Air France 447 - On topic only!

[Jpmkam]

The good news continues. First they found the recorders which I thought was a longshot, now they've successfully been able to access the data they contain. The ability of the recorders to withstand the depth and time underwater will serve as a new benchmark for future recoveries if necessary.

Accident investigators have successfully downloaded all of the information captured by the flight-data recorder aboard the Air France jetliner that killed 228 people two years ago when it mysteriously crashed into the Atlantic Ocean, French investigators said Monday.

Led by French air-safety specialists, an international team of government and industry experts over the weekend retrieved data and voice recordings expected to help pinpoint the causes of the June 1, 2009 accident, the Bureau d'Enquêtes et d'Analyses, or BEA, said.

The BEA said the team collected "all the data" from the flight-data recorder and "the whole recording of the last two hours of the flight" on the cockpit voice recorder. The BEA added that "all of this data will now be subjected to detailed in-depth analysis." The analysis "will take several weeks, after which a further interim report will be written and then published during the summer," the BEA said.

The flight-data recorder was in excellent condition and showed almost no corrosion from salt water or other types of damage, according to one person familiar with the work, despite spending nearly two years on the seabed at a depth of 12,000 feet.

The Airbus A330's cockpit-voice recorder, according to one person, suffered slightly more damage and required greater efforts by investigators to clean and dry its internal computer chips and memory boards.

Over the next few weeks and months, investigators will fuse the different strands of recorder data to reconstruct a single, precise timeline and virtual replay of the high-profile accident, which until now has confounded experts.

[Evan]

Yuck, this sucks! You are right. That's the worst of both worlds!
Load factor demand and no stall protection!!!

Yes, this is where it gets interesting. But, remember, two ADR faults were detected, so the systems were aware of the UAS condition. Therefore, any protection that requires reliable airspeed data would be inhibited:

You just have this:
A low speed stability function replaces the normal angle-of-attack protection

System introduces a progressive nose down command which attempts to prevent the speed from decaying further.

In the case of a failure of 2 ADR's, there is no low speed stability.

An audio stall warning consisting of "crickets" and a "STALL" aural message is activated.

Without airspeed data, will this function? Can it function from the AoA sensor alone?

But then you have this:
Certain failures cause the system to revert to Alternate Law without speed stability.

I wonder if an unreliable airspeed is such "certain failures"?

Yes, exactly, see above.

So leave it alone, slow down, the flight control computers will keep 1G until it stalls with no nose-up command from the pilots! (human or auto).

I want THIS feature removed NOW!!! Forget about the "gimme my frigging plane now" guarded switch.

This is from the KLM Flight Crew Training Manual:

When the PF performs sidestick inputs, a constant G-load maneuver is ordered, and the aircraft responds with a G-Load/Pitch rate. Therefore, the PFs order is consistent with the response that is "naturally" expected from the aircraft: Pitch rate at low speed; Flight Path Rate or G, at high speed.
So, if there is no input on the stick:

.

• The aircraft maintains the flight path, even in case of speed changes.
• In case of configuration changes or thrust variations, the aircraft compensates for the pitching moment effects.
• In turbulence, small deviations occur on the flight path. However, the aircraft tends to regain a steady condition.

So, going back to this:

That's the worst of both worlds! Load factor demand and no stall protection!!!

Not necessarily. The memory item for pitch (the one pilots never seem to remember) is to apply and maintain 5° ANU until more exacting calculations of weight and altitude are made (which call for 3.5° ANU by the QRH chart for 205t at FL350). So what happens if the pilot does make the necessary inputs? The FCC should maintain the new load factor commanded and the thrust should ensure a safe power setting to keep airspeed and AoA within range (Although if the memory items are applied and not corrected by the tables within a reasonable span of time, the manual indicates that overspeed can be expected). In fact, load factor demand can be a blessing here: because the FCC are maintaining the FPV with pitch corrections, once you are more or less stabilized, speed can corrected by observing pitch and correcting with thrust.

BUT if the pilots did not perform the memory items, or did not perform the thrust memory item correctly (move the thrust levers out of CLB detent and back into CLB—which is necessary to ensure full CLB thrust but NOT mentioned in the memory items or the QRH procedures) and thrust was at a reduced setting for turbulence penetration at the time of A/THR disconnect, thrust would be locked at that reduced setting and thus airspeed would be falling off. I can see how the combination of these two aspects of the system (computer maintaining both FPV and N1) would require increasing AoA and could lead to a stall. This assumes that the pilots:

• did not perform the memory items (which BTW are not required if the crew does not consider the situation to be immediately dangerous), and
  
• did not recognize the developing situation correctly due to other distractions.

Now, one more factor to consider about the situational awareness. Loss of airspeed data can result in unreliable altitude data. In this case, the 'bird' might be unreliable or disappear and the VSI might be unreliable as well. In this case, your options are pretty meager:

If FPV is not reliable and GPS altitude information is not available (no means to ensure level flight):
Adjust pitch and thrust according to table, and wait for speed stabilization.
Expect a significant stabilization time and important altitude variations.

That's not where you ever want to be. Especially in the middle of a MCC nightmare. And, of course, if you revert to direct law and get this...

• MAN PITCH TRIM USE

it's another bag of snakes entirely.

[Evan]

Got it!!!

Success:

PARIS — French accident investigators said Monday that they had succeeded in downloading all of the flight data and cockpit conversations from the so-called black boxes of an Air France jet that crashed two years ago in the Atlantic Ocean — a critical breakthrough that could finally resolve the mystery behind why the plane went down.

[Fear_of_Flying]

Success:

See posts 797 and 799.

[Evan]

See posts 797 and 799.

Right. Only saw the thing about the cards being dry. Whatever.

They've listened to the CVR. You have to wonder if they already know what happened. Hopefully someone will leak it to the public soon.

[Jpmkam]

Right. Only saw the thing about the cards being dry. Whatever.

They've listened to the CVR. You have to wonder if they already know what happened. Hopefully someone will leak it to the public soon.

I'm sure we'll hear some preliminary reports of what happened fairly quickly.

[Haendli]

A first statement regarding the results from the boxes is sceduled on Friday 21.May 2011.

Greets Haendli

[fmedina]

Is it hard to sync CVR data with FDR?

I read somewhere that it is somewhat complicated to sync the two? Why woudl that be, are they not using same timestamps? Or maybe this is difficult when there is missing data on one or both machines?

[Black Ram]

Thank God! From now on, the investigation should proceed smoothly, even though we will be anxiously waiting for the preliminary report.

[Gabriel]

I read somewhere that it is somewhat complicated to sync the two? Why woudl that be, are they not using same timestamps? Or maybe this is difficult when there is missing data on one or both machines?

That was mostly a matter of the past, when both the voices and the data was analogically recorded in tapes. Slight differences on the speed of the tapes, the tape slipping in the recording head, etc. made that the recordings were not in sync. The usual procedure was to use the 400hz noise in the CVR to get the speed of the CVR right, and then use mick keys events (that are recorded both in the CVR and the FDR) to sync both.

Now that the data is stored in digital format in non-volatile memory chips, things are easier.

[Gabriel]

Not necessarily.

I tell you, man. I don't want load factor demand with no envelope protection.

The FAR are very clear that there must be a positive force gradient vs speed: Starting from trim, it's REQUIRED that a push force is needed to increase speed, and that speed increase must grow with the push force applied, and the same goes for a pull force and reduction of speed. There is a very good reason for this requirement, and it's called "speed stability"

The load factor demand concept ignore this requirement. There is no force on the sidestick needed to change speed. Airbus somehow managed to get rid of that requirement, that they are clearly not fulfilling. Probably that was by arguing that the envelope protection offered an at least as effective means of preventing a dangerous speed excursion. And I agree. AS LONG AS THERE IS AN ENVELOPE PRETECTION!!!

Remove that, and the load factor demand concept just doesn't belong to an airplane. It's simply too dangerous. Figure this:

They had dozens of warnings to troubleshoot. The memory items of the unreliable airspeed were just one of many. There is the risk that they didn't notice that the A/T had reduced thrust shortly before the event. There is the risk that, as you've said, they forgot to move the T/L away from, and then back to, CLB. There is the risk that they got distracted by the other warnings, or by severe turbulence, or whatever. In a classical approach, and even in an Airbus in direct law with just a spring as the only force feedback, you need to apply increasing displacement of the yoke/stick, which also takes an increasing amount of force, to keep slowing down. No matter what you do with the throttles, the plane just won't slow down unless and until you pull up.

In normal law, on the other hand, with the load factor demand concept, the plane can slow down with no stick input: If the thrust less than needed to keep the speed, the FCCs will keep increasing the AoA to maintain 1G and won't let the nose go down (in fact it will go up), and the plane WILL slow down. But only so much. There are several layers of protections against the stall made by the FCCs, that includes adding nose down input first, then linking the stick deflection with AoA and not load factor (hence the AoA won't keep increasing unless the pilot keeps pulling up more and more), and finally the FCCs will just ignore the pilot's inputs that would stall the plane. Very much like HALL in 2001, the computer will simply not let the pilot stall the plane.

Now, remove the stall protection and the slow speed stability protection, keep the load factor demand concept, and the pilot is left with what? With no muscle feedback, be it in terms of position of the sidestick or force needed. Ok, he has some very nice screens in front of him that he can use for visual feedback. Except that they are filled with non-usual symbols and messages, some usual information is removed, and they still have to troubleshoot a few dozens of warning messages while fighting to keep the memory item pitch (and thrust) in the middle of some nasty turbulence. Ideally, that is doable. One pilot aviates, the other troubleshoots. The flying pilot might need to move some switches, look for some information, and answer some challenges from the checklists. To do that, ideally, the flying pilot would look away from the attitude indicator just 2 seconds each time, then return to the AI for a few seconds to check that things are still as he left them and check that the tendency is still that they are staying there, and only then look away for another 2 seconds.

Yes, it's doable, but it gives you no margin.

Make those 2 seconds 10, remove all feedback from the sidestick, remove envelope protections, add anxiety, turbulence, darkness and some confusion, and you are dead.

Unacceptable!

Imagine yourself driving a car where it could slowly turn with no input in, and no feedback on, the steering wheel. And that you are required to concentrate on the road 100% of the time to keep on it. Now go change the DC and I see you in the woods.

The speed stability requirement is there for something! And if the lack of speed stability becomes acceptable in lieu of other layers of safety, then it becomes instantly unacceptable the instant that these other layers of safety are removed.

As I've said, IMHO, if Airbus was working on the guarded "direct law switch" right now, I would tell them: Stop. Take these engineers. Fix this first. Then go back to the guarded switch. And that's a significant comment coming from me, a strong supporter of the need of that guarded switch.

[3WE]

I tell you, man. I don't want load factor demand with no envelope protection....

...As I've said, IMHO, if Airbus was working on the guarded "direct law switch" right now, I would tell them: Stop. Take these engineers. Fix this first. Then go back to the guarded switch. And that's a significant comment coming from me, a strong supporter of the need of that guarded switch.

I hearby proclaim that it is time for JetPhotos to un-ban In The Shade.

Do you belive he will be appearing in some other aviation fora soon, given the breakthroughs on crash of a backwards-wired, cheap French composite crackerbox?

[TeeVee]

Lol

Ok, we will await the autopsy reports... Then I will await a virtual humble pie from you.. As you NOW say you are qualified in nothing, yet question others ? roflamo !. so that is a good start. authorative is an opinion you have, contradictory without any supporting evidence is also authorative dude !!! .... I appear to be wasting my time explaining to you, in the nicest way. - agree to disagree until the final reports are made.
MMMMMMMMMMMMMMM ! Mom's Apple Humble Pie... I look forward to it..

I won't even GO to your post about 'yanks' not leaving their 'military' behind as a matter of 'pride' .. hahahaha, but then it is not appropriate for others to want closure etc ??? Jeeez...... My god / Allah etc or whatever you believe, it is no wonder the US is under constant threat and the world perception against you is based on people like YOU, NOT the general US populace... but, again, that is another story..
Have a nice day, would you like fries with that ?

since you have ow proven yourself to be incapable of carrying on a discussion without resorting to ad hominem attacks and ridicule, i'll simply place you on the ignore list. stick to hiding out in the EU on your "yacht."

[Gabriel]

I hearby proclaim that it is time for JetPhotos to un-ban In The Shade.

Do you belive he will be appearing in some other aviation fora soon, given the breakthroughs on crash of a backwards-wired, cheap French composite crackerbox?

Indeed. You can count on that.

[Evan]

I tell you, man. I don't want load factor demand with no envelope protection.

The FAR are very clear that there must be a positive force gradient vs speed: Starting from trim, it's REQUIRED that a push force is needed to increase speed, and that speed increase must grow with the push force applied, and the same goes for a pull force and reduction of speed. There is a very good reason for this requirement, and it's called "speed stability"

Do you mean static longitudinal stability? Are you referring to Sec. 25.173?

If I get you right, you are saying that the idea of the FCC maintaining a load factor that slows the aircraft below trim speed after stick force has been released, as opposed to it returning to the trim attitude as a result of static stability, is inherently dangerous in the absence of stall protections? Is that right?

I think Airbus gets around this (it was written in 1965, long before FBW technology was practical) by taking into account part (d)

(d) Within the free return speed range specified in paragraph (b) of this section, it is permissible for the airplane, without control forces, to stabilize on speeds above or below the desired trim speeds if exceptional attention on the part of the pilot is not required to return to and maintain the desired trim speed and altitude.

Meaning the FCC is designed to be doing some of that work via autotrim even in ALT2 manual flight.

I do I see your point about the danger of load factor demand without adequate speed protections. It would seem to defeat static longitudinal stability. In fact the FCOM does state:

In certain failure cases, such as the loss of VS1g computation or the loss of two ADRs [as in UAS], the longitudinal static stability cannot be restored at low speed.

I also see a danger in lack of control position feedback for awareness, as in the A/THR. I wonder why this is not servo driven. I think it is because the engineers want the pilots to think in terms of modes, not incremental values, and to let the aircraft systems sort out the details. But that probably isn't the best policy when things go wrong.

I don't know though. I would think that someone at Airbus would have anticipated this danger. There must be more to this that is simply over my head at this point.