Air France jet missing

[flyboy2548m]

Notice the warning cares nothing about what law or lawlessness you're in. It will keep blaring until you unstall the airplane. You're welcome, Gabito, keep on keeping on.

[Gabriel]

Noted. Thank you.

[Gabriel]

I wonder... Is that stall warning AoA different in normal law (where you have alpha max hard protection) vs alterante/direct law?

That, I think, would make sense, providing an earlier warning to an impeding stall in direct/alternate law, and avoiding stall warnings before alpha max in normal law.

I found this in an incident report (A319 actual stall in normal law) by the Civil Aviation Administration of China that may hint in that direction:

From this we can see that AoA increased rapidly, exceeded stall warning threshold (23 degrees under normal laws)

BTW, interesting incident, although the report is not very well written and at times confusing (to me).

An A319 encountered microburst/windshear on approach and stalled, apparently in normal law (although FAC1 and FAC2 came offline). The stall warning did sound.

While alpha prot, alpha floor and alpha max all activated before the actual stall, the magnitude of the loss of airspeed due to windshear and of vertical gusts due to microburst were so big that the protections (that included commanding 7 seconds of continuous full nose down elevator) could not prevent the AoA from reaching 33 degrees and indicated airspeeds to go as down as 74 kts (although that was probably inaccurate due to the excessive AoA, the actual airspeed must have gone below 89 kts, that is more than 37 kts below their normal approach speed).

According to the report, given the weight and configuration (full) the 1G stall speed should have been 99 knots, but at the time of the stall warning the load factor was 0.8 (the protection was already commanding full nose-down elevator) and the Chinese say:

the actual stall speed at that time should be 99*0.8=79.2kt.

This is wrong. Stall speed is not proportional to the load factor but to its square root (because lift = weight * load factor is proportional to speed squared). That's why in a 60-degrees-bank-turn (coordinated and constant altitude) the load factor is 2 but the stall speed is 41% above the 1G stall speed. 99*sqrt(0.8)=89.

Another interesting aspect is that the report hints that they would all be dead if it was not for the envelope protection. Not only they say plainly that the protection recovered the stall condition, but also criticize the pilots quite harshly in their decision making of attempting the approach in the first place and in their performance under the quickly deteriorating conditions and upset:

... the automatic protection system began to work, and saved the plan from the stall condition ...

When wind shear became obvious, pilots took no resolute action to stop the approach. The crew failed to respond as per manual. When the aircraft attitude was gradually deviating from norm, air speed decreased and low energy warning appeared, the pilots intervened in the AP control by choosing the speed manually instead of pushing the throttle forward as QRH2.03, leading to complex situation, including increasing attitude, low airspeed, Alpha Protection triggered and stall. The Crew Resource Management was a chaos. When aircraft was in complex conditions, pilots reacted out of their instinct, leading to dual side stick inputs for as long as 12s.

bea.aero/docspa/2010/b-54100914.en/pdf/b-54100914.en.pdf

[ocelot]

The point I think you may have overlooked is: the protection is meant to keep pilots from stalling the plane by pulling_up. It wasn't ever intended to (and can't) prevent the plane from stalling under any circumstances, as this windshear event demonstrates. Consequently, even though you may not be able to set off the stall warning by screwing around, it still needs to be active in normal law.

Whether the threshold is set differently, I have no idea. You could make a good case for having it lower in direct law. You could also make a good case for having it lower in normal law. So my guess would be that it's the same, but IDK.

I have a different question. Since part of the idea of the protection is that if you need UP you can pull_up without having to concentrate on the exact pitch... I would expect that the maximum AoA allowed by the protection is chosen accordingly. That is, in addition to not stalling it's also supposed to prevent getting onto the back side of the power curve, because that's not where you want to be if you're trying to not_hit terrain. Is that true? And if so, doesn't it interfere with flaring for landing? (And if not, isn't it making the terrain/windshear escape less effective? Or is the idea that eating that effect but making the manoeuvre easier will save more mediocre pilots than it kills skilled pilots?)

[Gabriel]

As I said, I agree it is good for the stall waring to work in normal law too. I just thought (wrongly) that Airbus decided otherwise.

The reason why I thought that was because I found references to the stall warning only associated with the normal law. There were also comments in PPRUNE back then saying that the stall warning proved that the airplane was not in normal law. I remember that user didn;t inspire a lot of confidence in me that he knew what he was talking about. Now I know he was wrong (and so was I).

When trying to find a justification of why Airbus would do something like that, I reasoned that a) to avoid the stall warning sounding before the hard AoA protection and b) because alpha max was still exceeded (for example like in this incident) the automation would apply full elevator down by itself and regardless of what the pilot does with the stick. Applying nose-dow elevator is how the pilot should respond to the tall warning anyway, so adding the warning would not change the effective elevator input.

Now, point a) can be solved by shifting the stall warning AoA to a value greater than alpha max. And point b) I hoped it was not true because, while the airplane would be able to save itself (like it did in this incident) letting the pilots know what is going on seems like a good idea.

Regarding your last question: alpha max is very close to the AoA of max lift. That is why a stall warning with normal margins over the actual stall would sound before the alpha max protection kicked in. That would be very bad because the stall warning has precedence to everything else. Even windsehar escape or "terrain terrain woop woop pull up". The reason is simple. If you are in a windshear or about to hit something, stalling can only make things worse. So in a normal law escape maneuver (thrust to TOGA and pull up as much as you want), if the stall warning sounded before alpha max and the pilot honored the stall warning as they should, alpha max would never be reached and they would leaving some performance on the table.

Regarding the power curve, it only applies for constant speed. In an escape maneuver you will not be climbing just by converting excess power into climb rate, but also you would be trading airspeed for altitude until you have no more airspeed to trade (you reached alpha max or got a stall warning) and they you would keep that airspeed converting all the excess power into climb rate. Once you reached this point, it is true that this will give you not the optimum climb gradient because you are in the back side of the power curve. But if you lower the nose to accelerate to the best climb gradient speed, your flight path will deteriorate and it will take some time until you accelerate and climb better and intersect the path of the status quo. In other words, the escape maneuvers prioritize short-term performance instead of long-term performance.

[3WE]

So I don't like how ANYONE writes/communicates things.

I THINK Mentour's video said the stall warning was going off and on as speeds went from very low to ludicrous low.

But now, flyboy says, the stall warning stays on.

Nevertheless, I STILL can't see my outsider, ass-hat, pontificator, idiot relentlessly pulling up (even though Evan has disdain for me sitting there clueless and letting the plane fly FDnH at IMPROVISED power and attitudes...)

[ocelot]

Regarding your last question: alpha max is very close to the AoA of max lift. That is why a stall warning with normal margins over the actual stall would sound before the alpha max protection kicked in. That would be very bad because the stall warning has precedence to everything else. Even windsehar escape or "terrain terrain woop woop pull up". The reason is simple. If you are in a windshear or about to hit something, stalling can only make things worse. So in a normal law escape maneuver (thrust to TOGA and pull up as much as you want), if the stall warning sounded before alpha max and the pilot honored the stall warning as they should, alpha max would never be reached and they would leaving some performance on the table.

That makes sense, although I was under the impression there was more room on the back side than that; that is, that you wouldn't get the stall warning until you're past the place you want to be for max escape.

Regarding the power curve, it only applies for constant speed. In an escape maneuver you will not be climbing just by converting excess power into climb rate, but also you would be trading airspeed for altitude until you have no more airspeed to trade (you reached alpha max or got a stall warning) and they you would keep that airspeed converting all the excess power into climb rate. Once you reached this point, it is true that this will give you not the optimum climb gradient because you are in the back side of the power curve. But if you lower the nose to accelerate to the best climb gradient speed, your flight path will deteriorate and it will take some time until you accelerate and climb better and intersect the path of the status quo. In other words, the escape maneuvers prioritize short-term performance instead of long-term performance.

That's true but... unless you're going faster than one would expect, the zoom climb part of an escape maneuver will only last a few seconds. E.g. suppose your best sustained angle of climb speed is 120 knots and you're tooling along at 150 when you break out of a cloud and there are sheep looking at you; you pull_up, but it'll be only something like ten seconds, if that, before your speed drops to 120; and you then want to keep it there. (That is, if you let yourself slow below that speed you won't be getting your best escape performance.)

I suppose the amount of time will vary some depending on how much reserve power you have, but it won't be measured in minutes.

(On a second look, though, you're talking about already being slower than best angle of climb and I was assuming faster. I guess slower is reasonable for windshear escape where usually you'd be on approach, whereas unpleasant terrain surprises probably happen at higher speeds.)

(I guess if you're already slower than best sustained angle of climb, you do better by first popping up as far as you can without actually stalling. But at that point it seems like there'd be some optimum transition to a sustainable climb, and it's not obvious what it is and maybe dependent on the airframe?)

[3WE]

[Discussion of various stall protections]

That makes sense, although I was under the impression there was more room on the back side than that; that is, that you wouldn't get the stall warning until you're past the place you want to be for max escape.

[/quote]

Just my total ass-hat opinion, but buffer discussions are kind of academic in my book...including "past the place you want to be".

I IDEALISTICALLY believe that you can recover from a lot of stalls including FULL ONES, as long as you act fairly promptly...theoretically a punch of nose over and you have healthy flight over the wings (even if you are a little slow).

With Colgan, I oftentimes think it was nasty that they leveled, flapped, geared and sped up the props...Their speed seemed to go from very healthy to stall in very few seconds...and yet there was SOME buffer.

Switch to Airbus...you want darn near maximum AOA for the terrain avoidance...and if you actually stall in a brief wind shear, I ass-ume HAL promptly reduces the nose up angle to make the AOA vane happier...

As for a human...what, 5 knots of buffer since we react a little slower?

/rambling on how much buffer is enough...I don't think we need OR GET that much.

And somehow the Colgan TIRED Colgan crew forgot that 1) Leveling, flapping, gearing and flattening the props results in a remarkable FEELING of braking and 2) generally REQUIRES a HEALTY power increase. (Simultaneous stupidity or tired/forgetfulness or some combination?).

[Gabriel]

And somehow the Colgan TIRED Colgan crew forgot that 1) Leveling, flapping, gearing and flattening the props results in a remarkable FEELING of braking and 2) generally REQUIRES a HEALTY power increase. (Simultaneous stupidity or tired/forgetfulness or some combination?).

You got this wrong. Colgan was NEVER about "oops, we slowed down too much". They were intentionally slowing down to a target speed that was below the point where the sticskhaker triggered, thinking that said target speed was of course above the stickshaker trigger speed.

The target approach speed they were aiming for was the correct one for non-icing conditions. But they selected "REF SPEEDS INCR" in the ice protection panel. This setting asumes that you are going to calculate the ref speeds using the values for icing conditions (which maintains the required margin over stall with the worst ice accretion condition defined in the FARs) and reduces the AoA that triggers the stickshaker and stickpusher accordingly. So they were aiming to the non-icing Vapp whe the plane thought they were aiming to the icing Vapp. The stickshaker triggered at a speed somewhere between the two. Because icing was not a factor, they were nowhere close to the actual stall when the stickshaker triggered. That is how they managed to slow down so much and pitch up and climb so much before actually stalling.


Seems that the image doesn't want to show embedded, try here: http://majesticsoftware.com/mjc8q400/sa ... mg_250.png

(Disclaimer: stall, stickshaker and stickpusher all depend on AoA and not directly on speed. You know the drill, add "1G" where relevant to convert AoA to speed)

[3WE]

And somehow the Colgan TIRED Colgan crew forgot that 1) Leveling, flapping, gearing and flattening the props results in a remarkable FEELING of braking and 2) generally REQUIRES a HEALTY power increase. (Simultaneous stupidity or tired/forgetfulness or some combination?).

You got this wrong. Colgan was NEVER about "oops, we slowed down too much". They were intentionally slowing down to a target speed that was below the point where the sticskhaker triggered, thinking that said target speed was of course above the stickshaker trigger speed.

I will review this. My recollection was that they did lots of slowing and leveling and the speed was dropping at a fast clip…

I acknowledge that the stall warning was early, but I question that they were competently throttle jockeying to nail a target speed and then, quickly returning power…I’d THINK they should have more gradually slowed up…

I’m pretty sure they were seconds away from “the normal” stall warning speed, and the plane did not_have auto throttles.

All this being said, I still favor 5 min of recurrent training: Are you aware of folks pulling up relentlessly and stalling perfectly good airliners? Do you remember the basics?

[Gabriel]

I will review this.

Let me help you:

https://www.youtube.com/watch?v=o6c3ENr_CRM

[3WE]

I will review this.

Let me help you:

https://www.youtube.com/watch?v=o6c3ENr_CRM

No.

That dude is more long-winded than you.

[3WE]

Here’s the efficient version:

https://youtu.be/lxywEE1kK6I

Fat & Happy, 180 kts +/-

Power back
Gear
Props to fast
Flaps 15

20 seconds of crisp slowing to 130 knots. No power increases.

It would only be a few more seconds until the no-ice target, were they really on top of their speed, or “dozing”?

Was forgetting the higher speed stick shaker a CRITICAL TURN of events, or was it not_watching the speed/major drag, zero power?

Anyway, the pull up was relentless, and the power even lagged a couple seconds and an extremely short time until the wings were doing big uncommanded drops.

And all of this happened about 50 times faster than Mentour’s analysis.

[Gabriel]

Here’s the efficient version:

https://youtu.be/lxywEE1kK6I

Fat & Happy, 180 kts +/-

Power back
Gear
Props to fast
Flaps 15

20 seconds of crisp slowing to 130 knots. No power increases.

It would only be a few more seconds until the no-ice target, were they really on top of their speed, or “dozing”?

Was forgetting the higher speed stick shaker a CRITICAL TURN of events, or was it not_watching the speed/major drag, zero power?

Anyway, the pull up was relentless, and the power even lagged a couple seconds and an extremely short time until the wings were doing big uncommanded drops.

And all of this happened about 50 times faster than Mentour’s analysis.

Lazy. Watch just under 1 minute from 15:36 to 16:33.
Here you have the link for the starting point:

https://youtu.be/o6c3ENr_CRM?t=936

Yes, they were decelerating about 2.5 kts/sec. They were aiming for a Vref of 118 kts with flaps 15, which they would have achieved some 5 seconds after the stickshaker had the deceleration rate remained the same. The captain had just asked for flaps 15 just 2 seconds before the stickshaker started so he was probably more or less aware of their current speed.

The increased falps-15 Vref for icing conditions was 138 kts, and if you ask why it had just 8 knots of margin over the stickshaker... well, it didn't. By when the stickshaker started the flaps were still in transit from 5 to 10. The stickshaker would have activated at a few knots slower had the flaps been already at 15. The timing for requesting flaps 15 was appropriate for their expected flaps-15 Vref of 118 kts.

What is undoubted, however, is that they missed the dotted red stripe coming up in the airspeed indicator. That was a clear cue that the stickshaker was about to shake even if they thought that they were going faster that stickshaker speed.

[3WE]

What is undoubted, however, is that they missed the dotted red stripe coming up in the airspeed indicator. That was a clear cue that the stickshaker was about to shake even if they thought that they were going faster that stickshaker speed.

Thanks for making my point, oh ye lover of relentless verbiage… They were doing a crappy job of monitoring/managing speed.

Also, in my numerous turboprop rides, gear, flaps leveling and flat props has pretty much always included a healthy power bump…. They were doing a crappy job of monitoring/managing speed.

…and if you are 5 seconds from v-ref, the power should be coming up…. They were doing a crappy job of monitoring/managing speed.

I find the stick shaker miscue an interesting FACTOR, but put THE blame on snoozy woozy flying (that I could have been doing if I were in their white epauletted shirts with their “great” work schedules.)

The oh-shit pull up…less likely…you know, that goes against the type-specific procedure- which is the only reason you shouldn’t aggressively pull up…

[ocelot]

That's quite the confusing panel, particularly the top right knob. I mean, sure, you're supposed to have read the directions before operating it, and I haven't, but stuff like this increases the risk of making stupid blunders when using it under stress in real situations, especially if also fatigued.

Seems that the image doesn't want to show embedded, try here: http://majesticsoftware.com/mjc8q400/sa ... mg_250.png

It's fine here...

[3WE]

***stuff like this increases the risk of making stupid blunders when using it under stress in real situations, especially if also fatigued.***

Yes! * a large number!

You also ‘forgot ‘to spew cryptic acronyms in this thread process (but several are depicted on the displays).

That’s where I see the Air France guys…their well trained minds analyzing 10 failures (and a geometrically more complicated order in which they should be addressed)…they studied and studied and studied but the “light a cigarette and Aviationate” procedure wound up on the reel to reel backup tape instead of the flash memory.

[elaw]

Hey out of curiosity, has anyone here read "Understanding Air France 447" by Bill Palmer (https://www.amazon.com/dp/0989785726?

I'm only about 40 pages into it so far but it seems like a good analysis from a pretty qualified person.

[3WE]

Hey out of curiosity, has anyone here read "Understanding Air France 447" by Bill Palmer (https://www.amazon.com/dp/0989785726?

I'm only about 40 pages into it so far but it seems like a good analysis from a pretty qualified person.

Do you think he will suggest that a relentless pull up was a bad course of action?

[elaw]

Do you think he will suggest that a relentless pull up was a bad course of action?

I'm gonna go with "yes" on that one but stay tuned...