@ Dozy
you misunderstood my post, which is probably my fault.
I'm with you in most of your posts in this thread here. My remark was pointed at those, who are either unable or unwilling to understand, that only in direct law the elevator deflection corresponds to the stick deflection, and that it is an basic principle of FBW C* law of AB (at least that's how i see it) that the pitch command of the stick represents an g command blended by pitch rate when applicable.
Therefore it is normal, that the pitch rate change of the increasing thrust enters the equation as just that, a pitch change. If the pilot commanded pitch change corrected by the protections is less than the actual pitch change by thrust, then the elevators will command a pitch reduction despite the stick position.

By the way, a pilot in the loop in a conventional aircraft in such a pitch up situation due to increasing engine thrust would react how (without available protections)? By controlling the pitchup by use of proper stick input, which might be stick forward!
Who wants to end like the B737 at Kazan?

Much appreciated - I know we've crossed swords in the past, so maybe I was also a little quick to rush to judgement.


Understood. However in the High AoA protection scenario, back-stick commands maximum or best possible alpha, based on the aircraft's previous, current and projected (via deltas/trends) state. If the current pitch attitude already corresponds with maximum/optimum AoA based on the current/projected state, then the command is to *maintain*, rather than change, that optimum attitude. So, as you say, if the EFCS detects even the slightest pitch-up tendency - e.g. from the gradual thrust increase - then it must counteract that, and because Alpha Floor was unavailable, the only aspect it could control to counter it was pitch.

I get your point, though I would say that there's no guarantee that a pilot *would* do that, no matter how "in the loop" or otherwise. It's worth remembering that the A320, while considered a very nimble airliner, is still nevertheless an airliner. The whole point of using digital technology in civil flight control systems is that they can, and do, detect and correct minute changes (as in fractions of a degree/knot etc.) in aircraft orientation more quickly and accurately than a human can. This is not because they're capable of outperforming humans in general, but because unlike a human pilot that is all they are designed to do.

In a pre-FBW fighter or a specialist aerobatic aircraft, a pilot will be trained and given experience in recognising responses and making those split-second
decisions/inputs - and because those aircraft are designed to be lightweight and nimble above all else, momentum is less of a problem. Even smaller short/medium-haul airliners like the A320 are still big and unwieldy aircraft by comparison - and this is the reason high AoA protections were a very important thing to demonstrate in the early days, being a scenario no previous type could pull off safely. I'd be prepared to bet that if you asked any one of Yeager, Armstrong, Cunningham or Beamont et al. to get in a conventional jetliner and hold 15 degrees of AoA precisely while flying slowly (as in near stall speed) at 1000ft, let alone 100ft, their refusals would have been both emphatic and unprintable.

The flight envelope protections afforded by modern FBW systems have been consistently characterised by the doubters as being primarily a technological vote-of-no-confidence in piloting skills, and this has been a very difficult idea to dispel amongst pilots. Again at the risk of repeating myself, the overarching premise is nothing of the sort - the main impetus is in allowing pilots to have confidence in making positive maneouvres without having to worry about the aircraft's ability to stay in the air.

Dozy,
To answer your recent question: yes, the (Habsheim) a/c was in a normal landing configuration, i.e., L/G down and "Flaps 3" (Slats 22 deg, Flaps 20 deg). BTW, the other normal landing config is "Flaps Full" (Slats 27 deg, Flaps 35).

Quote from RetiredF4:
"By the way, a pilot in the loop in a conventional aircraft in such a pitch up situation due to increasing engine thrust would react how (without available protections)? By controlling the pitchup by use of proper stick input, which might be stick forward!
"Who wants to end like the B737 at Kazan?"

Quite so, and - IIRC - there's a precedent on the A310. I imagine most twin-engined airliners with underslung engines have comparable characteristics in the go-around; partly because they are essentially overpowered when both donks are running, and it is normal practice to use TOGA thrust when going-around from a low height.

On the A310, which for this discussion we can regard as being in Direct Law, the amount of down-elevator to counter the pitch-couple effect of TOGA is cosiderable, even from Vref. This is highlighted best if the AP is off, when the PF has to exert an almost unsustainable forward force on the control column, manually trimming the THS forward from the landing setting as fast as possible to relieve the elevator load. (When in use, the AP and auto-trim obviously have the same task.) The down-elevator requirement would be much greater from Valpha-max. (BTW, the A310 also has an Alpha-Floor mode in A/THR, so the latter must have been demonstrated for certification, presumably at an aft CG.)

FWIW, I don't remember ever doing a G/A on an A320 in Direct Law (certainly not in anger, because in 14 years I never lost Normal Law). However, a glance at the relative profiles of the A320 and A310 suggest that the A320 elevators may enjoy a more effective pitch-moment than the A310's. (The A319 might be more comparable to the A310 in that respect, and the A320 more like the A300-600.)

In which case (and I'm looking for guidance here), does this not render the question of why the aircraft did not attain an AoA of 17.5 degrees somewhat moot?

He actually says 'in this configuration'. Do we know which configuration that is?

Dozy,

As you know, I'm not the best one to answer that, but I've no reason to doubt the BEA figure of Alpha-max = 17.5 deg, which is specific to Flaps 3.


AFAIK, the AoA is measured in relation to a specific part of the wing chord, ignoring the varying positions (angles) of the slats and flaps.

He says "..which is the landing configuration" at about 09:40.

Thanks for providing that info. What has tickled my grey cells is the "gear down" distinction. While not directly related to the calculation, would the increased drag from the gear require a reduction in maximum AoA in the protected mode to compensate?

Hi HN39,
Re the flap settings. Regardless of briefings and configuration commands from PF to PNF, the DFDR shows that the fly-past was flown at Flaps/Config 3.

BEA Rapport Finale, Annexe VII, Tome 2
Published DFDR data start at 1244Z, Time mark 240.0 sec (this may be 240 secs from selection of TOGA at the start of T/O).
At 240.0 sec, Slats18/Flaps10 (= "Flaps 1+F", maintained from T/O).
At 263.0 sec, Slats22/Flaps15 (= "Flaps 2").
At 279.0 sec, Slats22/Flaps20 (= "Flaps 3").
At impact (334.0 or 335.0), Slats22/Flaps20 ("Flaps 3").

Dozy,
Re possible effect of L/G on alpha-max: again, I think the BEA would have done their homework. L/G increses drag, which offsets thrust, but I don't think the aerodynamics are changed significantly. For example, alpha-floor remains the same at 15.0 deg in that config, according to my contemporary FCOM, which unfortunately gives no equivalent figures for alpha-max. Placarded VLS (on the ASI) does not change, IIRC, with gear position, whereas it does with speedbrake extension.

PS
HN 39, I now see your exchanges with Dozy refer to the video of the Gordon Corps demo flight...

I'd expect so too - however at the same time I must take into account that there was in all likelihood no single person more knowledgeable about the A320 systems specifics at the time than Gordon Corps - and from what I've heard about the man I doubt very much that he'd have allowed that demonstration video to be released with a verbal mistake on his part.

To be somewhat brutal, the only person I've known to make a big deal about the 17.5 degree figure is CONF iture on these forums. The BEA alluded obliquely to the figure but did not question the systems' behavioural integrity, however at the same time neither the SNPL nor Capt. Asseline's legal team (who had access to the report and those figures) questioned it during the court case. If this apparent 2.5 degree discrepancy was the 'smoking gun' CONF claims it is then why was it not brought up at the time?

The way I see it there are two main possible conclusions - either Capt. Corps was mistaken when narrating the video, or the BEA included the 17.5 degree figure in their report as background information only, not intended for comparison with the aircraft's actual AoA during the sequence.

Sorry Dozy, nothing I previously posted this evening refered to the Gordon Corps video, which I had not watched. I hadn't realised you and HN39 were refering to it in your exchanges about the landing config.

Have now watched the video, and cannot shed any light on which of the two standard landing configs Gordon was using, or why he implied that 15 degrees was about the maximum angle of attack for it. If he was using Flaps Full, I think alpha-max MIGHT be slightly less than for Flaps 3, but that guess is only based on the fact that my FCOM says the Flaps-Full alpha-floor is 14.5 deg - 0.5 deg less than at Flaps 3. That said, the small, electro-mechanical AoA indicator seems to exceed 15 deg occasionally in the video.

Hi Chris,
I agree. It exceeds 15 momentarilly when he rolls on some bank.
During the steady turn, the AoA gauge then returns to 15 and when he levels the wings again, the AoA shows less than 15. All that is done with full back stick.

In level flight, the system seems to be limiting the AoA shy of Alpha Max - so that some bank can be applied with the same initial pitch attitude without exceeding Alpha Max.

From my FCOM:
• αfloor is activated through the A/THR system, when
α is greater than αfloor (9.5 ° in configuration 0; 15 ° in configuration 1, 2; 14 ° in configuration 3; 13 ° in configuration FULL)

FCOM FCB5 Avoiding Tail strikes; “Assuming an 8 kt speed decrease during flare, and a -1 ° flight path angle at touchdown, the pitch attitude will increase by approximately 4.5 °”

So during the flare, the flight path is changed by +2° (-3° to -1°) for a loss of 8kt. Therefore the angle of attack increases by 2.5° (4.5°-2°) for about 8 kts reduction in speed.
http://www.blackholes.org.uk/PP/AoA.png
From the diagram, αfloor is about 8 kts from αMax, therefore in CONF 3, αMax would seem to be about 14° + 2.5° = 16.5°

A FLOOR doesn't work below 100'RA.

Contemporary FCOM extracts
 

Quote from rudderrudderrat:
"From my FCOM:
• αfloor is activated through the A/THR system, when
α is greater than αfloor (9.5 ° in configuration 0; 15 ° in configuration 1, 2; 14 ° in configuration 3; 13 ° in configuration FULL)"

I guess that must be a FCOM from a later era than June 1988? Here is the equivalent paragraph from the BCAL Tech Manual, dated 17FEB1988, around the time of A320-100 type-certification. [I'm stuck with transcribing it in plain script.]

Alpha-floor is activated when:
- alpha > alpha-floor (9.5 deg in conf 0; 15 deg in conf 1, 2, 3; 14.5 deg in conf FULL, or
- sidestick deflection > 14 deg nose-up and pitch-attitude > 25 deg or in angle-of-attack protection [sic...].

[The ambiguous mix of "and" and "or" in the final, un-punctuated sentence is as printed.]

A slightly earlier, late-1987 edition of the Airbus FCOM Flight Controls chapter (1.09.10 P8, REV 03, SEQ 001) omits any specific alpha figures; instead providing the following figures of the relationship between the alphas at the protection stages and the alpha at Vs.

All configurations at low speed:
alpha-prot >= alpha (1.13 Vs)
alpha-floor = alpha (1.1 Vs)
alpha-max = alpha (1.06 Vs)

Clean configuration at high speed:
alpha-prot = alpha (buffet)
alpha-max = Czmax
When alpha protection is active nose-up auto-pitch-trim is inhibited.

One might assume in the above that Vs is a speed associated with the alpha at CL -max, but it does not state whether that Vs is at 1.0G or a lower figure. And whether it is the same value used in the "Stalling Speeds" graph (FCOM 3.01.20 P5) is unclear.

Therefore, I think any attempt, using a combination of the above sources, to infer specific values of alpha for the three different protection levels might be ill-founded. Also, the two versions may originate either side of the FBW certification process.

Hi Jesse,
"A FLOOR doesn't work below 100'RA."

Yes, we have discussed that previously, although my 1988 manuals only seemed to mention it in the Power Plant section (A/THR automatic-engagement criteria). The Habsheim crew had planned to disable Alpha-Floor to ensure it did not intervene at their briefed display height of 100ft.

For a non-professional it is fun to read all these comments about official manuals for the same type of aircraft that change whenever the manufacturer decides and when those are used by professionals to analyze a sequence of a particular flight they come in of conlusions that contain the words .. ill founded ... ambiguous ... unclear ... probable ... etc ...
So I guess the results when everything is explained in a court where everyone tries to blame the other party :rolleyes:

Quote from jcjeant:
"For a non-professional it is fun to read all these comments about official manuals for the same type of aircraft that change whenever the manufacturer decides and when those are used by professionals to analyze a sequence of a particular flight they come in of conlusions that contain the words .. ill founded ... ambiguous ... unclear ... probable ... etc ..."

Well, that's not entirely true. Haven't noticed any (ex) professional recently using the word "probable" here... ;)

Quote from rudderrudderrat:
FCOM FCB5 Avoiding Tail strikes; “Assuming an 8 kt speed decrease during flare, and a -1 ° flight path angle at touchdown, the pitch attitude will increase by approximately 4.5 °”
"So during the flare, the flight path is changed by +2° (-3° to -1°) for a loss of 8kt. Therefore the angle of attack increases by 2.5° (4.5°-2°) for about 8 kts reduction in speed.
"From the diagram, αfloor is about 8 kts from αMax, therefore in CONF 3, αMax would seem to be about 14° + 2.5° = 16.5°

I take it you are assuming Nz is 1G before and after the flare? I'm wondering about the possible change of ground effect on the 1G-AoAs between (say) 30'R and 5'R.

What a Great video - Never seen it before.
The presence of the AoA indicator is perfect for the demo, only a simultaneous view on the sidestick and the flight controls page could be better.
Alpha max is at 15 deg in landing CONF FULL but 17.5 CONF 3
Absolutely no hesitation to go rapidly to alpha max.
Why no such intention in Habsheim ?

Quote from CONF_iture:
Alpha max is at 15 deg in landing CONF FULL but 17.5 CONF 3

That seems a big difference between the two configurations, particularly as my contemporary FCOM states that the alpha-floor figures are 14.5 deg and 15 deg respectively (see my post above).

I wonder, does your late-1987 (i.e. pre-cert) edition define V2min in relation to Vs (in the A330 FCOM it's in 3.04.10 - Characteristic speeds)?

Hi CONF iture,
Maybe the system can rapidly achieve aMax during +ve delta g, because it is able to avoid exceeding aMax by simply reducing the rate of rotation.
Whereas at Habsheim, there was only the last 2.5° (if 17.5° is correct) of rotation available, and would only approach the limit gradually.

Plus on the Airbus video demo he was at low speed, but keeping the thrust on (i.e. the engines were not spooled down - unlike Habsheim). Also the demo airspeed was stable and not decelerating prior to full back-stick.

Do you have a reference handy for that, or is it a guess?

Being a programmer, I am naturally averse to numbers without knowing what they are and where they're from (aka "magic" numbers).

Airbus contemporary FCOM "Operating Speeds Definition"
 

Dozy,
That's an interesting hypothesis, i.e., the EFCS might deliberately keep a margin to alpha-max in idle thrust. I wonder.
AFAIK, however, we haven't established what config/flap GC was using for that demo, and in any case I currently disagree with Confit's figure of 15 deg for alpha-max with config FULL (see my post).

HN39,
That's a good call, although 3.04.10 was re Flight Instruments, and not included in my issue. However, 3.04.01 "Operating Speeds Definition" may provide the best explanation of how the peculiarities of certification for the A320 in Normal Law - which (as you know) was unique in being incapable of meeting the contemporary norms of stall testing - were addressed, and a compromise reached to avoid it being unfairly disadvantaged, field-performance-wise, relative to types like the B737.

Airbus FCOM 3.04.01, REV01 SEQ001 (still current at SEP 1987, prior to formal type-certification) (my emphasis):
"For a conventional a/c the stall speed which is used for reference is VSmin based on a load factor lower than 1G, which therefore gives a stalling speed less than that obtained at 1G. All operating speeds are expressed in relation with this speed (for example Vref = 1.3 VSmin). Since the A320 icorporates a low speed protection feature (alpha limit) which cannot be overridden by the flight crew, the airworthiness authorities have reconsidered their position regarding the stall speed definition.
"All the operational speeds have to be referenced to a speed which can be demonstrated by flight test. This speed is VS1g for A320 and VSmin for previous a/c types. As a result the Authorities have agreed to allow A320 to have the following factors:
V2 = 1.2 x 0.94 = 1.13 VS1g
VREF = 1.3 x 0.94 = 1.23 VS1g
"It can be seen that these speeds are identical to those that would have been achieved, had the a/c been conventionally certificated to the 94% rule. Compared to a conventional a/c the A320 has exactly the same manoeuvre margin at its reference speeds.
"In the FCOM VS1g will be designed [designated] VS. " [*]

* VS = "Reference stalling speed (equal to VS1g)."

I wouldn't know about that - under the 'contemporary norms' Vsmin would have been equal to Valpha-max.
If "alpha-max = alpha (1.06 Vs)" and Vs=Vs1g, how was Vs1g demonstrated in flight test?

Also, if "alpha-prot >= alpha (1.13 Vs)" and "V2 = 1.2 x 0.94 = 1.13 VS1g", then V2>=V(alpha-prot). So V2 can be equal to V(alpha-prot)?

Hi HazelNuts39,
Well spotted!

FCOM, Aircraft Systems, DSC-27-20-10, Protections:
"At Take off aProt is equal to aMAX for 5s."
By which time, the SRS will have commanded about V2+10.
During OEI, aFloor is disabled.

@Chris Scott
Hi Chris,

Thanks for those figures in post #213 because if we assume the Cl is proportional to AoA (straight line graph between afloor and aMax), then we can guesstimate aMax.

Lift = 1/2 rho * A*Cl*Vsquared.
Since the lift remains constant at 1g, and A is constant,
Cl(amax) * V(amax)squared = Cl(afloor) * V(afloor) squared.
Rearranging Cl(max) = Cl(afloor) * V(afloor)squared / V(amax)squared.
If we assume Cl is proportional to AoA then,

AoA(max) in conf3 = 15 * (1.1)squared/(1.06)squared = 16.15°.
Using the same assumptions, astall = 18.15°
Since the accuracy of the AoA probes is 2.5% (see post #46), then the error could be +/- 0.45°.
18.15 - 0.45 = 17.7

I wonder if that where the 17.5° came from?

With one engine inoperative?
Is that a valid assumption? In the landing configuration AoA for Cl=0 is almost certainly some significant negative value. Also Cl-alpha is non-linear near Clmax.

Hi HazelNuts39,
OEI - SRS commands V2, but aFloor is disabled so there is no problem.
Correct.
That is why I've only considered "straight line graph between afloor and aMax" (see post #211 for the picture).

Inconsistencies in contemporary FCOM
 

Quotes from HN39:
If "alpha-max = alpha (1.06 Vs)" and Vs=Vs1g, how was Vs1g demonstrated in flight test?

It could not have been (in Normal Law), which is inconsistent with the text I quoted. The plot thickens... One possible explanation is that the "Operating Speeds Definition" comes from REV01, whereas the factors you mention are from REV03 in a different volume. All three volumes were issued to us at Blagnac in early January 1988, but we were warned that amendments were already pending.

Also, if "alpha-prot >= alpha (1.13 Vs)" and "V2 = 1.2 x 0.94 = 1.13 VS1g", then V2>=V(alpha-prot). So V2 can be equal to V(alpha-prot)?

Presumably not!

Hi Chris,

Vs1g was demonstrated in flight test in ALT LAW.
It was demonstrated at 1g, whereas in a conventional aircraft, there is a tendency for the nose to drop and the demonstrated value is around 94% of 1g (allegedly).

- apparently it can be.

If the system is fast enough to efficiently control the elevators to rapidly command achieve or control alpha max in one case I don't see how it would not in another case.
Also, an excursion over alpha max is not an issue, as demonstrated in the video.

rrr

OK, but in Conf 3 the zero lift AOA Would be -6.5 deg

In earlier stall demonstrations Vsmin was taken as the minimum speed measured in the stall and stall recovery - this usually corresponded to about 0.94g

Vs1g=reference stall speed is the current norm for conventional aircraft. In ALT LAW the A320 can be stalled as a conventional aircraft, so why would it need a different treatment of stall speed? The explanation given in the FCOM makes no sense if Vs1g was demonstrated in ALT LAW.

Hi HazelNuts39,
A conventional aircraft stall speed is demonstrated by trimming the stab only as far as Vref, then decelerating at 1 kt/sec whilst applying back elevator until the nose drop. It was found that this always produced slightly less than 1g (94% ish) at the measured stall speed value.
In Alt Law, the Airbus continues to trim and level flight was achieved right up until the speed where you will lose control - all apparently done at 1g.

See page 36:http://www.skybrary.aero/bookshelf/books/2263.pdf

"Airworthiness Authorities have agreed that a factor of 0.94 represents the relationship between VS1g for aircraft of the A320 family and VSmin for conventional aircraft types."

Not so much that it keeps a margin for idle thrust (though it may do so) as it takes into account whether the aircraft is accelerating/decelerating and by how much, then uses the trends to stay ahead of the aircraft.

I tried to illustrate this earlier:

Capt. Corps alludes to the idea that the overriding ethos behind the EFCS in this particular mode is retaining stability and keeping the aircraft flying.

[EDIT : In fact, if we go back to Chris Scott's earlier post:

What if the hypothesis that 17.5 degrees should be consistently held with full back-stick in High AoA mode is incorrect?

Is it not possible that 17.5 degrees AoA represents the absolute "do-not-exceed" maximum alpha from a normal limit of around 15 degrees AoA? If Chris's translation is correct, the wording of the BEA report is certainly consistent with that scenario. This interpretation would also be consistent with Capt. Corps' narration on the video.

]

I would go for that one - I did not find anything in the FCOM to state otherwise.
As you mentioned, CONF 3 selection on the MCDU will modify VLS and VAPP, but it is also a signal for the GPWS in its FLAP MODE operation.

Alpha max at 17.5 deg CONF 3 as per BEA
Alpha max at 15 deg in landing configuration as per video
CONF FULL is the usual landing configuration as per FCOM.

To lose 1kt/sec is very soft and normal and certainly nothing to restrain the AoA at 15 deg.

And nothing in the FCTM states that 17.5 degrees AoA is guaranteed in such a scenario, but you still seem to take it as gospel.

The deceleration delta was nothing like stable at -1kt/sec throughout the sequence - try not to think about it as "restraining" AoA as anticipating where it will be if the aircraft continues to slow (which it would do until the engines have spooled back up).

How much was the deceleration in the last 10 sec as the airplane was level ?
When the elevators movement oppose the AoA increase, it is a clear restriction.
But nice theory of your own making, too bad it is not detailed either in the documentation or the BEA report.
What if the Airbus documentation is wrong ?
17.5 deg of AoA for CONF 3 is the target to maximize the performance, that's why Airbus chose to call it Alpha Max.
The video demonstrates how rapidly Alpha Max (15 deg for CONF FULL) is obtained.
But again ... nice theory of your own making.

Hi CONF iture,
The video also shows (at time 10:50 ish) an AoA of 16° after rolling on about 15° of bank. I would guess from that demo that Alpha Max is >15°, but the system remains on the conservative side of Alpha Max to allow for rapid changes in bank.

I can even read 17 deg for a very temporary excursion over alpha max.
Then it would be publicized as such ... Any reference ?

Hi CONF iture,
It simply does this rather well. No promise of achieving aMax - only of not exceeding it.
FCOM 27-20-10 Protections.
"HIGH ANGLE OF ATTACK PROTECTION

Under normal law, when the angle-of-attack becomes greater than αprot, the system switches elevator control from normal mode to a protection mode, in which the angle-of-attack is proportional to sidestick deflection. That is, in the αprot range, from α prot to αMAX, the sidestick commands α directly. However, the angle-of-attack will not exceed αMAX, even if the pilot gently pulls the sidestick all the way back. If the pilot releases the sidestick, the angle-of-attack returns to αprot and stays there."

You and I must be reading this differently: