Machinbird

Long experience with the perversity of aviation accident causes makes me extremely suspicious when I see anomalous behavior.
While I absolutely agree that the 8 degree initial bank was no big deal, I am not so comfortable in describing the PF's initial wing wobbling as minor gyrations. PF was using a substantial portion of the aircraft's lateral control authority to drive this oscillation, and any pilot with high altitude handling experience knows that this is a bad idea for a number of reasons. Although the bank angles achieved were not very spectacular until after the stall, the roll rates were very high. I do not believe an experienced pilot would have exceeded 1/10th of these lateral control inputs in "normal" Alt 2 flight. (But I have to provide the caveat that I have never flown the 'Bus)
From this chart I provided earlier, the first 4 roll half oscillations were on the order of 15 degrees each. Each one was accomplished in just over 2 seconds and PF then accelerated his inputs to drop his average time between reversals below 2 seconds. This is too much movement for the average pilot to calmly ignore.

I suspect his initial difficulty in controlling the wing roll attitude grabbed his full attention. There are hints of this in the type of strategy he used to control the oscillation. Instead of backing out of the loop for a moment and letting things calm, he accelerated his inputs to "get ahead" of the oscillation. You can see this by just graphing out the reversal intervals. As I have said a few times before, he actually got so far ahead with his large inputs that he caused at least two phase reversals of the oscillation. Yes, he did finally get the aircraft under a semblance of control- but he did it with high gain inputs. This is very indicative of an agitated mental state (IMHO) at the point just before the stall. You don't fly smooth when the adrenalin is pumping.

This second sentence is very concerning to me. From observing the posts of others, I know that stuffing the nose down is avoided like the plague in airline flying-and for good reason when you have people in the back of your aircraft.

For a tactical experienced pilot, there is a concept called unloading the wing. If you establish a ballistic trajectory with your aircraft, you do not require lift and this allows your aircraft to accelerate much more quickly than it would while maintaining high AOA. In a ballistic trajectory, your aircraft will not stall. This is the core idea behind many tactical maneuvers. In airline use, you would not have to fully adopt a zero g ballistic flight to recover from a potential low speed excursion. 1/10th g is sufficient to keep most things stuck to the floor and still provide almost all the benefit of zero g. Unloading is a concept that every jet pilot should have in their back pocket through practical (airborne) flying experience. (You cannot do it in a Sim). A pilot who understands the concept won't have difficulty getting the nose down to get flying. The only limitations are likely those built into your flight control system.

roulishollandais

Hi Maching Bird,

Unloading the wing , was one of the art of Henri Giraud, who did it once... to land in the mountain, "like a bird on the branch", on a 40% slope with fresh snow, who had two times the length of his tail-dragger... The most difficult for him was to turn his aircraft 180° to take-off toward the valley of Grenoble : That way he saved the life of a skier who had fallen, very near from the death, from a cliff.

This art of Giraud has never been and will never be teached officialy in France : Henri Ziegler for his confidants put the hand on the airspace administration, after he tried to send Giraud to the "STO" during the WWII : Giraud prefered to enter in the french Resistance (he was the chef-adjoint from the 7th camp du Maquis du Vercors during two years). After the war he became the fabulous glacier pilot, alone to land on the top of the Mont-Blanc with an aircraft (june 23. 1960).... and landed 53 five times on the "Mont Aiguille". He died in his bed .

The french aviation is just totally unable to do that... near of the ground.
She is also unable to do many other things that Monsieur Giraud did !

Maching Bird, teach that to your students, AF and Airbus will never teach that. They are just unable !

Thank you so much for this very refreshing and sparkling post in the middle of the despair of this modern ... AF447 crash and crew (I dare not to say "pilots" ).

mm43

Interesting discussion on "unloading" the wing - thanks Machinbird and roulishollandais.

One thing I have mentioned before is the obvious 1.5g as the PF loaded the wing shortly after taking the controls, during which the two brief bursts of SW occurred. At the top of the initial climb the wing was unloaded and 0.75g resulted as the aircraft went over the top in ballistic fashion, to be loaded to 1.15g as he pulled into the final climb and triggering the SW continuously.

The reduction in 'g' as he pulled for the second time was in direct relationship to the remaining CAS, i.e.

1.15 / 1.5 = 0.767 or,
265 * 0.767 = 203 KCAS.

That's not the real point, as it is the crossing into the stall during the 15 secs after 02:10:57, where the aircraft was "ballistic" and continued to gain height as the normal acceleration dropped to 0.8g, is where the real opportunity to change the AoA rapidly and recover airspeed existed. It wasn't to be, and the opportunity to do so became limited after 02:12:00 (IMHO).

jcjeant

Hi,

All this discussion on the roll of the AF447 is very interesting ...
Where the pilot see the banking angle (instrument) ?
Why when checking this instrument he (they) don't see the longitudinal angle (climb .. this must be plenty blue there .. with 15° and more)
Why they discuss (all 3 pilots) all the way down about keep the wings level and the speed.. but not about all the blue they must certainly seen on the instrument ?

RR_NDB

Hi,

What kind of "inputs" they received?

PS

We have all, related to their "outputs".

What BEA have today would be enough to understand crew actions?

PS2

PF acted "immediately" when was suddenly inserted in the loop. Crew actions are to be based on reliable inputs.

AS was lost due A/C resources limitations (simply not measured) during a period.

The (relevant and necessary) information PF (and crew) received was also lost?

PS3

Are current FDR able to record all required information? In order to take into account everything necessary for the analysis? (e.g. ALL pilot inputs)

Lyman

Human Factors

Blue/Brown. Neither color is reliable to show the Angle of attack, although at first, Pitch was not unreasonable. The BEA language make it unclear whether the ship climbed prior to, or after the ten degree 'plateau' right after manual control.

That is not to say the data are wrong, only that the Pilot was flying, not the autopilot. For all, did he realize LAW had degraded? The first we know the lack of NORMAL was at 2:10:22. Virtually all his experience was non ALTLAW, and STALL had been trained sporadically, without a syllabus that included High altitude upset leading to STALL. Say what you will, and the evidence is damning surely, but the fulcrum of recovery was immediate and positive reaction to manual control.

NOT ALTERNATE LAW. I must disagree with no less than PJ2. The ROLL was 'part of', not a singular challenge. We have the benefit of nearly three years of perusal. The Flying Pilot had one second. Well and good to frame the discussion based on trickle down data, and that parsed; we appear to have relinquished our initial hold on objectivity. The passage of time does not change History; it changes only opinions.

mm43. What about that 1.5g 'load'? I believe the report was 1.65g, and my question is how did it increase that high? Doesn't the Bus protect PITCH with One G commands?

Similarly, a maneuvering aircraft can't be felt in de-brief. Not accurately. ROLLING is a form of UPSET, regardless the parameters. I think that might be what Machinbird is on about. All three axes were quite active, and the a/c was in turbulence. Quoting the book, and emphasizing how 'easy it is' to recover UAS is misleading, borderline dishonest.

I will rely instead on the report of post wreck simm. For one only line pilot to have crashed '447', is too many. For several to have failed remains a benchmark for those who prefer to frame the discussion as problematic, not simplistic.

regards

BTW, re: damage and config. At least one FLAP track is referred to as "extended", in BEA's damage inventory.

It's in there. At some point, did they wish to "recover" the Rudder's authority with a F/S select?

Machinbird

Simple answer to this question. When the bank became excessive, the sensation of down became well off to the side. If the Captain was standing between the seats, holding on to head rests, he would have trouble remaining there. The wing was no longer carrying much of the aircraft's weight. The missing lift was replaced by the drag of the fuselage so that the aircraft was at a relatively stable speed (but velocity vector pointed very down).
To summarize, it was as if the fuselage had rolled partially onto its side while on the ground. Hard to maneuver inside like that. That was having an immediate effect on everyone's ability to function.

RR_NDB

Hi,

Some days ago Clandestino mentioned similarities (buffet misleading) between both cases.

Yes, there are many similarities:

1) UAS was the "trigger factor"
2) Both planes stalled
3) Crew perception failed
4) Recovery was not possible
5) Crew errors were present
6) HF factors "played important role"

The Pitot heater was the cause leading to UAS (Alt and Climb, indications) degrading the man-machine interface and misleading the crew.

Subsequently the buffet "confirmed" (a wrong perception) of "overmach".

In F-GZCP a sophisticated "interface" seems not helped the crew in their perception (perception relates to survivability). Despite (or because it?). PF put (diligently ) the plane into an stall (through persistent NU).

In CVR there is a mention to "crazy speeds". This shows they were receiving conflicting indications. From interface? From cockpit "environment", certainly: Noise (probably a different noise) and buffet (post stall).

The similarity probably is not with the meaning of "buffet misleading" interpreted as overmach buffet. In 447 the buffet, i suppose was clear after the stall (high AOA descent). In 727 the buffet (low speed stall buffet) "led to the full stall". In 447 the "misleading" Clandestino suggest, as a possible reason for "nose up" after plane was stalled. But i ask: During many minutes? From apogee to FL100 or lower?

As i understand your rationale was: Overmach buffet requires climb the plane. As an explanation for not lowering the nose. (misleaded by an overmach stall "thinking")

Is it possible? This could partially explain what crew "maintained" after apogee?

IMO there was misleading, by at least, lack of reliable information (scan failure, HF, etc.). Aggravated by Noise and high AOA buffet. In a turbulent environment (outside) and in a "stressing" cockpit (man machine interface aural and visual outputs)

A recipe for serious HF issues.

But, i agree, in both cases there are several similarities.

PS

I didn't find my post when i commented the Thiells case.

Lyman

Machinbird

That is what I was trying to say. A maneuvering a/c is disorienting.

Disoriented, on an airliner, is (can be) disastrous.

jcjeant

Hi,

I agree .. but what is the sense of vision is reduced or affected or impaired by such feelings?

Lyman

I can tell you this: Situational awareness is not overdone, ever. Once the status is lost, (forget the instruments), the odds go long. Recovery from upset is taught, and tested, ab initio. As is STALL. When the important instruments go sour, and the airplane is not recovered, quickly, the rest is guesswork, random.

It boils down to "Sit on your hands" or "recover the bank angle and PITCH."

This flight was doomed well before the STALL. At least one pilot, the one flying, was done in rapidly, he lost his grip on the situation. The second one shortly thereafter. (Else why the frantic call to Captain?). The folks who continue to claim this was business as usual are holding an illusion. The a/c needed correction, at least in the mind of the PF. Can that be argued? To what point?

The Captain had to climb up the aisle at quite an angle. One g, in the pilot's seat, is quite different than having to ascend a ramped aisle.

He entered to STALLSTALL, and a confused crew. "Er, What are you doing?"

Shortly thereafter, he lost the plot. Check that, he never had it, and it was not forthcoming.

Bend over in the crosswalk, tie your shoelace, get flattened by a truck. Same-o.

roulishollandais

I can teach you nothing : I may only answer to the french half of your question :
1. For the instrument/system mesuring vertical speed, in french, we use the common acronym VSI only in airliner, to be easely understood for maintenance papers and books. We use it also on small aircrafts to be pedantic !
2. IVSI is the instrument/system too, mostly used in gliders or by paragliders for the instantaneous vertical speed indicator with accelerometer.
3. "variomètre" is the prehistoric world ... still used for the instrument, as french aviation is still on Mermoz century, and stopped with Concorde...
Despite the word seems to refer to the meter, it may be calibrated in FT/mn as well as m/mn. (~meter refer to latin word for mesure and not to metric system)
4. French speaking pilots use "vario" or "Vs" or Vz for the name of the information showed by the system/instrument. It is shorter to say "vario" then "vitesse verticale", or "vitesse ascentionelle".

I hope a english speaking linguist/pilot will answer to the second half of your question, and despite my user name I am not dutch...

Machinbird

Perhaps not the sense of vison, but the management of priorities. If the roll is making it hard for the crew to even hold their position in the cockpit and to maintain a semblance of control, then the roll attitude takes priority over the pitch. With a nose up pitch, the crew is being cradled by their seats. With a nose down pitch, gravity is trying to take them off the front edge of their seats. If PF was not using his shoulder harness, there was nothing for him to lean into, and he would find it was easier to hold himself in the seat in a nose up attitude.

From the Captain's perspective, the aircraft generally follows the nose in all his past experience. If the nose is above the horizon, then even though the attitude is not optimum, it shouldn't be falling out of the sky and he can concentrate on trying to make sense of the instrument displays. But if the roll was excessive and if he was not properly seated, he would be almost unable to function in any useful capacity.

RR_NDB

Hi,

Machinbird:

Makes a lot of sense.

RR_NDB

Hi,

Machinbird:

Supposing the limitations are disabled: Questions:

1) Using this "approach" AF447 could be saved? If so, the maneuver could still be decided (and initiated) at what minimum FL?

2) This stall recovery technique should be incorporated in (SIM) pilots training?

Machinbird

RR_NDB, unloading the wing is used to prevent entering a stall, even when the aircraft is well below its normal level flight stall speed.

You cannot effectively teach unloading the wing in a simulator. A key element of the training is the feel of near zero g. This type training goes hand in hand with use of AOA indicators. Once a pilot understands the idea, and if the aircraft is equipped with an AOA indicator, then some Sim reinforcement of the training can be performed by pushing to lower AOA indications to near zero actual AOA, (not necessarily indicator zero.)

Once you are actually in a stall, unloading the wing should break the stall and allow acceleration provided you do not progress into a spin. The sooner AOA is reduced to un-stalled ranges, the lower the likelihood of entering a spin.

Once a pilot understands unloading, the concept of pushing the nose down to get flying again seems natural. Then it is just a matter of do you have sufficient altitude to recover? Sometimes it is a delicate balance between getting the wing flying and not generating too much sink rate. In those cases, an AOA system or a system such as the Airbus Normal law protections is what makes the difference. If your 'Bus is not in normal law though, you are back to basic Stone age flying technique.

Lyman

Mach

NORMAL LAW protections end up creating what 447 experienced. In NORMAL, the wing will not STALL, even with full back stick, held. It "merely" increases sink rate. It is a technical "NON STALL", only, for with increased sink, the result is what happened to 447. So Bus pilots are taught that the a/c will not stall, even with continued back stick. (In the Normal Law).

Is it just a trick of nomenclature? Do they believe that even with the STALLWARN, they are not STALLED?


For whatever reason, if the pilots are granted even minimal skill, they did not consider they were STALLED, in evidence in CVR. They knew they were descending, but could they have possibly believed they remained in NORMAL LAW? At the very last, they PITCHED down. They then increased PITCH before impact. On the face of it, it appears at least that they may have considered the STALL was not real?

Per your unloading description, they were 'flying', at one g, the wing is not "STALLED"? I have never experienced a one g STALL, developed, though I know you have. Had they? "How can we be STALLED, the NOSE has not dropped down?" ( No "0" g )

I suppose it is an aerodynamic exercise, help?

RR_NDB

Hi,

Machinbird:

The concept is clear: Stall is a function of weight to be lifted (by the airfoil) and beyond a certain AOA there is a fast degradation. Unload means "help" the wing (to do the lift) "reducing the weight" using physics.

Let me put both questions a little different (better):

1) Using this approach nearing the stall could avoid the stall? Sure, Yes.

2) Using this approach after the stall could make you exit? As you put, Yes. In this case with a pilot trained (to "feel" the zero G) in an airliner (similar conditions of this flight) do you think the "day could be saved"?

3) In this case at what safe minimum FL the maneuver (by the same trained pilot) should be initiated (at what FL) to present reasonable chances to recover in time. (before SL )

Obs. After a fully developed stall chances reduce fast due height loss, spin threat, noise (), etc.

Unfortunately the "zero G" feeling is impossible (for longer duration) in the SIM

RR_NDB

Hi,

Bear:

Disturbing to imagine (a training could do) this.

May match elements we have so far.

(*) trajectory with HEAVY LOADED WINGS

Machinbird

In the AF447 case, they could have recovered if they had initiated somewhere around 10,000 feet, maybe even less, but there would be no room for secondary stalls. It would have to be done perfectly, and if you don't have Normal law, then you need an AOA indicator to do the recovery perfectly. Your flight control system can screw you up if it trims the THS too slowly.

Not only that, there is also a 'cornering' speed that you would not want to exceed either. If the aircraft got into g limiting, that would result in less than ideal results.
Once flying (un-stalled), you would want to accelerate using maximum thrust up to the cornering speed and then go to near idle to avoid exceeding it.

One of our aircraft performance guys can probably come up with some realistic numbers, but I have just described the basic tactical approach to the maneuver.