's not about you, it's about taking the beach.
The pilots could have also had this reflex training on the effected systems, but Boeing and the regulators decided not to bother with training, and the training they did have was inadequate because Boeing's system was fatally flawed.
> They didn't need to manually trim it:
This procedure you outlined goes directly in the face of Boeing's guidance to 'follow the procedure for runaway trim'. Since the system wasn't documented at all, I assume people would trust that the procedure should work. It's not until the plane is diving into the ground due to uncorrectable trim did they discover the procedure was going to get them killed.
- WalterBright 7y agoThe EA pilots followed the procedure, and when it didn't work they went off procedure (i.e. they turned the stab trim system back on) which led to the final plunge. If they had used the trim switches after turning the system back on, it would have corrected the nose down. It seems the natural thing to do. Why didn't they? I do not know, and none of the accounts I've read addressed this. I suspect that will be a crucial point in the NTSB's report. Your assessment about pilots not being ordered to their deaths is incorrect. My father served as a navigator on B-17s in WW2. He was ordered to fly 30 missions. At the time, the 30 mission survival rate was 20%, as bad as any beach assault. The aircrews knew this, and it put a terrible strain on them. But they went anyway. My father told me he was convinced he was going to die, made peace with it, and simply decided to do the best job he could before he bought it. The Luftwaffe pilots had it even worse. There was no 30 mission maximum. It was fly until you die. Not many survived.
- linuxftw 7y ago> Why didn't they? I do not know, and none of the accounts I've read addressed this. I've wondered the same thing. My assumption is that the MCAS disables the input of the trim switches used by the pilots when it's active. We know that MCAS resets itself after certain conditions and 'forgets' what it did previously to override it's intended maximum authority. It's also be stated that the MCAS is 'reset' when the pilot uses the trim switches, but I would think that this would have been abundantly apparent as well: press switch, nose stops going down, unless they just did not notice the trim being activated in the first place (which based on other people's accounts, that would be hard to miss due to a giant wheel). > Your assessment about pilots not being ordered to their deaths is incorrect. My father served as a navigator on B-17s in WW2. He was ordered to fly 30 missions. At the time, the 30 mission survival rate was 20%, as bad as any beach assault. Fair enough, that does sound pretty brutal. My point was more about the reflexive training to overcome fear to function. It might be impossible to think critically in such a moment for most people, thus the goal is to put people on 'autopilot' when they encounter a specific scenario. But, WW2 was USAAF, USAF didn't exist yet, so no way for me to get that context ;)
- salawat 7y ago> Why didn't they? I do not know, and none of the accounts I've read addressed this. I can give you a good hunch as to why. Because the absolute last thing a pilot wants to do with a plane load of people is experiment. That's what a simulator is for. I stumbled across the "suppress MCAS by trim switch spamming" within the first few seconds after I read a complete description of the system's functionality. It took a complete description of the problem, and a background in programming/systems design, engineering, and an enthusiast+ level of aviation familiarity to come to that course of action in that amount of time, under ideal information processing conditions, with the initial information uptake immediately prior to the advent of reasoning. Airline pilots had how to fly, and no information initially that there were dragons there, and even once they knew, they had no way to try to build new situational reflexes due to the lack of availability of accurate MAX simulators, because Boeing fought tooth and nail to get the regulator to agree that they weren't necessary. They had to manually fault inject on an NG simulator to even get an understanding of the conditions involved. Military pilot training, from my understamding, bypasses the simulator, and dumps you into the seat of some of the most disagreeable aircraft ever designed, with an instructor who sabotages you regularly, leaving you in a spin, and you have 30 seconds to recover please. Begin. In that case, it's you, and the a-hole who just put you in a flat spin. Experiment away. Not 100+ innocent civilians depending on your ability to reason through what this damn plane is doing. Point being, information about the airframe was withheld due to an absolutely crushing need by Boeing for the aircraft to not require even the smallest degree of retraining. The pilots were absolutely not at fault for not being able or willing to experiment at the drop of a hat with a system they suddenly realized they had no way to understand. Putting them in that position in the first place is undeniably Boeing's fault.
- WalterBright 7y ago> It might be impossible to think critically in such a moment for most people Indeed. Pilot training should attempt to wash out those people. There's an audiotape of a test pilot who's airplane went into an uncontrollable spin at high altitude. He was in conversation with the ground, and methodically tried one thing after another, all the way down, voice even and calm to the end. It's astonishing.