OldAeroGuy wrote:Here is why i believe you're wrong.
Training and adherence to check lists/procedures is required for every phase of airplane flight.
There are no evolved human traits that teach a pilot to rotate an airplane at the proper rate at VR or flare an airplane at the proper height for landing. It all comes from training.
Likewise, there are many failure conditions that a pilot must recognize and take the appropriate action. Most are single failures and many are memory items. Failures to take the correct action can result in hull loss and loss of life.
- Engine failure below V1.
- Auto speed brake failure on either a refused wet takeoff or a contaminated runway landing
- Rapid decompression at cruise altitude
- Cargo compartment fire
- Flight with unreliable airspeed
- Runaway stabilizer trim
Just as you depend on crew training to successfully deal with all these failures, why would you feel that a properly trained crew would not be able to deal with an MCAS operating failure, particularly since MCAS.v2 will both decrease the likelihood of such a failure and limit possible MCAS failure mode action?
This applies to kalvados post as well.
Well, this an engineering psychology question. Humans make mistakes, that is their nature. If an engineer has to rely on trained human in emergency - they already lost most of situation control, it is now a lottery.
It is a very good practice to give people a chance to think and do things again if they fail after first try. If failure is not an option - drill, drill, drill. And it has to be straightforward situation to drill, and getting into that situation is already a high risk.
Engine out at V1, as far as I understand, is practiced often enough, and it is about very few symptoms, minimum diagnostics. How many rejected takeoffs didn't end up well?
Same with rapid decompression - once alarm sounds, better safe than sorry, you go down first diagnose later..
Problem of MCAS is that it required significant instant action (hold that damn yoke!) AND diagnostics of the issue is less than straightforward as it is a cascade failure. Specific indication is a bare minimum.
I bet even if you drill MCAS 1.0 response, success rate will be low, 90% at best. What's worse, since this high profile situation has to be drilled well enough to run on muscle memory - there will be false positives, when MCAS procedure will be called for even in non-MCAS scenario, and I am not ready to say if that would be a "better safe than sorry" case.
So what I am saying, threat of erroneous actuation has to be minimized, to at least 1e-8, with clear understanding that proper response will happen in 9 out of 10 cases if we're lucky.