The Seattle reference was to post 1747, which has a nice Seattle ring, at least imho.

That said, no argument that anytime one flies, one is exposed to all dimensions of human error, both active and passive. Still, certification is expected to weed out the most sensitive cases, if need be at the expense of efficiency.
Bearfoil makes the case that the pilot had 11 seconds to make the right call once the a/p dropped out. That is in an airplane encountering no extraordinary stresses, at least afwk, in excellent repair, in routine flight and with no serious mechanicals other than the pitot icing. That is scary, at least to me.

Bear presents the case.

He hasn't made it for unrecoverable in re the 11 or 3 seconds.
Evidentiary shortcomings to confirm causes and effects.

More light won't be shed for a while, and may never be.

Curious!
 

After I did post the ‘BUSS’ article last week, I realized that when you’ve lost all the speed signals this Back-Up Speed Scale wouldn’t work either.
So went into the books again and found out that modified ADIRU not only send its AOA through IR bus i.s.o. ADR bus (crew needs to switch of ADR part of all ADIRU’s to activate ‘BUSS’) but also the AOAc signal is always present (even if CAS < 60Kts !!!!), only the SSM remain set to NCD if CAS < 60Kts.

Aural Stall Warning:

In Normal Law: Max AOAc = 23 surface degrees.

In Direct Law (think this is a typo error in the docs and should be 'Alternate & Direct'?)

This is the sum of S/F term, a Mach term and a Mach term when Speed brakes are extended. Clean Config base angle is 10.8 and M correction is summed.

There are 4 tables given for all S/F positions and several M numbers. Too complicated to type it over and can’t make copies or duplicate any other way.

Table interpolation in clean config results in:Aural Stall warning is inhibited on ground (except in functional test mode) and above M 0.866.

If all ADR’s are switched off to engage ‘BUSS’ a backup stall warning table is used based on S/F only.
Clean Config value = 8.6 AOAc.

(‘BUSS’ was not installed in this A/C)

https://docs.google.com/leaf?id=0B0C...thkey=CILGt_QN

That looks a bit like the graph HazelNuts39 posted a few days back. :)

As I understand his intent, it wasn't BUSS info related. He was mapping Mach and AoA to show the relationship to when stall warning would go off at various Mach numbers.

Simple question: How is M number computed? As related to the discussion on Stall Warning in the preceding posts.

:confused:

"Aural Stall warning is inhibited on ground (except in functional test mode) and above M 0.866."

What caused the STALLSTALL to attenuate at PF's inputs?

Why was it not available at <60 (but airborne?)

Was PF's (initial) reaction to 'overspeed' (as in from bad airdata)? Is there Inertial record of airspeed at >.866M?

Was CL a reaction to O/S? What else could have kept her "flying" to FL380, especially after STALL at "cruise" (a/p) managed parameters? If the STALL was true at a/p drop, how could she have held that much (aero derived) ENERGY?

Some peeks at the "code"
 

Finally getting some engineering stuff versus the basic things I see in the 'bus manuals.

- We may need to define some terms.

- Smilin' Ed may help us ( a squid golden arm), but the mach corrections to Aoa look a bit harsh to me. OTOH, the plane prolly has an efficient subsonic camber and could be more susceptible to mach effects than what he and I and 'bird and Retired flew.

As several of us have observed, if AoA data is available regardless of speed, we could have a case questioning the design. I ain't going that way........ yet.

- Interesting, the use of tables. Of course, original Viper used analog circuits, so everything was "instant" and the outputs were the result of op amps and function definitions mechanized with discrete components. Outputs were basically voltages. Ten years later the system was digitized. Don't know how the digital system handled frame rates and "mixed" the outputs, but control system technology was moving along at a rapid rate.

thanks for the insight, A33

Bear, where does the aircraft get the energy to accelerate from Mach 0.82 to Mach 0.866, particularly when there are indications that they reduced to Mach 0.8 or so before things began to go wrong?
*smart arsed remark* nose down pitch? (sorry)

Bear, that sequence of events looks to conflict with recorded sequence of events. How do you arrive at "be stalled at A/P managed paramaters" there? The sequence looks more like a/p held parameters -> something -> a/p off -> (we are in alternate law) -> parameter change -> something -> stall. As noted previously, with the momentum of Mass (205 Tons) and V (upwards of Mach 0.8) you don't bleed off the kinetic energy all at once.
I think you answered your own question. IF "stall at A/P drop," THEN no climb. When you stall, you tend to fall, since you no longer have sufficient lift to defy gravity. Also, it means that "at ap drop" all of that airspeed is already gone, or most of it. That doesn't seem to fit the data provided.

Put another way ... when you are stalled, you don't have the energy to get that other 3000 feet up.


Can someone show me how to make an aircraft climb while it remains in a stall?

Caveat: Hmm, maybe there's a bunch of left over energy if you have an accelerated stall in a high G turn ... I'll need to think about that one ... but there is no evidence that AF 447 was in accel stall nor in a high g turn.

OK, I have an idea. Can I gain altitude after I stall? (Well, one way to do that is to bounce off of the pavement after I hit it. Another is to eject from the stalled jet before it hits the ground ... but none of those apply here ...)

Yes, I probably can, but not much, given the energy I bled off getting to stall in the first place.

Let's see ....

I could be going real fast, pull up on the stick, pitch the nose up, and hold that attitude, and my monemtum carries me up ... before I reach apogee, as energy bleeds like a severed femoral artery ... I stall by reaching critical AoA, and then just past ... a few more feet of "up" are the result of residual momentum from my up vector, rapidly being countered by Mother Earth's gravity.

That does not appear to be what was going on with AF 447, in terms of the altitude increase from FL 350.

Bear, you can't manipulate the data into a stall, then a climb to FL 380. Not until you dig up Bernoulli and make him rewrite that law. :p

A couple of thoughts have occurred to me (I am guilty of two things - much preferring the bus to the boeing, and instantly judging the FO when reading the BEA report):

1 Alternate law was not helpful.

When UAS is detected at high altitude, then maybe direct law would be safer. The stationary stabiliser would help the pilot keep the attitude stable. If he got slow, then he would need constant NU to hold the pitch - and the clue is there for him. Auto-stab is a nice luxury, but is best when combined with Normal Law protections.

2 The "illusion" of turbulence.

I have had nasty mountain wave twice - once in a Boeing and once in an Airbus. The event in the bus was a lift combined with a massive increase in speed. I had to disconnect, pitch up, and then dive when the speed vanished. By the end of the event I was more or less back where I started, save a slight bead of sweat or two.

Having felt that lift, might I (if I felt it again) associate it with a sudden overspeed? It's so easy to imagine UAS in smooth conditions. But what if the PF sees the speed indication go and then feel a huge lift? Pitch up because he has seen this before? Once he has pitched up, he has lost his "feel datum" because the stab is trimming for him. Still with no speed indication the confusion has set in, and there is no escape from the confusion.

It will be very interesting to see the plots. In particular G forces on the a/c and sidestick deflection. So much will be answered when these are available. In the meantime, more high altitude training, including dealing with various serious faults at altitude.

..

Incidentally, I have been shocked at the number of FOs I fly with who have no idea of approximate pitch settings in the bus. They have all gone straight from Piper to Airbus, and have never had to fly by pitch. One of our Boeing FOs was chatting to me about this accident, and he could specify all the pitches for all phases of flight (happy memories!). As others have said, the bus is great but when it all goes wrong it reverts to a basic jet. If you have never flown a basic jet, then this can be a little confusing.

MACH
 

@ GY:
Not sure if I build the formula correct, but something like this is mentioned. Couldn't find the Root sign:8

MACH = ROOT(5 * ((Qc/Po+1)^(2/7)-1))

http://upload.wikimedia.org/math/c/e...d03e6ee91f.png from wiki (thx LoneW 50)
Qc/Po = ratio of impact pressure to static pressure.

@ Bearfoil:

Maybe because pitots (Qc) where clogged with Ice?

Newton obtains. Bernoulli was a Swiss plumber.

No, the Stall was bogus, wait, it was in Alternate Law, so Stall was legit. No, she climbed with gusto after Stall, so it was bogus, no, it was below 60 knots, wait, that can't be right, etcetamundo.........

Something told the PF to climb, or not, and he inadvertently climbed to 380. Was it on a hunch?

1. Cavalry Charge

2. Master Caution

3. "I have the controls"

4. Roll left

5. Climb

6. STALLSTALL

7. Time elapses

8. PITCH (Assiette) increases "progressively"

9. After increasing to Ten Degrees, she STARTS to climb...

10. Speeds reacquire

11. PITCH 16

12. AoA 6

13. STALLSTALL

1 through 13 are reported (Observed) by BEA. In this sequence there is the Death of an a/c and 228 souls.

BEA would have us believe that the procuring cause of this accident was the Pilot's initial Climb (NU) input. If not, then they owe some more data.

Old Chinese saying: "Half the Truth is a Whole LIE...."

Quote:
Why was it not available at <60 (but airborne?)
Maybe because pitots (Qp) where clogged with Ice?



I meant the STALLSTALL. It is computed by AoA not a/s. What hath the AoA have to do with Pitots? Frozen AoA vanes? Quite the coincidence, No? Wait, there's this: The AoA vanes could not have frozen in STALL angle, not without at least a chirp. And they "Started" @ angle< STALL, so they froze at the very moment of Vane excursion into STALL angle?

Stallstall
 

@ Bear

I see, was focused on Qc, if M = 0 = no correction = STALLSTALL trigger @ 10.8 AOA.
But if probes open up again AOA STALLSTALL could be @ 3.8 AOA to trigger = intermittent STALLSTALL between AOA 10.8 and 3.8.

Didn't check the report if AOA was in this range.

hundredpercentplease

Be interesting to see if the g trace predicted by HazelNuts39 fits what comes out in the report, if reported ...

Bear: Step 14 includes gravity at work, and a variety of things attempted (I presume) that didn't work. That too would be of interest to learn from the good folks at BEA. Sometimes learning what doesn't work helps identify what does.

Garage Years:

This wiki article has an easy to understand calculation of Mach numbers in air:

Speed of sound - Wikipedia, the free encyclopedia

Start at "Practical formula for dry air" and work your way down.

Lonewolf

".....Bear: Step 14 includes gravity at work, and a variety of things attempted (I presume) that didn't work. That too would be of interest to learn from the good folks at BEA. Sometimes learning what doesn't work helps identify what does..."

Indeedy. Thusly: From the git, the authority had far more data than any civilian. Then they found stuff, and they had that much more. Then they had a powwow with the Brasiliens, and had Medical and even more stuff. Then the wreck, the boxes, ad pano.

They will always have more than anyone, that is the game. The longer this continues, the less they will have to give up. That too, is the game. Safety is less important than Control, in the game. On the Flight Deck, Safety is everything. The twain shall not meet, and that is the endgame.

A33Zab Thanks. For everything. :ok:

Valid Indications?
 

Just a reminder that sometime between 02:11:51 and 02:11:59 both PFD's had the FPV page selected. Reference to the ACARS messages will show these two sequential FPV messages timestamped 0211. They had precedence over the FLR message which was received at 02:11:55, and there would appear to have been a queue/link glitch which delayed their transmission.My point is the comments that both pilots made could be directly related to the FPV pages having no valid data, or in more general terms there was nothing on their PFDs which seemed valid.

The question should be - how is Mach computed in the A 330 ADC?
Note how temperature is in the theoretical calculations: – TAT in the ADC?

mm43

Hiya.

Quote:
At 2 h 12 min 02, the PF said "I don’t have any more indications", and the PNF said "we have no valid indications".

At this point in the recording, the pilots have known for two minutes that they are in Stink. I see the PNF's use of the word "valid" as a courteous correction to the record. (CVR) IE, "Nothing is making sense"."Je ne comprends riens."

IOW: We have indications, none of which make sense.

Possible?

Hi Bear;

Possibly the Captain asked to see the FPV page, and the PNF/PF comments confirmed no FPV and or any other valid data. I don't think their selecting the FPV pages just prior to 0212 was a coincidence - more likely related to what they said.

Yep I remember now the FPV was N/A. Thanks.

TAT Calculation
 

ADR internal calculations:

Just as I mentioned before:

http://upload.wikimedia.org/math/c/e...d03e6ee91f.png

TAT calculation is back to Elementary School: :O




The Calender-Van Dusen equation for a 500ohm platinum sensing element is used.

• RT/RO = TAT + α[TAT - δ(TAT/100-1)(TAT/100) - β(TAT/100 –1)(TAT/100) 3]

where

• TAT = temperature in °C
• RT = resistance at temperature "T"
• RO = resistance at 0 °C (500 ohms)
• α = .003832
• δ = 1.81
• β = 0.0 for T > 0.0° C and 0.1 for T<0.0 °C

Thanks A33Zab and Lonewolf 50:

Ok, so to compute Mach speed, which is a conditioning value for the Stall Warning, we need dynamic pressure which we know (believe) was lost (for some duration) due to pitot icing. So the Stall Warning would have done what? Defaulted to the low speed limit? That doesn't seem smart.... this whole issue has been whirling around my head for a few days. Can anyone shed some light on what the heck the Stall Warning system does once the speed input becomes unreliable? Agreed the system would recover with the return of the dynamic pressure.

A33Zab, so there could be Mach problems (amongst others) with a blocked pitot probe, but what if TAT rose erroneously to zero coincidently or even before the pitot problem – due to TAT probe blocking – a feature of ice crystal encounters?

Ah, Machinbird, I'd separated the roll and the pull up as two different problems. The roll correction should be automatic. The pull up was wrong. So I suppose "do nothing at all" is probably wrong. Correct roll once you recognize it is real and then do nothing about speed.

And, of course, doing nothing about communications with DAKAR or Brazil trying to reach the plane without using selcal is also the wrong thing.

All that said, if the roll was not dramatic, why not wait a half dozen or dozen seconds to see what happens? The ride for the SLF might be a little rougher. But, they might have lived through it. Diagnose then do. "Don't just do something (reaction), sit there( think), then do (action)."

It will be interesting to find out how fast the plane rolled when the AP kicked out. Was it urgent? Was it slow enough to ask why then act? That much tonnage seems like it'd take quite awhile to develop a really sincere roll or pitch change absent strong control surface actions. Is my intuition here right?

Aviate, navigate, communicate, comes to mind. How did they aviate to a 10 degree + nose up attitude at FL350? Did it have something to do with the Airbus can not stall teachings?

Thing of it is, according to BEA, the a/c rotated progressively up to more than ten degrees NU, and "then started to climb." There's a real poser, eh? It takes 10 degrees NU to initiate Climb? There's a sluggo. From 0.80M ?

Bearfoil,

Isn't that the whole problem with that damn stupid report. Essentially hundreds of pages of posts that still haven't actually understood what that repoort is actually reporting. Why such a vague description of the events? What nonesense is this :ugh:

He is probably referring to a different accident.
In this one it took more than 1 minute to deeply stall the aircraft (Alpha >>10°) and there is no indication it was unrecoverable.

Hi bubbers44,
I think so. Maybe it is over "confidence" in FBW to the exclusion of intelligent flying technique? e.g. the QRH Windshear procedure has two references to the use of "full backstick".

"Airborne, initial climb or landing
Thrust levers ... TOGA
...
SRS orders ...follow
This includes the use of full backstick if demanded.
Note2. If the F/D bars are not available, use an intitial pitch attitude of up to 17.5 degs, with full back stick if necessary."

If they mistakenly diagnosed their problem as one of CB "windshear", their actions could be explained by performing that particular QRH procedure. After the stall warning ceased, maybe they were just waiting to get out of the "downdraft".

It was a note (not a report, to be published before the end of july) released under the pressure of leaks, speculations, and of Airbus (which wanted the preliminary data to be published before the Bourget air show). Even if uncomplete, we now know who was the PF/PNF in the cockpit, what were the crew reactions, and how the A/C stalled. Now we have to understand why these multiple NU inputs in the presence of stall warnings (in many of the past Pitot incidents with early stall warnings, ND inputs put the A/C in descent. None of the crew attempted to set a CLB/5°). Here:

-In a first sequence (from the Pitot signature at 2 h 10 min 05 to 2 h 10 min 50), the A/C climbed very rapidly to FL 375 but was not yet stalled (AoA~4°). Whether this climb was voluntary or not is not clear, even if the PF expressed regrets during the brieffing for not being able to climb due to the temperature profile, there were NU inputs as soon as the PF got the controls, even before the 1st stall alarms, then ND orders to stabilize the trajectory at the FL 375.

-The following 50 sec sequence between 2 h 10 min 50 and 2 h 11 min 40 (CPT reentered the cockpit), beginning with stall warnings, and then TO/GA + maintained NU inputs, effectivelly stalled the A/C (AoA=pitch=16° @ FL380). I am puzzled by these maintained NU orders by the PF (~800 hrs on A330 type) when stall warnings sounded for the 2nd time, but even more by the fact that the PNF (~4500 hrs on A330 type) did not react to the increasing pitch 2.5°->4°->10°->16° at such a flight level (over FL 375), leading to the stall. The BEA note says that during this phase the PNF tried several times to call the Captain back into the cockpit: was his attention devoted to help aviate/navigate or to communicate with the cabin crew ?

bear - do take another look at the A340 AIRPROX zoom-climb Appendix B FDR plots, there is a 10 second lag between pitch up and establishment of climb. There's even a hint of the classic "non-minimum phase" control response - up elevator pushes the tail and the whole a/c down slightly before the lift/thrust vectors and Newton's second law build vertical velocity.

[NB no UAS different laws/protections active in the A340 vs AF447, not too clear what commanded the pitch up, lots of elevator action with A/P off controlling G? Is pilot sidestick trace inverted? Who/what adjusting throttle with A/T off?]

I'm sorry but if I induce a 10deg NU input on an aircraft flying M.8 the climb rate will come on pretty much as the input is applied. We simply don't know enough to ascertain (data wise) to make anything sensible out of that statement by the BEA. For all I know it can just be lost in translation.

The only reason why I can see an experienced Captain/Crew maintaining a NU input with stall warning going off is that they believed that the aircraft could not be stalling (it's an Airbus right?) and therefore deduced that the warning must be erroneous.

Mode and Law confusion are the most common things that pop up in the simulator:
'Wtf is it doing now'' seems to be a common phrase.

What Garage Years has been asking for a while now is the value of the critical AoA (AoAc) which is used by the confusers to trigger the stall alarm in ALT law when airspeed (including Mach number) is unreliable.

Unfortunately, it is not documented in the A330 FCOM. Only those who have access to technical documents not publicly available could answer the query. Others and I can only speculate hopefully trying to use sound logic.

The Air-Caraibe incident report mentioned a critical AoA of 4.2 deg. Cross-checking with the data just posted by A33Zab, this corresponds exactly to M 0.80, the recommended turbulence Mach number i.e. the speed at which the ACA A330 was supposed to fly at the time. It don’t see here a confirmation that it was the AoAc really used but just an assumption.

Obviously, confusers cannot just cancel stall warnings on the basis that Mach number is unreliable. They have to know what to do in this case but different solutions are possible like:

- Use the last Mach value considered valid.
- Use the highest current indicated Mach among the 3 sources.
- Use a combination of the above.
- Use MMO instead.
- Your choice …

None of these possible solutions are ideal.

No it doesn't take 10 degrees nose up to initiate a climb but if you applied a large aggressive input with less aerodynamic damping (at high alt) then you would reach 10 degrees very quickly. There's a-lot of momentum to overcome before the climb begins and I would say that the BEA comment tells us about how the controls were apllied and the rate of change of attitude that was achieved as a consequence.

DJ77:
Well the stall warning -is- cancelled if the sensed speed goes below 60Kts and then comes back when > 60Kts so maybe not so obvious to the designers that cancelling an in progress stall warning is suboptimal.

This snip from the report, well into the descent with the plane undeniably stalled, is the second instance the stall warning coming and going with invaled speeds could have further confused the pilots:

The report then mentions dual sidestick inputs at that point but frustratingly does not specify if they match or in what direction.

Murphy:
 

The dual sidestick input(s) are bracketed time-wise with Prim 1 fail or switch off, and Sec 1 fail or switch off. Where are the switches situated to switch off/isolate/disable these units? I recall somewhere that dual SS use has to be enabled to allow summing of stick movement. Could the movement be the result of a hand finding it's way back to the SS?

Gums, if you are around: when this thread is all wrapped up and shifted to the dusty vaults, don't wander back to your Sipping Whiskey and fishing pole- we need a memoir writing!

Minor request :
With all the quibbling about words like "assister", and 'around', I wanted to check back to the original French, but, with well over 150 pages of threads, I've now lost the link to the original of the latest BEA note (in French).

Could somebody post it again? Thanks in advance!

I think this is what you are asking for:

http://www.bea.aero/fr/enquetes/vol....mai2011.fr.pdf

Weeds round the prop,

What you recollect and are asking for can be found here, Pages, 35,36 & 37

A330 Flight Controls

Turbine D,
That's it. Thanks!

I believe the 30 second time grab with the a/p disconnect square in the middle would give it up. The Captain has said a climb is not available (I think they were cleared for an eventual Cruise @ FL370). Warm Air. He has also claimed the a/c will (in two minutes time), enter increased turbulence. He is not making small talk. Someone wants a left turn. With an aft load of Trim fuel, one expects at least a slight THS ND to load the HS to lessen Drag (oa). But the THS at handoff is 3 degrees NU.

Warm Air, sink, probably ND (descent)? No big, the Trim is correct, until, Pitots supply discrepant Pressures to AD(s). The autopilot gets out of whack, perhaps attempting a renew of Trim (airframe), and in doing so can't keep up, and for whatever reason, it is disconnected. Now it isn't that important how the Pitots started to read in disagreement, let's assume ICE, though that is not necessary, imo. ICE is patent in UAS, so we go with it. How much has the nose dropped since the beginning of the Temperature change? Unknown, but why would it be less? The a/p (and a/thr) are about trim, so let's say increase in Thrust and a notch more of NU.

The a/p quits, Cavalry Charge, and Master Caution. "I have the controls" says PF, and so he does, for he (right away) inputs NU and Left ROLL, one assumes from what he sees on the FD? Now this is a big a/c, and it is Nose Down, accelerating, and may have excess energy, even too much Thrust. At handoff, does the a/c remain at current Trim? That is my understanding, so if the Pilot has to correct, the a/p was unable anyway. So far so good. We don't know if the a/p had a chance to trim speed back to .80M so may be the PF has to think about that at this point also. He just got the stick, and he hasn't hand flown for awhile, and perhaps never after an emergency loss of Autopilot. He's a little aggressive, but quickly relearns a soft hand. Fighting this soft hand might be indicated, though, for he needs to make some corrections to regain the Plan. Did he overshoot Pitch? Was the THS quick enough to retrim the Tail?

If PF maintained NUtrim to slow, recapture altitude 'lost' in the temperature exchange and possible variable and substantial airstream fluctuations (vertical), then to me he's got it right, and the a/c will catch up.

"What's it doing now"? I don't give this sa remark much credence, and I never did. It is, like Dozy said, almost certainly doing what it is programmed to do.
How often is this type exposed to the combination of factors present for this upset? Computers don't 'forget', and that is a good thing. Neither do they 'think'.

The last thing on Earth anyone wants is to not understand the a/c, maybe worse is to misunderstand the aircraft. At least in confusion lies inaction?

For me, I'd like some more data, especially what happened in those thirty seconds. I think the Pilot turned left, and the aircraft turned right. (Metaphorically).