RegDep

http://libraryonline.erau.edu/online...s/AAR75-13.pdf for Welsh Wingman

BOAC

- no bigger a pill than usual There I disagree - it IS there to protect from 'bad piloting as such' and it does it very well in normal circumstances - things like excessive pitch, too high an AoA, too much 'g' etc etc.
- you need to remember that normal 'manual flying' in an Airbus is still 'protected' by the system. Thus they do NOT need to know the limits of safe flying, even manually, as the aircraft controls those.
- no. That was EXACTLY my point!

FE Hoppy

Welsh Wingman:
Flugunfall 01 DEC 1974 einer Boeing 727-251 N274US - Stony Point, NY

Litebulbs:
Not sure what you're asking? The indicated airspeed whether accurate or not isn't the issue in determining the energy state of the aircraft. We know what the airspeed was before the incident. We know that the crew are not reported to have added thrust and that the AT was disconnected. You can't climb for free so the energy came from speed. They stopped climbing when they were at the ceiling for their thrust setting at the alpha they were commanding. Note that when TOGA was selected they were able to climb a further 500ft. In this case they added energy to the equation.

Tailspin Turtle

Note that skydivers, before they open their parachutes, have a glide ratio at a very high angle of attack, just not a very good one. With a fabric "wing," the glide ratio improves. Click for an example.

Similarly, the wing is still providing "lift" beyond the stalling angle of attack, just less and less of what it was before, and more and more drag. The airplane should pitch down at stall because the center of gravity is ahead of the center of lift but in this case, the center of gravity may not have been been very far ahead of the center of lift and you've obviously got engine thrust, an almost fully nose-up trimmed horizontal stabilizer (which is probably sized to keep the cg range as wide as possible), and nose-up elevator creating enough nose-up pitching moment to offset any nose-down moment. I am surprised that there was still enough roll control authority at that angle of attack to keep the wings more or less level.

FE Hoppy

Tailspin Turtle wrote:
Not really?

CoG could be some way behind the "centre of lift". In this case the CoG was at 29% MAC. The "centre of lift" would generally be at 25%.

Lonewolf_50

Graybeard, to get the AoA value you suggest, would not the aircraft have to be moving backwards?

See the diagram here.

Angle of Attack Indicator

If you move (orange) line (marked number three) down, you increase the the angle. To get to the value you assert, the orange line, resultant relative wind, would have to be coming from behind the aircraft's nose, behind and below the leading edge of the wing. That would need for the plane to be dropping not just straight down, but straight down and roughly moving in the direction (with nose up 16 degrees) that the tail was pointing. If it were doing that (and I doubt it was) the ailerons would not provide lift. The flight path described (see takata's picture some pages back) by the data BEA provides, indicates a long, wide, right turn (almost a teardrop) from apex of climb to impact with the water. Nowhere does the BEA report indicate the rapid change of heading associated with a spin.

So if it didn't spin, but was stalled, and was in a rate of turn that described the track over the ground from that reconstructed picture, the aircraft was moving in the direction that the nose, not the tail, was pointing. More to the point, your estimate would require the aircraft (whose ground speed was reported as 107 knots) experience a massive microburst of airflow to hold that AoA (relative wind from behind) and 107 knots forward.

Your AoA estimate there does not fit what has been reported.

Can you sketch out why you come up with that estimate, or were you tossing that against the wall to see if it would stick?

Litebulbs

I am confused, that's all. At 2:10:05 and M0.8 there was a L Up input with a warning and a massive drop off in IAS. At 2:10:16 a climb started, increased to 7000fpm then reduced to 700fpm with an increase in ALT of 2500ft by 2:10:50 at M0.68.

So in 34 seconds at 7000fpm you could achieve a 4000ft increase in ALT. But surely you have to start the climb and then reduce it to 700fpm too, all within 34 seconds, whilst still having a forward speed of M0.68 and all from all from energy contained within the airframe.

grity

after creating the "zoom-boom scenario" up to FL400+ to bring down the kinetic energie of 205 t with 240 m/s to an flightlevel zero without a highspeedchrash.... within 4-5 min... within less than 10 miles

we all thought: yes it seems physical possible, BUT it can not be because no pilot will pull in this situation....

and now we know the way, it need only a pull at the elevator a bit over some time and the right or wrong (too-loose-)lav for the follow of the trim to the max position and with a 270 deg segment of circle a climb to FL380 is enough to bring the kinetic energie to FL zero... BUT by all RAZORs I can not believe that the PF has pulled up over 3 min with his hand, pulled with all skills he had in his hands

you have more sensors in one hand than all airplanes together have in all ther highly sophisticated technic,

even if the PF sits his first day on the left side and had used the stick only with his right hand on the rhight side before, he will not pull so long..... no no no

so are we shure we can roll out all other possibility reasons for the up move of the trim?

icing AoA vane ???
the g-sensore
is the trim lav absolutly save, in all moments during the disconnect of the AP while changing the lavs???
something else.......I do not know

if an actuator like the auto-trim has a faulty circuit it either stops at the last position or it run to the end common with low energy an actuator can run to the end(my first RC-modell crashed in first flight, the elevator moved full down, I just plump forgott to recharge the akku before flight....)I also did not understand the climb of the A340 after disconect the AP, they say nothing over the trim, the say the flightpath followed a g-protection in this case, the path will oscillate in altitude, my feeling is even this jet was nearly before stalling and the protection was the PF ???

if the g-protection work over the elevator will the trim than beginn to follow also?

work the g-protection together with the airspeed? or the AoA?

Lazerdog

RegDep.... Things haven't changed much since the '70's have they? As I posted several dozen pages back, an aircraft can "feel" pretty normal in a stall on a dark and bumpy night. You have to look for clues like the altimeter unreeling to know what's happening if airspeed, FPV, and AOA are suspect or missing. Thanks for posting the report. Hopefully the world's instructor pilots are reading.....

FE Hoppy

Well the pitch input was at 05 and the climb commenced at 16. The peak value was 7000fpm but the mean was about 1400fpm. If we are to believe the the quoted speeds they lost 60kts in this manoeuvre and ended with the same thrust setting and 4°alpha.

How about this scenario:
PF makes his correction for roll and doesn't notice the rate of climb or is confused by the stall warning. When he notices the 7000fpm he corrects with nose down and levels at 375 all without adding thrust. He's now at 4°alpha but having not added thrust his speed is reducing which induces the second stall warning. Now he adds the thrust and pulls up. The aircraft doesn't have much more thrust to give and his alpha is increasing all the time while the speed is reducing. We know that the isis speed comes back at 2:11:06 (15 seconds later ) and shows 185 which matches the PFD.

Doesn't appear unreasonable and doesn't require any external energy input.

Welsh Wingman

Very grateful re: your link to the NTSB report on NW6231. That's the flight I have always thought of when you mix pitot tube icing and aerodynanic stall. Probably a generational thing. Showing my vintage.

But for the BEA, and from the stance of investigating the loss of AF447, I still think TE901 and have done since my very first skim of their 27 May report. Why would pilots continue to pitch-up the nose on a stalled aircraft from 38,000ft all the way down? It's been done all the way down before, see above, but other factors will surely come into play in the full CVR and FDR (with details of the aircraft and its systems responses).

As initially inexplicable as flying a perfectly good aircraft into a 12,500ft volcano at 1500ft feet? Eventually, you get to the truth - ANZ flight crews unaware of the MSA or else ignoring them on "sightseeing" flights, McMurdo ATC giving permission to descend well below any ANZ SOP MSA (the radar not picking-up the plane/interrupted HF communications as clues to an intervening land mass?), no Antarctica "whiteout" training for the flight crews, and then the plane being programmed the night before to fly towards Mt Erebus rather than down McMurdo Sound (as per the earlier flight crew briefing) and without the flight crew even being notified.

The flight crew doing something which appears inexplicable at first sight, training and operating issues coming out "in the wash" (haven't Airbus already changed their recommended high altitude stall procedures, focussing less on power application and more on a willingness to sacrifice height and get air flowing over the wings again, in the intervening period between crash and CVR/FDR recovery?) - sound familiar to AF447? Intense media activity (albeit nothing as compared to the pressure that McDonnell Douglas were under in November 1979 in relation to their DC-10s, post-Paris and Chicago crashes - I bet the engineer who downloaded all the INS data from the wreckage to prove a controlled flight into terrain got one hell of an Xmas bonus that year!). BEA may additionally have aircraft issues to add to the mix with AF447 (there are a number of troubling issues, particularly in the short time between A/P cut-out and the stall beginning at FLT380).

Alot of speculation in the meantime, on still limited known facts from the CVR/FDR, but some excellent technical input from this website. Just don't get overly fixated with that last sad few minutes (I know it goes against the grain for us aviators), because most of the chain was completed and in place before the A/P computed "do not compute" and handed the plane over to the flight crew. Think how small a trigger, the ice-defective FOHE grill on the Trent 800 engines, was to bring down a B777 on final approach to LHR. The margins are tiny.

And a salutary reminder to today's aviators, airlines and manufacturers of what the great Captain D P Davies wrote (well, more or less) over 4 decades ago - if you are left with a choice of causing an aerodynamic stall or causing anything else, probably best to go with the latter...........

That's all gentlemen (and probably a few ladies), from me, until after the next BEA report. Keep up the good work/keep the posts coming.

We have travelled less since December 1974 than everyone had hoped (the Colgan Air propliner, not just AF447, and even with the Aeroperu and Birgenair "wake up" calls in the 1990s).

Tailspin Turtle

Thanks. My bad again. I should have written neutral point instead of center of lift.

gums

1) Whew! Glad some folks corrected/confirmed my geometry concerning flight path and AoA.

Every airplane since the one I checked out in back in early 1904 ( Wilbur was my IP), has depended upon the shape of the wing and the AoA ( relative wind over the wing) to get airborne and stay there.

Why is AoA not a basis for all thoughts here concerning the ability of an airplane to fly?

2) I initially homed in on a classic, uncontrollable "deep stall" scenario due to my experience in the Viper. After seeing many charts and graphs and verbiage concerning the 'bus, I backed off.

Then I learned about the actual flight path of the jet and some limited tidbits concerning pilot inputs. I do not think the jet was in an unrecoverable stall. It was simply stalled!

Unlike our little jet, which could find its way to a true deep stall that was beyond human or aerodynamic capacity to overcome, the 'bus does not seem capable of "getting there". It has to be "held" there - by the pilot, the basic control law implementation, or a combination.

The role of the THS in this accident will become a major finding. That's my story, and I'm stickin' to it.

How did the THS get to max "up" position? Well, the jet has a gee command for pitch, not attitude or AoA. So if I hold a gee command greater than 1 gee, the THS moves to allow the elevators to behave "normally". Pull back, go up. Push forward, go down. Pull back a lot, go down quickly.

Somehow the THS remained in the max position for a nose up command. Cause is to be determined.

3) Philosophy.

The FBW systems and many other schemes are designed to a) protect the plane or b) protect the "cargo".

In my case, the design philosophy was not to protect the plane!! We could have had a plane with a 15 gee capability, an effective AoA of 60 degrees, etc. But what we got was a plane that would roll and pitch and turn at a very effective limit that no other plane of its time could equal. We could pull as hard as we could, but the jet would only reach the AoA and gee limits. We could command max roll at any condition and the "system" would limit that command in order to keep the pointy end forward. So we had a system that protected the pilot from doing stupid things, but perform better than any adversary we'd be likely to face.

I see a combination of protection with the 'bus. It tries to protect the plane from over-gee, max speed, max roll angle, max pitch attitude, etc. It also protects the pilot from demanding more of the plane than it is capable of delivering - to a point.

So we now come to unique conditions at FL350 and pilot inputs that meet conflicting "laws", and the "laws" themselves.

So is the problem training, airmanship, inappropriate control laws for "abnormal" sensor data, ad infinitum?. And the beat goes on.......
______________________

I am disappointed about the lack of understanding here concerning AoA and stall characteristics. Stall recovery and recognition of stalls are paramount capabilities all pilots must master.

DJ77

Citing FCOM 1.27.20 P2:
“When angle of attack protection is active, THS is limited between setting at entry in protection and 2° nose down (i.e. further nose up trim cannot be applied).”

This makes sense since AoA protection is to prevent stall and doesn’t need more nose up trim.

Question is: Why is THS allowed unrestricted nose up move in ALT law when the stall alarm is active? Is there a logical answer to this?

Had the AoA parameter value that triggered the stall alarm simultaneously preclude more nose up THS, the AoA would probably not have reached 30 + deg, whatever the pilot was trying to do. Elevator alone efficiency is limited .

spagiola

FE Hoppy: There's no reason to think the climb didn't start shortly after the nose up input just after 05. A big weakness of the narrative format that the BEA adopted for this report is that it makes it hard to tell when the events being described are basically simultaneous and when they are sequential, and you really have to be alert to subtle tense differences (eg TOGA was set at 2:10:51, but the throttles are already at idle at 2:12:02, so we don't know when in the intervening period power was reduced).

Other than that, your scenario is not implausible. And could it be that they attributed to turbulence the upward acceleration that they must have felt?

So this puts us at the top of the climb, at the edge of the stall (or at its beginning). What is much harder to understand, though, is the following three-some minutes. Other than power being reduced sometime before 2:12:02 and some nose-down inputs at 2:12:02+15, there seems to have been little effort to break the stall. What was going on in that cockpit?

bearfoil

Lonewolf50

As you point out, the g trace will be interesting. The first input (?) by the "new" Pilot Flying, was NU and roll left. Accompanied by two Stall alerts, this is an a/c at its maximum energy and densest air (for the climb). After this input, everything is bleeding off. Except altitude. We see ROC that varies, and no specific time points for each value. The duration of the climb is thirty seconds which must include an entry, max, and end of 700, with a third Stall Alert that means business.

I am picturing a climb with a very emphatic Elevator input, to initiate a g loaded Stall warning, and an ensuing rise in AoA that is remarkable. I also entertain the thought that some damage to the airframe may have occurred, or damage to actuators, something that made recovery more difficult, and may even have prevented ND. From Machaca's picture of the inside of the tail cone, and some familiarity with the mechanicals inside, the THS is a massive beast. The elevators by comparison look like tacked on tabs.

None of this conjecture will command attention unless and until further data is released.

The system by definition allowed this command, for it tracked it with THS inputs (NU). So a discussion of LAW here is important, but perhaps not relevant, for here the evidence is the salient issue, not the LAW, and what the a/c "would be doing". (Paraphrasing DJ77.....)

Seems to me, regardless of the controls 'domain', this a/c trimmed for Stall, and held it. IF PE, not too very well protected from the PF after all.

Even in Direct Law, or mechanical, this is possible?

Addendum. Since Takata says the ACARS stream can happen all at once, (but reported in sequence, and spaced out) could this radical climb have initiated the Cabin Advisory Vertical Speed? I still do not grok this ACARS system........

Tailspin Turtle

About a decade or so ago, I was involved in the substitution of a digital engine fuel control for the old but very reliable hydro-mechanical control on a single (turboshaft) engine helicopter. In order to simplify it and reduce cost, it was in effect, a single-channel system that failed to a manual fuel control. In other words, when the electronics went off line, the pilot had to twist the throttle to maintain rotor rpm just like in a piston-engine helicopter instead of the fuel control maintaining a constant rotor rpm based on the pilot's control inputs.

Our test pilots accepted it, the FAA evaluated and certificated it, the training academy developed a training curriculum for it, and out the door to operators it went.

Unfortunately, the unit was initially a lot less reliable than we thought it would be and when Joe/Jane average pilot had to take over and twist the throttle, he or she often proved inadequate to the task.

After the first couple of crashes following reversion to manual mode were reported, I arranged to fly one for 30 minutes or so and didn't understand what their problem was. My first helicopter flight, like all older helicopter pilots, was in one powered by a piston engine and manually maintaining rotor rpm in takeoff, cruise, and landing was not a big deal even though I wasn't current and certainly not as good as I used to be (although probably never was).

So for a while we tweaked training, improved the unit's reliability, published advisory articles, etc. Nevertheless, the crashes continued. We were expecting too much from a generation who were raised on turbine-powered helicopters and in at least a few cases, older pilots who hadn't twisted a throttle to maintain rotor rpm in a long time.

DozyWannabe

Hi BOAC,

Apologies if that sounded facetious, it was intended in a friendly manner, I assure you.

I don't know if I'd read it that way - the protections are there to prevent the aircraft from getting into a dangerous situation, whether that be pilot-initiated or otherwise.

As I said before, it's a very emotive issue for pilots and the narrative of the dastardly French getting around the table and saying "zese pilots, zey are dangerous, so we will build an aircraft zat will nursemaid zem and tell zem 'ow to do zeir jobs" is a sadly prevalent one, despite it having very little basis in fact.

Pilot mistakes are but one facet of what the protections were put in place for. From my perspective the whole philosophy centred on the fact that we now had space-age technology at our disposal, but our aircraft were designed around an ergonomics system that was largely unchanged since the war years. You put an astronaut in an A320 flight deck and I'm sure he'll feel right at home. As such, it may be helpful to think of the Airbus FBW philosophy not so much as taking authority away from pilots, but trusting line pilots with the kind of technology used to fly to the moon.

First sentence, absolutely. Second sentence, I couldn't disagree more. If the aircraft was capable of remaining in Normal Law indefinitely and that was proven, then whoever tries to put that notion forward might have a case for that - but the fact is that it isn't and therefore the limits of safe flying *must* be included in training. IMO any airline that does not include coping with FBW/FMC failure modes in their recurrent training is playing Russian Roulette with the lives of their crews and their passengers.

From my SLF point of view, I wouldn't want to see wholly-automatic pilotless airliners even considered until they are entirely composed of failsafe components - which means as long as they are built of material that can fail, have engines that are vulnerable to bird strikes and FOD and have sensors that can be blocked or otherwise rendered ineffective, I want a human being upfront who can diagnose the problem, come up with a solution and get us down safely.

Litebulbs

Ok, accepted and thanks. Next question if I may? At this point would it be acceptable to assume that the ground speed would be about 400kts?

FE Hoppy

Using the back of a fag packet (are we still allowed to do that?) TAS at FL380 for IAS 185 would be about 325kts. I don't know what the wind was so add or subtract as required.