AF 447 report out

[DozyWannabe]

@BARKINGMAD:

No probs.

I don't want to bang on too much about this - but AoA gauges were also discussed, ad infinitum, on the Tech Log threads. I'd suggest looking there.

@SSD:

Agreed - one of the more thought-provoking aspects of the Tech Log discussion was that the PF in this case seemed to be suffering a startle response so profound that he immediately started pulling up before even trying to assess the situation, thus making said assessment harder for both himself and his colleague. Given that they were unable to diagnose a stall with the warning going off, ADI indicating a nose-up trend and a rapidly unwinding altimeter - would an AoA gauge have improved things or merely heaped more data on them when they were struggling to process what they had?

[Fox3WheresMyBanana]

AoA is particularly critical on a thin, narrow delta at the high end of the range.

It's important, but even more so at the other end, e.g. single engine heavy weight swept wing approach in the Tornado F3 - ideal AoA 16, Limit for control 17, Limit for u/c 15...no pressure!

[Shaggy Sheep Driver]

Fox 3 that's what I meant; "at the high AOA end of the range". On Concorde, a small increase in AoA induces a big increase in drag at that high end (like on the approach). VZRC lurks! Especially if there is an engine out. Bad news near the ground.

[BOAC]

As Dozy says, 'AoA indicator' has been done to death (no pun intended) in Tech Log. Yes, a very useful piece of kit. In my (limited) experience in partially jet-borne operations, vital, BUT it does require an understanding of how aeroplanes and wings actually work which so often seems to be missing on PPrune and particularly in the cockpit of 447, and of course, that nightmare !MONEY SPENT! on installation and training.

[gazumped]

Bubbers44

I did some very interesting test flying many years ago in some light twins. The reason was to explore a "deep stall" ie, after provoking a stall(@ 10,000 feet or so) wait for the nose to drop and then pull in full back stick for one second.

The results were amazing, and suffice to say, enough to gain a very healthy respect for speed control at anywhere close to the ground.

These light twins were generally very docile aircraft, PA-31, EMB110 , B200, and similar. The height loss is these deliberately provoked stalls were universally in excess of 2,500 feet. This, from aircraft that regularly lose no more that 50 to 100 feet in stall training.

I have never stalled a heavy jet at or close to its service ceiling, but I would say it is high time this was included in the standard syllabus.(of course only in the sim). I would have a guess that once "deep stalled" a A330 would require in excess of 15,000 to recover, and only then if the correct control Imputs were done.

Thoughts anyone?

[DozyWannabe]

Originally Posted by gazumped

I would have a guess that once "deep stalled" a A330 would require in excess of 15,000 to recover, and only then if the correct control Imputs were done.

Thoughts anyone?

If you have a look through the Tech Log threads, you'll see that we did some experiments in an A320 sim (didn't have access to that of an A330/340). So with that caveat, plus the fact that the sim works on extrapolated flight test data in the stall regime and a non-pilot (i.e. me) was at the controls for some scenarios - we had the aircraft recovered from a stall at FL380 within about 18,000ft using just the sidestick to recover, and about 12,000ft using the trim wheels to centre the THS plus sidestick to initiate nose-down/descent.

[BOAC]

Way back on thread xxx- I ventured my opinion that around 20k was probably the last point at which they might recover, based on nothing but a 'feeling' and a thought about the 50 degree nose-down pitch change required....................

[Ye Olde Pilot]

This is what happens when you have system managers and company SOP's in charge of flying instead of training pilots the old way. I'm sure a ten year old kid on a computer game would have spotted this problem.

The whole transcript should be drummed in to anyone who aspires to fly a commercial aircraft.

[gazumped]

Well lets do some real world estimations. Your stall recovery in a sim in a A320 lost 18000 , and you knew beforehand you were going to stall and were mentally prepared for it (probably silently went over the actions in your head before the event), confident your life and limb weren't at risk, and still lost 18000!!!

Now a dark and stormy night, UAS, unexpected autopilot kick out, startle factor, fear for your life, as your past flashes before your eyes, would have to add a50% factor to it, makes a recovery possible in say roughly 27000 feet!

This should be added to every heavy jet training syllabus immediately!

Our company made a mod to the sim cyclic after thus event, on climb out "one " IAS failed and you had to determine which was the faulty one. Then we did approach to stalls at 37000 with a full panel. I don't want to be highly critical of my own company but for heavens sake talk about underdone!!

I have voiced my opinion to anyone who will listen and just get some vague platitudes, and then the listener's eyes seem to glaze over. It seems all too hard!

There have a number of high altitude loss of control events over the years, and AF447 seems to be just another one.

[HazelNuts39]

Originally Posted by gazumped

I would have a guess that once "deep stalled" a A330 would require in excess of 15,000 to recover, and only then if the correct control Imputs were done.

The correct control input is initially quite simple: full nose-down maintained until the airplane is unstalled and the stall warning stops. Then comes the tricky part - recovering from a steep dive without overspeeding and without entering a secondary stall.

A near optimal manoeuvre would have recovered the airplane within about 11000 ft to level flight at a speed close to Vmo.

[gaunty]

Click, click, click.

[gazumped]

Hazelnuts

I don't think you could say that this crew of AF447 , were ever in danger of performing a near optimal manoeuvre.

The sim tests in a 320 mentioned earlier give 18000 as an achievable figure, my fudge factor of 50% I think would be reasonable netting a cool 27000!!

So somewhere between 11000 to 27000, I can guarantee you these numbers would be known by very very few line drivers(or senior management )

The argument for inclusion in sim cyclics can hardly be stronger.

I seem to recall a China Air B747 losing control because of an un-noticed outboard flameout, the resulting auto pilot unlock and VMC/stall/over speed was not recovered from till the aircraft broke out of cloud and became visual. The aircraft diverted to Anchorage, and was subsequently a total write-off.
This crew ad a complete panel of serviceable instruments.

It would seem the sheer lack of hand flying skills has been around for some time.

[NWstu]

gazumped

I seem to recall a China Air B747 losing control because of an un-noticed outboard flameout, the resulting auto pilot unlock and VMC/stall/over speed was not recovered from till the aircraft broke out of cloud and became visual. The aircraft diverted to Anchorage, and was subsequently a total write-off.
This crew ad a complete panel of serviceable instruments.

Link to AAR here

[Brian Abraham]

diverted to Anchorage, and was subsequently a total write-off

Landed at San Fran. Aircraft repaired and returned to service, with a little extra dihedral.

[HazelNuts39]

gazumped,

in the sentence quoted you were talking about the airplane. The crew is another matter altogether.

[llondel]

It would seem the sheer lack of hand flying skills has been around for some time.

With reference to the China Air B747, it's fair to say they were distracted by the engine failure and none of them was flying the aircraft, but once it had stalled, there was a fair bit of skill involved in getting it back to level flight and putting it safely on the tarmac.

[DozyWannabe]

The difference there was that the crew worked together to diagnose the problem in time for a recovery to be performed. That didn't happen for AF447.

[etudiant]

Equally important, it was daytime and the crew could see what the problem was.
The crew of 447 did not have that unequivocal input, they got instrument failures and stall warnings that came on when they tried to do the right thing. I think there but for the grace....

[Capn Bloggs]

The correct control input is initially quite simple: full nose-down maintained

Not forgetting to retrim forward off the full nose-up stop...

[jimjim1]

Trim

Not forgetting to retrim forward off the full nose-up stop...

My understanding is that was not necessary, although it appears that manual nose-down trim input would have speeded the recovery.


@DozyWannabe

we had the aircraft recovered from a stall at FL380 within about 18,000ft using just the sidestick to recover, and about 12,000ft using the trim wheels to centre the THS plus sidestick to initiate nose-down/descent