Toe original AF447 thread is closed so...

[Evan]

https://images.app.goo.gl/Lui8FyubDubBrK7H6

Make that a donkey and I agree.

[Gabriel]

Correct me if I"m wrong but I remember hearing that Air Bus didn't require stall training during initial training. That was subsquently changed after this accident.

I don't know for a fact but I cannot believe that that can be possibly correct.

[Gabriel]

If they had the "FLY THE AIRPLANE FIRST BEFORE YOU EVEN START THINKNG ABOUT THE SPECIFIC PROCEDURE" principle installed in their heads, I think things would have gone very differently and we would not be discussing this anymore than we are discussing the many other times procedures were faithfully followed.

Fixed.

[Gabriel]

In ANY emergency or abnormal situation, the 1st memory item and #1 priority ALWAYS is FLY THE PLANE. And I am going to go full black and white here and point out that PERIOD at the end of the sentence. THEN you can start working on the diagnosis and corresponding troubleshooting.

That is exactly what the procedure calls for. It's the THEN that the procedure is there for.

Are you kidding me?
The UAS procedure says that you have to keep the plane stable before identifying the UAS condition and executing the UAS procedure?

This is the not-written part of any procedure:

Autopilot self disconnects
"I have controls, plane stable, ok Scotty, what do we have?"
"Unreliable airspeed"
"Ok, unreliable speed procedure then, let's go"

You are not to go to line 2 and 3 before executing line 1. Capice?

[3WE]

In ANY emergency or abnormal situation, the 1st memory item and #1 priority ALWAYS is FLY THE PLANE. And I am going to go full black and white here and point out that PERIOD at the end of the sentence.

When you have never ridden a bicycle or ventured out of a padded, sterile bubble, you think that the answer lies in uttering cryptic acronyms. In that case, Gabriel’s comments and mine and especially ATLs comments are absurd.

[Evan]

Fixed.

They did, obviously.

[Evan]

Correct me if I"m wrong but I remember hearing that Air Bus didn't require stall training during initial training. That was subsquently changed after this accident.

A friend of mine, my assistant chief pilot in the DC-8, worked at an operator where he flew the 737 and later the 757. He then went to work, where he retired, flying the Air Bus 320. He did do stalls in the Bus but I don't remember what time frame. I sent him that video and he said it was very accurate.

Or course there was stall training. What changed was the procedure. When the accident occurred, it was assumed that stall would only occur at low speeds and low altitudes where trading altitude for airspeed wasn’t a great idea (although nothing is worse than stall, except perhaps a granite wall). After the accident, the procedures were revised to emphasize reduction of AoA over gaining airspeed. If stall occurs at critical altitudes, well, ask Gabriel about that.

[Gabriel]

They did, obviously.

They did what? (I fear the answer)

[TeeVee]

Or course there was stall training. What changed was the procedure. When the accident occurred, it was assumed that stall would only occur at low speeds and low altitudes where trading altitude for airspeed wasn’t a great idea (although nothing is worse than stall, except perhaps a granite wall). After the accident, the procedures were revised to emphasize reduction of AoA over gaining airspeed. If stall occurs at critical altitudes, well, ask Gabriel about that.

eh, according to mentour, the only stall training these guys had was in the type rating early on, and as you said, at low altitude. afterwards they apparently had none. my guess is because AF put all its faith in the "sorry human, you can't stall me" airbus.

[Evan]

eh, according to mentour, the only stall training these guys had was in the type rating early on, and as you said, at low altitude. afterwards they apparently had none. my guess is because AF put all its faith in the "sorry human, you can't stall me" airbus.

The A320 and A330 share a common cockpit and the stall procedures are the same.

[Evan]

They did what? (I fear the answer)

PF: “I have the control. Plane stabl… whoa, plane not stable! Correcting for roll and descent. Whoops, too much left roll… too much right roll… too much left (damn this thing handles very strange up here).”

PM: “What the devil is going on… hmmm… what should we do about it—Hey, you’re climbing! Don’t do that!”

PF: (whoops, I am climbing) "OK, I’m going back down (reduces pitch but not enough) Now, where was I with this ECAM thing…”

etc.

Zero CRM. Zero task-sharing procedure. Tunnelling down the wrong tunnels.

[Gabriel]

PF: “I have the control. Plane stabl… whoa, plane not stable! Correcting for roll and descent. Whoops, too much left roll… too much right roll… too much left (damn this thing handles very strange up here).”

PM: “What the devil is going on… hmmm… what should we do about it—Hey, you’re climbing! Don’t do that!”

PF: (whoops, I am climbing) "OK, I’m going back down (reduces pitch but not enough) Now, where was I with this ECAM thing…”

etc.

Zero CRM. Zero task-sharing procedure. Tunnelling down the wrong tunnels.

Exactly. That is NOT flying the plane. I mean, even just for the pilot flying (leaving CRM at a side for a second)... Focusing on ONLY 1 parameter (roll)? Not even a second super critical parameter which is presented in the same graphical cue than roll? They were NOT flying the plane. They never stabilized the plane (in fact they artificially unstabilized a perfectly stable flight). So the step ZERO of the procedure (which is common to ANY procedure) was never executed. And you are expecting that they followed the UAS memory items that start a couple of steps later?

I don't understand your expectations.

[Evan]

Exactly. That is NOT flying the plane. I mean, even just for the pilot flying (leaving CRM at a side for a second)... Focusing on ONLY 1 parameter (roll)? Not even a second super critical parameter which is presented in the same graphical cue than roll? They were NOT flying the plane. They never stabilized the plane (in fact they artificially unstabilized a perfectly stable flight). So the step ZERO of the procedure (which is common to ANY procedure) was never executed. And you are expecting that they followed the UAS memory items that start a couple of steps later?

I don't understand your expectations.

Flying the plane badly is still flying the plane, just badly. I agree that it isn't flying the plane correctly or safely. As I said above, it is flying the plane dangerously. But the god question here is: why so badly? Right?

Focusing on ONLY 1 parameter (roll)? Not even a second super critical parameter which is presented in the same graphical cue than roll? They were NOT flying the plane.

Focusing is exactly the issue... It seems apparent to me that the PF was not focusing on pitch and the PM was not focusing on monitoring flight path. Again, why not....

(in fact they artificially unstabilized a perfectly stable flight).

Not true. The autoflight was adjusting for significant turbulence. When the autoflight quit, the plane immediate went into a pronounced roll excursion. The instrumentation also indicated a departure from level flight. Both of these happened before the PF made any inputs. So no, not a perfectly stable flight.

Then the PF takes over with roll in direct law at high altitude and little of the natural damping he is practiced with at low altitude.

Now study the sidestick inputs. Note that the PF does not make a sustained full stick back input and hold it there until 02:11.41. Before then, there are varying degrees on ANU and number of very pronounced AND inputs. But more importantly, look at roll inputs. Crazy. A sustained, ongoing fight with roll excursions. Bad, inexperienced hand flying under increasing stress. Tunnelling.




I know you are familiar with the phenomenon of tunnelling. It occurs when a pilot is under great stress and considerable difficulty. One thing is focused on to the exclusion of everything else.

I think you are mistaking my attempt to understand this with an attempt to justify it. There is nothing justifiable about this. The actual first step of any procedure is not FLY THE PLANE. It is LEARN HOW TO FLY THE PLANE. In the case of a passenger transport airplane, that means LEARN CRM as much as LEARN ELEVATOR. This flight crashed because of pilots who lacked the full training to deal with this situation. I hope we can agree on that much.

[Evan]

Now play this back with the PM calling out the UAS procedure and going to pitch and power values. Disengage autothrust at CL setting. Pitch at 5deg. And monitor that. The PF is forced to stop tunnelling in roll and focus on achieving and maintaining the pitch value as well. Sure, you climb a bit but you don't get anywhere near a stall and the thing is soon over.

Procedure is there to prevent or break out of the human factors that distract from flying the plane properly. And all human pilots are vulnerable to these human factors. So learn them, memorize them and just do them.

[Gabriel]

Not true. The autoflight was adjusting for significant turbulence. When the autoflight quit, the plane immediate went into a pronounced roll excursion. The instrumentation also indicated a departure from level flight. Both of these happened before the PF made any inputs. So no, not a perfectly stable flight.

OH MY GOD 8 degrees roll and 300 ft, we are going to die!

I didn't say PERFECTLY stable. You knw very well that if all the pilots did was wind the watch and light a cigarette, they would be all alive.

And if you call fighting to keeping roll in control while wildly pulling back "bad flying", then I am going to call hitting the X in the keyboard repeatedly "bad Shakespeare", or using the gas to keep try to adjust the speed, while turning the steering wheel to the stop for no reason, "bad driving".