Air France 447 - On topic only!

[guamainiac]

I wish people would stop using the "pilot bashing" lines especially in this thread.

I have only been here for a short time compared to some and most but frankly there is not a person among you that I think is honestly "pilot bashing".

Did the pilots seemingly make mistakes?

Did the system break down regarding training?

Did some person screw up? Of course.

It is possible that BEA and/or Airbus would like us to think that pilots were at the root. The job of the (French or Euro), ALPA will defend pilots and all of this positioning is to be expected.

How long are we away from the "final" report?

[TeeVee]

You see what happens when you actually read the posts TeeVee?

But I don't agree at all with the notion that pilots can be completely excused due to lack of training. Pilots need to assume individual responsibility for studying every system interaction and every procedure they might one day require, with or without prompting by training programs and regulators. The information is widely available to them. It is their job to know these things and comes with the heavy responsibility of piloting mass transportation. It goes back to what I call 'airline pilot material', something that must be present before any training begins. I fear that the industry has been populated with pilots infected with a task-oriented, bureaucratic 'need-to-know-basis' mentality, who do not take enough (if any) initiative to learn on their own to supplement their knowledge and preparedness. Therefore, I also blame these pilots to some degree.

yeah, but you don't really want to blame airbus do you? they built the damn thing, and to an extent, it failed. you appear to want to defend the technology at any cost. well, ya know? the technology failed miserably. for all its sensors, processors, redundancies etc, etc, it couldn't even tell the pilots the most basic of things, or am i getting it all wrong and three pilots with more experience than almost all the folks on this forum just didn't/couldn't read the instruments?

[guamainiac]

In a matrix of complex systems, how do you know where you are deficient until you are in the middle of becoming "deficient" or allegedly deficient?

If you could take and duplicate exactly what those individuals were seeing on the the panel or feeling through "feedback" on the stick, how can they be judged?

Can you duplicate what was being seen and yeah, there was a stall warning but what were they seeing to make them not react properly?

Remember, was it on two occasions they pushed down on the stick, it really sounds more like they were doing things so that something made sense and started reacting as if it should.

Where are all those folks who do the fancy mock up's of the cockpit?

[Black Ram]

When it comes to flying the plane, they couldn't read the instruments. It's not just about looking at a few displays with some numbers. No matter how sophisticated the bus is, the pilot is actually not eliminated. These highly-trained professionals (at least we like to think of them that way) are there to cope with a lot more serious emergencies.

[Jingogunner]

Struggling Again

I'm struggling with this discussion. The altimeter measures increases and decreases according to a pre-set standard temperature/ density for each FL? So if it reads 35000' it may not actually be that height if the air temperature is much lower than what the altimeter has been given for that height, I could be flying at 34000'. Have I got that right? Is there a belief that the aircraft might have stalled because the air temperature was higher and the density less than needed to maintain lift at that speed? And the computers are unable to warn, advise and compensate for these discrepancies?

[Gabriel]

Gabriel,

A lot of what you wrote a few posts ago is true, but I really dislike this concept of saying the crew pulled up DESPITE a stall warning.

The failure they had (loss of airspeed) is the very failure which renders the stall warning questionable. It is the one condition in which it may be entirely proper to ignore a stall/overspeed warning.

Lets take the hypothetical situation where one airspeed indicator fails to 80kts (say the F/O), and the other remains valid.

The stall warning activates.

MCM, don't understand. IT DOESN'T ACTIVATE!!!!

In the Tomahawk I used to fly, the stall warning was triggered by a small vane in the leading edge of the wing: When the stagnation point went below that vane, the vane was moved up by the airflow and that switched the stall warning on.

In the Cessna 150 and 152 I used to fly, the stall warning was a kind of harmonica that was blown by air that flowed between two holes in the leading edge of the wing, one above the other.

In just about any jet that I know about, ranging from Cessna Citation to the A330, the stall warning is triggered by a sort of weathervane, except with an horizontal axis, that protrudes from the side of the fuselage usually near the nose.

What do all these different systems have in common? Just two things: 1) They are not sensible to speed and 2) they are sensible to the angle of attack. Thus, the stall warning doesn't know or care about the speed AND is ALWAYS triggered when you exceed a critical angle of attack.

That's why when you pull up some Gs even at a speeds safely above stall (for example, Vref) the stall warning can be triggered, even when the speed remained above the stall speed during the whole sequence.

The airplane will stall at different speeds depending on the Gs that you are pulling and the airplane's weight (a plane at MTOW stalls faster than when lighter), but always at the same AoA.

That's why the stall warning will NOT sound at low AoA's even if one, more or all of the speed indicators show "below stall" airspeed, be that indication real or false.

And that's why the A330 FCOM says that the stall warning must be observed during a UAS event.

Or am I completely misunderstanding your post?

Should the crew follow the stall warning procedure?

Yes, but don't believe me. Airbus says so.

How long should they follow it for?

Until the AoA becomes below the critical AoA and the stall warning stops.

Or are you talking about a simultaneous failures of at least two of the three airspeed indications AND the AoA vanes which are fully independent of the pitot tubes?

And, finally, even in that case, which is NOT what the AF crew was facing, there are ways out. Since the AoA is equal to pitch minus path, if you keep the nose relatively level and a not too high descent rate (which is what they SHOULD have done by following the memory items or at least the common sense of the most basic airmanship), you are not going to stall.

[Gabriel]

what about stall recovery? simple procedure: nose down... . do you mean to tell me that the vaunted airbus engineers couldn't program a simple set of instructions like that??

Stick pushers have been instaled on some certified airplanes since about 50 years ago.

maybe i'm so far off some of you will laugh. so be it. but my pilot friend has told me countless times that the "computer" is better at fine adjustments and precision flying than a human pilot. if that is so, and some of these procedures require a bit of precision and finesse,

I don't agree. These pilots could have followed the momory items with a gread deal of imprecision and they would have not crashed or even triggered the stall warning. The memory items are just the center of the safe envelope for a large range of altitudes and any airplane weight. There is quite margin around them that are still perfectly fliable (meaning no overspeed, no stall).

clearly, the computer can calculate and analyze the raw data it is receiving far faster than any human. clearly it can scan and note many many more variables simultaneously than any human can. just look at the CVR in the report: the amount of data listed is impressive. surely the pilots were not aware of half of it. in fact, it doens't appear that they were aware of anything and were quite lost and confused. but good ol HAL knew--and he did nothing except record the last agonizing moments of the humans existence. all HAL needed was a few more lines of code, then, with his clearly superior ability and speed, he could have fixed the problem, smoothed out the flight, and 200+ plus humans would have landed in Paris.

That's correct, but the Airbus was not designed as and unmanned plane but as if it had a barely sufficiently trained flight crew behind the controls.

What do you propose to do with all the less tecnically advanced planes that are flying now? The 747, the turboprops and regional jets (most of which lack even an autothrottle), the 747-400, the 757, 767, DC-10 and MD-11, DC-9 and MD-80/90/717, Bae 146, etc, etc, etc? None of them are capable of following the UAS memory items, none of them is capable of following the stall recovery procedure (although some have stick pushers). We ground all them?

[Gabriel]

Sound attenuation joke spoilt. The bait was out and he is usually so quick on the trigger.

The sound attenuation joke is an old one:

Delta 1234, turn left to 090 for noise abatement.
Center, Delta 1234. What are you talking about? Noise abatement at 35,000 ft?
Delta 1234, turn left to 090 NOW or there'll be a big Bang!

[Gabriel]

I wish people would stop using the "pilot bashing" lines especially in this thread.

I have only been here for a short time compared to some and most but frankly there is not a person among you that I think is honestly "pilot bashing".

Did the pilots seemingly make mistakes?

Did the system break down regarding training?

Did some person screw up? Of course.

It is possible that BEA and/or Airbus would like us to think that pilots were at the root. The job of the (French or Euro), ALPA will defend pilots and all of this positioning is to be expected.

How long are we away from the "final" report?

For the record, as much as I am angry and mad with what the pilots did here, I've said several times that "pilot error" is never ever the cause of any crash, unless the "error" was intentional. In cases where pilot performance was extremely bad, like this case, Clogan, Pinnacle, and others, the question is how did that pilot managed to get and stay where he did.

Because there will always be bad pilots around, it's the responsibility of the airline to protect itself from them by diligently selecting, training, monitoring and, if needed, getting rid of pilots to ensure that only pilots that are (made) good enough are allowed to fly their planes.

[Gabriel]

I'm struggling with this discussion. The altimeter measures increases and decreases according to a pre-set standard temperature/ density for each FL? So if it reads 35000' it may not actually be that height if the air temperature is much lower than what the altimeter has been given for that height, I could be flying at 34000'. Have I got that right? Is there a belief that the aircraft might have stalled because the air temperature was higher and the density less than needed to maintain lift at that speed? And the computers are unable to warn, advise and compensate for these discrepancies?

Short answer: No.

Long answer:

Altitude ASL (above sea level): Vertical distance from the plane to the mean sea level directly below.

Altitude AGL (above ground level): Vertical distance from the airplane to the ground (or water) directly below.

Elevation: Vertical distance from the ground to the local means sea level.

With these definitions, Altitude ASL = Altitude AGL + elevation.

But, the planes doesn't actually measure "altitude". Not for nothing the standard altimeter in any airplane is called a "baro-altimeter". It measures a pressure. How does it work?

If you get a barometer (or a manometer, i.e. a device used to measure pressure) and you climb a mountain or a building, you'll see the pressure drop as you climb. If you knew the relationship between the pressure and the altitude, you could get a marker and draw on the instrument a scale in ft with the corresponding altitude for each pressure. That's a baro-altimeter. But, do we know how the pressure changes with the altitude?

There is a model, called ISA (international standard atmosphere) that gives the profile of several properties of the air with the altitude ASL: temperature, pressure, density (calculated from the previous two), speed of sound (calculated from the temperature) and viscosity. Let's take just the pressure by now (we'll also use the temperature later).

So, using the ISA model you can calibrate the baro-altimeter to show a given altitude for each atmospheric pressure around the plane. That's it?

No. The ISA is not the real atmosphere. It's a model. And the real atmosphere is not always the same (otherwise we would not have weather). In fact, the ISA can be thought of an "average" atmosphere, with the real atmosphere any given day (or second) seldom (if ever) being exactly coincident with the ISA model.

Then? What do we do? Two things:

1- We add a new definition:

Pressure altitude: Altitude in the ISA model that would give the same pressure as the real pressure that's around the plane now.

At somehow high altitudes (meaning above 3000ft if you are in Europe or above 18000ft if you are in USA) you just go with it. You use the standard atmosphere as if it was the real atmosphere at that place and time. That basically means that you'll not be at the altitude your altimeter shows. But that's not a problem: All airplanes will be flying with the same "error", and hence, as MCM said, two airplanes flying at 20,000 and 21,000 ft of pressure altitude (indicated in the altimeter) might not really be at those altitudes ASL, but they'll be 1000ft apart. Also, the minimum safe altitude that you can fly in any given zone (and that's shown on the aviation charts) take into account all the possible variations in the atmospheric pressure, so if you fly that altitude you'll always clear the terrain and obstacles in that zone, although maybe one day you'll clear them with 1500ft and another day with 2000ft (even when the obstacle has not moved and you altimeter indicates the same altitude in both times).

2- For take-off and especially for landing, that's not enough: Here you are flying too close to the ground and don't have so much margin to cover all the possible pressure variations. Se we add an adjustment knob to the altimeter. With this know, that can be thought as an offset setting, you set the pressure that there must be around the plane for the altimeter to show "zero ft".

So if you set the pressure that is there on the ground at the airport, the altimeter will show zero when you are at the airport and hence the altimeter is now showing not the pressure altitude but the altitude AGL. If, instead, you set the mean sea level pressure at the airport, the altimeter will show zero when you are at sea level near the airport, and when you are at the airport it will show the airport elevation, hence the altimeter is now showing the altitude ASL.

This raises another question: And how do we do to fly the pressure altitude? We set the ISA pressure at sea level: 29.92 inches of Hg or 1013.3 hPa.

Here you have a baro-altimeter where you can see the adjustment knob and the "pressure for zero altitude" setting:





And finally, what about Air France?

Well, the Physics the airplane flies on doesn't care about the altitude or even the pressure. The engines need enough air to mix with the fuel and burn it. The wings need enough air to make lift. And enough air means DENSITY. More dense the air, more mass of air that is flowing through the engines or around the wings. The airplane will have the same performance whenever the air density is the same.

It's time for a new definition:

Density altitude: Altitude in the ISA model that would give the same density than the real density around the plane. So now we can say that the plane will have the same performance whenever the density altitude is the same.

The problem is, as much as the plane doesn't care about the pressure, the altimeter doesn't care about the density, or density altitude. You just can't measure the density altitude with a baro-altimeter. But there is a way out.

Think of pressure, density and temperature. They are linked by the formula:

pressure / density = constant x temperature, which can be written as density = constant x pressure / temperature

It is just a variation of the famous Boyle - Mariotte law for the ideal gases: PV=nRT

So, if we knew the pressure and the temperature, we would know the density.

Do we know the pressure? No, but we can know the pressure altitude. In fact it should be displayed right there in your altimeter unless you are flying somehow low. And for every pressure altitude there is one pressure. That's the ISA model, remember?

Do we know the temperature? Yes. An outside air temperature is a required instrument in the cockpit.

So we could take the pressure altitude, look up in an ISA table for the pressure corresponding to it, read the outside air temperature in the corresponding instrument, grab a pocket calculator, make the simple calculation shown above to get the density, grab the ISA table again, and look up for the altitude corresponding for that density. That will be your density altitude.

That, or, we can make a double-entry table or a chart that gives us the density altitude in function of the pressure altitude and outside air temperature.

Or, modern avionics give you the density altitude right away, automatically calculated from the outside air pressure and temperature that they constantly monitor and the ISA table stored in its memory.

In the Air France case, they had planned for an altitude of 37,000 ft assuming standard conditions. But now they were actually flying at a pressure altitude of 35,000 ft but the temperature was above the standard temperature for that altitude. More temperature means that the air is less dense, so the density of the air was close to that corresponding to 37,000 ft in standard conditions. In other words, their density altitude was already close to 37,000 ft while flying at a pressure altitude of 35,000ft.

But they knew it. That's why they didn't climb to 37,000 ft of pressure altitude as planed, but instead stayed at 35,000ft and mentioned that they were not able to climb to 37,000 ft as planned because the temperature was warmer than standard.

There's nothing wrong with that and the stall didn't happen because of that. The belief (and I cannot think of anything different to this) is that they stalled because they pulled up too much for too long.

[EconomyClass]

I wonder out of the thousands (perhaps millions) who are cocksure they know the failure that doomed 228, how many are willing to put themselves and their lives on the line to show it was so easy to avoid the accident. Sully has weighed in repeatedly modestly asking for humility for those who are passing judgment, but you wouldn't know it reading the threads here.

[Fear_of_Flying]

I wonder out of the thousands (perhaps millions) who are cocksure they know the failure that doomed 228, how many are willing to put themselves and their lives on the line to show it was so easy to avoid the accident. Sully has weighed in repeatedly modestly asking for humility for those who are passing judgment, but you wouldn't know it reading the threads here.

The term "passing judgement" really doesn't apply here, since no one is suggesting the pilots did anything intentionally wrong, or impugning their character or morality. As a passenger, however, I am appalled at their apparent lack of competence, in the same way I was appalled reading the Colgan transcript, and that impression is validated by people who know far more than me about aviation who seem to feel the same way. The plane was not doomed; they crashed it. If a baseball player is hitting .120 and has 14 errors in half a season, am I "passing judgement" when I assess his performance as poor?

Each time you get on an aircraft, or wave goodbye to your kids as they go up the ramp, you put your faith that the professionals in that cockpit know what they're doing and have the ability to protect their passengers. It seems clear that these pilots did not, and now it only remains to determine why, and how to avoid this happening again. At this moment, however, I see no reason to believe that history will not repeat itself again.

Don't forget, Sully has also expressed concern over the declining skills of today's pilots.

[Gabriel]

I wonder out of the thousands (perhaps millions) who are cocksure they know the failure that doomed 228, how many are willing to put themselves and their lives on the line to show it was so easy to avoid the accident.

I would. Anyday. In the sim.

[Jpmkam]

New note from BEA:

http://www.bea.aero/en/enquetes/flight.af.447/pressrelease03august2011.en.php

The English version of the interim report is promised for Thursday.

I am still shocked there was almost no communication at all from the Captain when he returned to the cockpit. Wouldn't you expect at least some questions by the Captain as to what was going on, and offering some guidance as well?

[alasta]

In just about any jet that I know about, ranging from Cessna Citation to the A330, the stall warning is triggered by a sort of weathervane, except with an horizontal axis, that protrudes from the side of the fuselage usually near the nose.

What do all these different systems have in common? Just two things: 1) They are not sensible to speed and 2) they are sensible to the angle of attack. Thus, the stall warning doesn't know or care about the speed AND is ALWAYS triggered when you exceed a critical angle of attack.

Gabriel,

What you have explained here is consistent with my understanding, but I'm struggling to get my head around why the A330's stall warning alarm reportedly goes silent at airspeeds below 60kt even if the AoA is still over the stall threshold. You've stated yourself many times that if AoA is used to detect a stall then airspeed is irrelevant, so why would the stall warning system make reference to the IAS at all?

Alternatively, have I misinterpreted the BEA's description of the A330's stall warning behaviour?