Let me agree with this particularly with regard to the aircraft pitch trim system.
An aircraft that keeps trimming the force off the stick/wheel as you slow might be convenient, but it loses natural speed stability and requires increased attentiveness/awareness in manual flight-or an automatic speed control system.

Any aircraft that keeps automatically trimming nose up after the stall AOA is reached has a fouled up design.:*

Maybe so, BUT ... would the sidestick input alone not have kept the airplane in the stall?

Another weakness of the design. A large transport aircraft should need two hands to hold it in a stall, not a crooked pinky finger. :rolleyes:

Granted, there are other transports that do not meet this force criteria as well.

Not so fast.

The Automation was built to ASSIST mankind not to second guess him. The Automation was in Alternate Law (NOT in Normal Law) but still allowing it to ASSIST the pilots as in "YOU WANT TO GO UP?. My God you shall go up with all the might of this airplane (THS full up)". But the instruction from the pilot was and still is WRONG cause a High Altitude STALL will be imminent and guarantied, as it did. Machines in full automation do one (and that one might be composed of many, mind you) thing only. And if that one thing can't be performed at its fullest, it must relinquish control back to humans immediately. And in the AF447, that is exactly what the Automation DID as per DESIGN.


But the way, speaking of Automation in those less than 4 minutes of the tragedy, none of the three guys had the smart idea of trying to bring the Auto Pilot back on line and then if successful the Auto-throttle later. Someone should set an A330 simulator and emulate conditions as in the AF447 at level flight100 and throw the AP back on and if successful throw the Auto-Throttle just as well. See what happens. You know what would be funny but not really that so, don't you? Just in case, I am presuming that deploying the spoilers was a great idea of course:ugh:.


:confused:



Cheers,

Uhmm I hope a bit better than that.

I allways cringe when I hear something discounted as "that's a 1 in a million scenario". If you do the math on departures per year one in a million is a bit on the low side...

Folks, was there a diagram posted on here of the sidestick mechanical layout/operation?

I remember seeing it but what with all the moding/editing have lost track of it...

Let us not confuse automation, basic irreversible hydraulic controls, and the benefits of good aerodynamic design.
What told the THS to go up was not the high level of automation that the 'Bus can deliver, but a flawed version of basic aerodynamic common sense that used to be designed into the basic airframe, but now resides in a black box.

In days of yore when flying 'big iron', you would have to trim up next to the stall either manually or with the trim button, and then give a firm heave on the controls for a prolonged period of time to get well into a stall (Think something like the DC-6)
Then along came the jets with their higher control forces and hydraulic boosted, then completely irreversible controls followed causing our control forces to be generated artificially.
Then we had the T-tailed aircraft with their pitch up in the stall that need protective devices to certify the airframe and we became used to flying aircraft with bad aerodynamic characteristics as long as we were protected from the consequences.
Now we have fly by wire aircraft with their C* (attitude stable) and C*U (speed stable) control laws that have permitted straight forward implementation of flight management systems, but what have we lost? Could it be that the smarts that used to be built into the basic airframe design that limited control authority to that which was necessary for the worst condition, (full flaps and forward cg in the case of nose up authority) didn't make the migration into a little black box because it was too hard to do so? That is what I think happened in the A330 Alt 2 law THS limits.

Glad Rag, look over in the Tech AF447 thread for the sidestick diagram.

depends on the consequences

It it were to be a 100% probability that the consequences would result in an inability to complete a safe flight and landing (somewhere) I would agree with you.

Typically the 1 in a million failure category still has other systems available to get you back on the ground 9 out of 10 times.

In the long run it's easier to work this 9 out of 10 back up redundancy than to try to make a ten fold improvement in the original malfunction causes

Glad rag,
You'll find it at AF447 thread no.7 page 27 post 523.

I think there is a 100% probability that if 100 airliners lost airspeed at 35,000 ft all of them would end up the same way if you pulled full back and held it for over 3 minutes. All pilots but one would have held attitude and power and survived.

One pulled full back and crashed. The others aviated as trained. Seems simple to me. We don't need better automation, we need better pilots. How are the new hires today doing? Can they hand fly or is that not required any more?

I gave a list of stalls (off top of my head) that were mostly not sidestick and not FBW but all (I think) involved pilots pulling back in reponse to stall warning.

I quote an investigation report of inappropriate response to stall warning that states that pilots were not trained in stall recovery and that this was a "minor safety issue".

I say it'll carry on until it's acknowledged that pilots pulling back in a stall is a major problem, and you come back with "it's a FBW problem".

Meanwhile others post stuff like "None of us in any Boeing would have stalled". [ yeah and the Ethiopian guys were shot down ].

Q.E.D.


You can open your eyes and mind to the real problem, or carry on believing the engineers have screwed up between the stick and the surfaces and there's no problem between the stick and the chair, carry on stalling, carry on pulling back, carry on dying. Gotta keep the accident investigtors in a job...

I think it is the most basic thing a pilot learns is stall recovery. It is so simple, release back pressure and if necessary push forward. If a pilot doesn't know that he isn't a pilot, he is a robot. Unfortunately some Airbus pilots have become robots that can't fly any more without a computer to do what they ask it to do.. SAD

No auto trim up trap and a much better chance to improve things, intentionally or not :

MACHINBIRD

That is where you are totally wrong. Tried it in the sim and it takes great effort on the sidestick and and effort that has to be maintained!!:ugh:

Wrong again it makes it more stable.

Thanks Hazel, thanks very much!

:ok:

I did here.

No effort at all on the sidestick to hold in the stall ... the almost full NU trim was taking care of it.
Apparently, same thing happened for AF447, THS was already at 13 degrees up when full back stick was maintained.
That undesirable auto trim operation has played a crucial role in the outcome of AF447.

A good friend of mine, president of our local Tuskegee Airmen chapter, is a check airman for Continental on 737-800s. He told me the other day that he encourages pilots to hand-fly an entire arrival and approach whenever possible.

Report: Cobham B712 at Kalgoorlie on Oct 13th 2010, stick shaker on two approaches

Now what? Redesign the B717-200? Expend gazillions of $ so the FMC/FMS warns the pilots when they introduce/punch the wrong/bogus data?

The buck has to stop somewhere and the sooner the better.:ugh:

Cheers,

Which was supported by Owain Glyndwr's theoretical analysis - in part...

Iceman.
My first impression was that you had done your sim evaluation in a Boeing type sim, but I see you list yourself as a 'Bus pilot, seat 0A, so will give you the benefit of the doubt.

It takes effort to hold the stick back in the 'Bus to get the nose up, but once it is there sufficiently high, and the trim runs up, it takes little effort as Confiture describes.

Do not confuse the ease of holding an airspeed in your 'Bus due to the attitude stability with possession of natural speed stability. Your 'Bus is a C* aircraft and is attitude stable, not speed stable. That got AF447 into trouble in Alt 2 law. Once the nose got pointed well up into the sky that aircraft was going to decelerate, even with no further stick input, and the trim was going to run up to the limits in blind obedience-no protections.

If you don't believe me, try flying the 'Bus sim at 12-15 degrees nose up in Alt2 law at altitude. Once there, it won't make a difference what you do with the power. As long as you keep your hands off the pitch input, that aircraft will decel to the stall and keep on decelerating, and the trim will run as far up as it can.

We can never make an aircraft idiot proof. No matter how they try Airbus won't either. Basic airmanship like back 20 years ago will get you out of most anything. Just make pilots the same quality we had 20 years ago and we won't have this type of screw up. Stop the pilot mills and start making real pilots again. I know this will piss a lot of people off but it needs to be done.

Wouldn't hurt to go back to making real airplanes again either !!

Question from SLF. Is that not a problem right there? I mean by way of a disconnect between man and machine.

The handling qualities of earlier aircraft stipulated stick force per "g", and it was up to the pilot to trim out any stick force. If disturbed the aircraft naturally sort to return to its trim speed.

It's hard to imagine the pilot of a non Airbus holding full back stick and manually trimming full nose up no matter what the circumstance.

The power of a trimable stab is such that it becomes the primary pitch control, with the elevator assuming the role of a short term pitch modification.

The pilot of old automatically knew the trim state of the aircraft because he put it there, where as the Airbus, I gather, the pilot is only aware of where the automatics have put the stab is by reference to instrumentation.

You may be SLF, but you have made a very salient point, that the pilot knows the trim state because he put it there.
Earlier designs (thinking DC9/737) kept pilots in the loop whilst auto-trimming, by either having an irritating honking horn, or a chunky great blo@dy wheel clanking round. In this accident of course neither would have auto-trimmed, although the Turkish accident in AMS showed us another danger of auto-trimming, even on a dinosaur like a 737.

Brian

I totally agree, Sir. However, having flown Boeing 737s, as well as Airbus 320 and 330, I would add that it is equally hard to imagine the pilot of an Airbus holding full back stick.

It seems to me that the initial error was the PF's failure to perform the Unreliable AirSpeed (UAS) drill, which is a memory item. He then pitched up to an unrealistic attitude. As the aircraft would have been in trim prior to his sidestick input, the force required to hold the sidestick fully aft would, initially, have been large. That effort would diminish as the autotrim took effect, but it would take a significant time.

Let us correct this popular misconception. Initially PF did not yank full back stick. His average aft stick was about 6 degrees out of 16 degrees possible for the first 10 seconds. What he was mostly doing was chasing a roll oscillation (PIO?).

It was not until later when the aircraft became fully stalled that he began to hold full back stick-probably to prevent the nose from falling through.

Does he even need to hold in back stick to keep the nose up? or doen't this apply to the law they were in?

http://www.ahrtp.com/EG_Images6/EADS...267_Airbus.jpg

Was he perhaps trying to maintain 15° pitch attitude?
http://i.imgur.com/krjIV.gif

What PF was doing would make better sense if his PFD had a ~ 12 degree pitch bias, but I don't think that happened. Instead, he was probably trying to maintain around 1 g and the harder he tried, the worse his problems became.

Cwatters, your diagram is essentially correct except there wasn't much good aerodynamic data available, so no protections that would have helped. Since they had no idea of Mach number (which affects stall AOA), they didn't even give the AOA system authority to control the aircraft and push the nose down, instead limiting it to shouting, "Stall" repetitively (until silenced by the A/S <60 knots).
Roll however was in Direct law so that was a whole new ball game for PF's level of training.

I am personally of the opinion that other than the airspeed system, the airframe systems performed exactly as designed.

I am not ready to cast PF into the entirely unfit to fly bucket yet. I have this uneasy feeling that many of our weaker brothers in the cockpit would have succumbed to this same scenario, particularly because there is probably something about the total scenario that we do not properly appreciate.

The human factors group at BEA will be having an interesting time if they do their job correctly. I only hope they can explain their understanding adequately.

Looking at that graph, I guess it's entirely & depressingly plausible.

He selected toga as well. Toga/15° is, I believe, the first pitch/power setting on the AF uas procedure (entirely the wrong one to use for cruise, but did they know that...)

Not that hard. It is actually what Airbus wants to force in your DNA for a number of situations. no thinking, just full back stick to get the best of the protections ...
That gives very good performance ... as long as data are reliable.

Let me disagree here again :
the force required to hold a sidestick fully aft does not change a iota whatever the situation, whatever the trim setting.

Machinbird
 

The poor PF was not only the PF, he was also the deputy Captain at the time.

He was trying to correct a PIO when by the book he should (as Capt) have called Unreliable Airspeed.

PNF didn't realise that all this was way too much for the PF. He could have done more to save the ship.

I suspect, without proof, that AF procedures (seemingly assuming that nothing would go wrong) contributed a lot that night.

Really ? Pop-quiz - which (now somewhat fragmented) aircraft is this then:[hint - it's one of those hard-to-imagine non-Airbuses]

Auto-trim is damned whatever it does.

447, auto-trim stays on, pilot pulls back through stall warning, trims up in stall -> "auto trim should hand back to pilot in stall"

Perpignan, auto-trim does hand back to pilot, pilots never re-trim, leads to fatal secondary stall. Would have been much better chance of recovery had auto trim followed pilot commands. -> "auto trim should follow pilot commands in stall"


It's not just an airbus problem however:

Ethiopian, Beirut - mis-trimmed, pilots never re-trim, pulled back through fatal secondary stall.
Turkish, Schipol - auto-trim does hand back to pilot, pilots never re-trim (recovery fails because they let throttles go back to idle again however)
Colgan, Buffalo - auto-trim might have had a hend in setup, but they pulled back all the way down and pulled the flaps in just to be sure...


Sure, there may be a pattern there somwhere, but "airbus" ain't it.

Auto-trim is a has been with us some time and has been known for causing trouble. Pilots first began to run afoul of it with the early autopilots that would trim unbeknownst to the crew-until it then kicked out and left the crew with a handful. :eek:
Some implementations of FBW require autotrim to function properly and it can simplify the job of flying. Autotrim has teeth however, and pilots who regard it as piece of the furniture in the cockpit and thus lose awareness of what is doing stand a great chance of being adversely surprised by it.

That is why most of the modern aircraft have noisemakers associated with trim movement. But one major airframe maker skipped this step. Why is that?:confused:

And the answer to the pop quiz is the Ethiopian B-737 at Beruit. For whatever reason, PF didn't want to handle the aircraft that night, but the A/P never relieved him. Don't you guys want to earn your pay for flying?

I realise that but..

For me the important bit is that neutral side stick means "maintain this attitude". eg It's not necessary to hold in back stick to keep the nose up once it's up.

What does the equivalent diagram look like for other aircraft? Do you have to hold the stick back to maintain a nose high attitude or will neutral stick do it?

For conventional non-FBW aircraft and Boeing FBW, it looks like the well known phugoid where the nose drops as you slow from trim speed. (But let me preface this with the statement that my Boeing FBW understanding is theoretical.)

Auto trim may be nice to make flying easy but it doesn't do much to help piots fly when things go amuck. These pilots were not the brightest so let it lead them into a deep stall because they were stupid.

We all know why they crashed. They were incapable of flying an aircraft with unreliable airspeed. All aircraft have charts and pilots should know approximate attitudes and power if airspeed goes bad. Nobody pulls full up on the stick hoping the automation will save them but they did. When they went into alternate law the airbus was no longer idiot proof.

Machinbird

Very condescending!

Perhaps you could also grace us with your practical / hands on knowledge and experience of the flying the Airbus. ( An Airbus pilot TRE/IRE, + 20 years military- Rotary(piston/gas turbine) fixed wing (piston and fast jet), 6 years Boeing (757/767) and 16 years Airbus flying A340/A330)

I believe that this is correct provided a thrust is set that maintains arspeed for the chosen trajectory.

A Bit of History
 

Positive aircraft stability means that when you let go of a control, the nose of the aircraft returns to the speed that you had when you first moved the control. This speed is called the trim speed. Neutral pitch stability means that when the nose is moved, the pitch angle remains where you put it.

When the Wright brothers started developing aircraft a century ago, they initially worked with the premise that aircraft should exhibit positive stability in yaw but be neutrally stable in pitch and roll. As the first test pilots, they quickly discovered that it is easier to fly if you have positive stability in pitch. If you pull or push on the stick in pitch and let go, the aircraft returns to the trim speed and settles down. This is particularly helpful in turbulence. If a gust displaces the nose in pitch (or yaw) the plane simply returns to the trim speed without help from the pilot. Accidentally bumping the stick or yoke is no problem.

Some PPRuNers have asked why the AF447 PF didn't simply let go of the side stick and let the plane resume the trimmed speed. Why indeed? In every plane those PPRuNers had flown, letting go would have allowed the plane to return to trim speed. In the AB system, letting go, which neutralizes the stick, does nothing but stop the movement in pitch. It does not cause the aircraft to resume the previously trimmed speed because the autotrim has set a new trim speed. The trim speed set by the autotrim in AF447 (in response to the PF's side stick input) was below stall speed.

With the AB flight control system, because it incorporates autotrim in pitch, the aircraft essentially becomes neutrally stable in pitch because the autotrim follows the sidestick command. The airplane does still retain positive stability in pitch when the nose is moved by external aerodynamic forces, but if you move the stick in pitch, the nose stays at the pitch angle commanded.

With a century of flight experience under our belts, why did AB decide to alter the flight control philosophy? Good question.