DCA A321 FO

Agreed.....

JamesNoBrakes

The whole "stick and rudder skillz" is kind of a cop out and at best blown way out of proportion by pilots who think they are god's gift to aviation and everything in the sky...birds too.

What does it mean, they are able to feel just the right amount of rudder needed by not looking at anything, ever? They are able to perform precise lazy 8s in a 757? Maybe they have to be aerobatic pilots? Or maybe even aerobatic pilots don't always have good "stick and rudder"? The relationship of the controls and pilot's ability to interact with these is always important, but even in some of the supposedly best "stick and rudder skillz" pilots I've still found some pretty large perceptional disconnects. The best ones probably were the ones that knew there limitations and didn't exceed them, worked to get better, even if they weren't the best "stick and rudder" pilots.

Crazy Canuck

I don't think it's about flying like Maverick or Iceman, or even thinking you can. The simple matter of truth is that a lot of pilots get very complacent due to the use of automation. I know personally of a lot of people that are at a point in their career where they are so dependent on automation they are no longer able to fly a visual approach without feeling an uncomfortable amount of pressure. Never-mind flying a full out STAR.

A good point was brought up earlier, about being proficient in both hand flying, and programming and monitoring automation. A lot of the former seems to be missing in this industry as career progression happens due in not really any part to a pilot's competence. It happens as a result of industry trends that are pushing for increased fuel savings and efficiency ($$$$). Also minimal amounts of training, as previously mentioned.

I have never used A/P in my life (save once for a very short period of time) because the company I fly for does not put them in our planes. I'm by no means the best hand flyer out there, but I am at least competent enough that I know I can do it to any airport, familiar or unfamiliar, and in some very challenging conditions too.
That being said, if you stuck me in a 737, or even put some kind of A/P (even a very basic 3-axis + nav version) and asked me to fly it from point A-point B, I imagine I would get a bit flustered with the A/P and just try to hand fly the darn thing instead. Is this a good thing? Maybe not, but we'll see what happens when I actually get to try it one day...

jungle

Just a guess, but you don't visit RVSM airspace very often do you?


This accident was not so much stick and rudder vs automation.

My guess is a plane full of tired pilots, one not yet familiar with the particular subtle qualities of new automation in a new[to him] aircraft and the lack of attention to airspeed. It can happen to anyone.

Most of us will catch it early enough, but despite a great deal of excess thrust if allowed to progress too far and too low, recovery
can take more time and thrust than you have.

A basic mistake not helped by CRM in this case and possibly hurt by not understanding or expecting the automation to correct it without input on his part.

The past decades show without any doubt that automation has saved far more people than it has hurt, but there are glaring examples that point to the need to back it up with a good scan and always understand what the automation can and cannot do.

Gaspaccio

How did the aircraft balloon 200 ft. ? I read they let their speed decay 30 kt below approach speed. Then tried to recover by adding power. When the tail struck, engines were on their way back up, just over 50%,
too little too late. And that was after a 17 mile final (glide-slope out of service). Non flying pilot asked for go-around few seconds before initial crash. Apologies if I don't have my stories straight.

e6bpilot

CRM doesn't cause accidents, it only prevents them. It is fairly clear there was a catastrophic breakdown here. Stand by while the investigation is being completed and the schoolhouse CRM instructors finally get good material to update their presentations from the 1970's and 80's scenarios they currently use (at least on the mil side).

I wouldn't say it appears that lack of stick and rudder skills led to the crash but instead an over reliance on automation, a lack of understanding on what mode was selected wrt thrust, and a complete lack of a basic scan on short final of a day, VMC, visual approach. The "stick and rudder" mistakes that were made weren't advanced stuff...I learned the importance of maintaining proper airspeed on final on my third flight in flight school which happens to coincide with about 3 hours of flight time and my first time in the landing pattern.

Say what you will about CRM, but I would put money down that if you had any American crew with even 1/4th that experience level, one of the three would have spoken up or intervened before it got that out of hand. It is just the way we are brought up.

atpcliff

Ideally, you have a PF and a PM, and no one makes mistakes. But, this is Earth, not Ideally...so we have stick shakers, aural stall warnings, and all kinds of other stuff to help us out.

Is there a practical reason not to have an aural stall warning that goes off 10 kts below target? I envision a switch to cancel it in such circumstances as turbulence, when your speed is continuously varying from 10+ kts over to 10+ kts under target speed.

Timbo

Has anyone (at NTSB or elsewhere) said exactly what they had loaded into the FMS yet?

I haven't been into SFO in about 7 years but on every visual approach, we always have something loaded in the FMS for reference. If there's an ILS, we'll have that, if no ILS, we'll have the RNAV-GPS approach, if there is one, or perhaps if there is nothing for that runway, then most guys will load the runway/elevation/speed, with another point about 5 miles out with an altitude and speed loaded on the Legs Page, so we have something to look at on the ND, for vertical guidance.

We have heard that the Asians like to use all the automation, all the time, but does SFO have a RNAV-GPS approach to 28L? Did they have that loaded, and were they in LNAV/VNAV until autopilot disconnect at 1,600'?

DeadHead

With the recent runway construction there it appears as though the threshold was displaced/relocated approximately 500'-750' down the runway. I wonder if the FMS data that they may have used (if that is) was calculating to the older glide path. If my math is correct that would bring then in approximately 1200'-1500' low. I get how that could possibly throw then off at the last minute trying to stretch the glide path to make the new aiming point.

atpcliff

If you are doing an RNAV STAR, and even more so on a SID, you need to have the automation flying the plane to avoid noise and/or track violations in much of the world's airspace.