But again, why do you still persist equalling Valphamax and stalling speed ... ? And if you do think Valphamax is Vs1g at 110 then what is Vstall ? True that all these speeds, all of them, including Vls Valphaprot, should have been published in the report for the weight config altitude at the time ... Where are they ? For the graph, things don't add up. How do they or you get identical speed traces from so different attitudes ? An obvious penalty on the speed there must be by pulling the nose up ... but also a benefit to the altitude ... Where are the ? |
This accident was 26 years ago, yet reams of fascinating info regarding the secrets of the 'bus "logic" continue to fill these pages.
And all this in addition to the acres of apparently "new" knowledge following the report on AF447. I admit I'm an ex-Boeing driver, but the whole process of revelation and presumably education of 'Bus drivers following a major accident, gives me no confidence as l board my next Airbus flight as SLF. If the beast really is that complicated, and so much info remains in the filing cabinet at Toulouse, then surely those who are NOT 'Bus enthusiasts are entitled to ask is this the best way to go in aircraft design and operator instruction? 787 design team please note?!?! :ugh: |
Originally Posted by OK465
For an aircraft without a cockpit AOA indication I'm not sure I agree.
|
Originally Posted by BARKINGMAD
(Post 8427197)
This accident was 26 years ago, yet reams of fascinating info regarding the secrets of the 'bus "logic" continue to fill these pages.
I think one thing that is coming out of the discussions above all else is that the only guaranteed way that accident could have been avoided was to not point an aircraft full of pax at a forest and chop back the power in the first place! That's an airmanship issue, not a technical one. |
Dozy, "It's not really that complicated, to be honest."
Nearly 800 postings on the latest Habsheim thread and almost 1,000 on the last AF447 thread surely scores higher than any Boeing, excepting of course the mystery of MH370? Admitted it's not an exhaustive analysis of the tech aspects of either breed, but maybe there's something in it? It would be interesting to know from 'bus line dogs how much more they have learned about their 'frames if they've followed the discussions in these fora. Yes, l learned things about the 73NG which otherwise would have remained hidden after reading reports various, but doubt that the volume of previously unknowns gets anywhere close to the Airbus genre. Tin hat is on........... |
Originally Posted by BARKINGMAD
(Post 8427397)
Nearly 800 postings on the latest Habsheim thread and almost 1,000 on the last AF447 thread surely scores higher than any Boeing, excepting of course the mystery of MH370?
In fact, Airbus's original EFCS technology as a whole was somewhat less complex than the T7 - the force-feedback logic being almost as complex again. The issue with Airbus discussions on here is that there are definitely a core of pilots - usually, though not always, from France and allied with SNPL - that insist Asseline was wronged by a French government cover-up, and no matter what the actual discussion is about, some of them will try to turn the conversation back to that. I'm hoping this thread will answer any remaining issues and end the ridiculous excuses in future... [EDIT : Put it this way - the level of software implementation we're talking about here would be roughly the equivalent of knowing about the tensile properties of metal cables and hydraulic flow rates on conventionally controlled airliners of yesteryear - which pilots didn't really need to understand in order to fly the aircraft!] |
@ OK465
The simulations appear to be well conducted and reasonably high fidelity with respect to that specific aircraft and incident....but as previously noted, the plots not so much. In addition to the pitch attitude inaccuracies, I particularly like the Pg 13 IAS plot that is moved upward halfway thru the simulation and connected with a vertical dashed line. I think there's possibly a little more pride taken in these type of presentations nowadays. The four "simulator" charts bear all the hallmarks of having been produced on an analogue pen recorder - the sort where paper is drawn past a line of pens which move only transversely across the paper. To analyse the data subsequently one had to annotate and add scales by hand so these are inevitably "untidy" compared to a printed version. To add to the misery the gridlines on the plotter paper were, IIRC, printed in a delicate shade of pink, which was great if you were reading directly off the plotter paper but which all but disappeared when passed through the sort of copying machine then available. Both these "faults" are present in the published charts, but they could have been then state of the art :8 About 10 more seconds of the sim plot on Pg 13 might have been interesting and actually I don't see where it would have cost anything to continue it to purported alphamax and present it. The sim was already set up and rolling. I would guess that the extension you advocate was looked at - in fact the BEA report (p.15) talks about "other studies", but these were treated as background material. If one is prepared to speculate a little it is possible to get some idea of what might have happened - check your PMs. |
Originally Posted by Owain Glyndwr
(Post 8427838)
@ OK465
I take your point, but I do wonder if we aren't being a little harsh when we judge a 1988 presentation by the standards of reproduction equipment available today. ... Both these "faults" are present in the published charts, but they could have been then state of the art :8 Even if it were possible, the current version of Windows was 3.0, and the current version of Word was 1.0 (for Windows). In June 1990, these would have been effectively brand new products, and while it was technically possible to insert Excel data into WFW 1.0, Excel could not generate charts - that didn't really come into its own until about 1993-4. Obviously the realisation that these tools could be used in this way did come, and modern reports use that method routinely these days. However, this was a result of a near-revolution in desktop computing power throughout the 1990s, and simply wasn't available for the publication of this report. As BARKINGMAD says, we're talking about an accident that happened almost 26 years ago, and a final report that is almost 24 years old! |
@ Barking
It is interesting that we are still here talking about the FCS implementation and such after such a long time. The point is that many heavy pilots have never flown their planes to within a hair's breadth of the programmed limits. The accident was caused by a low power setting and lack of practice and plain poor judgement ( demonstrating the jet at its max AoA with pax on board at a ridiculous altitude). Sheesh. I go with Doze and Okie for the most part. And I think Okie flew the same jet I did in his previous life, which was full FBW and very clear limits defined as to the aerodynamics. The mentality that the cosmic flight control system will "protect" you is dangerous. We don't all have to be super engineers or test pilots, but we do have to understand the "limits" and then all the reversion modes of both the FCS and the autothrottle ( Asiana). So I feel that good training in the real jet as well as the sim could help here. I know that $$$ count, but seems to me that a few hours and maneuvers in the real jet would pay off big time. AF447 PROVED that you could stall the jet if you climbed at such an angle and ran outta speed before the magic FCS could help. The plane was still recoverable, so it was not truly in a "deep stall", it was "deeply stalled". That's all I gotta say here from the peanut gallery. |
Originally Posted by gums
(Post 8429164)
AF447 PROVED that you could stall the jet if you climbed at such an angle and ran outta speed before the magic FCS could help.
I think the reason this particular accident still causes discussion and debate today hinges - at least partially - on the fact that the report was only ever published in French (which was standard practice when no international body was involved), and most of what the Anglophone world knows about it was filtered through the press. Asseline's argument was that the aircraft did not respond to his full back-stick command. The BEA research indicates that it did respond, but due to the late application of thrust and even later application of the back-stick command, the rate of response was initially slower than it might have been. This was not a routine operation - far from it - and if it had been, then maybe the details which this thread has uncovered would have been more widely known. As it was, the investigation concluded (rightly so, in my opinion) that the systemic lack of oversight was the primary factor in the accident, with the somewhat poor airmanship during the conduct of the flight also contributing. A radical overhaul of the practices regarding commercial aircraft performing demonstrations at airshows was thus sufficient to prevent a recurrence. [EDIT : I must say that while I'm appreciative of the namecheck, IMO the dynamite work on this thread has come from Owain Glyndwyr, HazelNuts39 and Chris Scott as well as OK465! :ok:] |
gums
447 and Habsheim are two different things. Protection was never an issue in AF447 as the aircraft was in alternate law and you could always stall. It was lack of procedural knowledge on part of pilots was the cause. In normal law you have the protection as was the case in Habsheim but here the issue that was being raised was did the protection cause the accident. Both side have gone considerable distance with interesting inputs although only agreeing to disagree. |
@ vilas
I don't disagree with you about AF447. My concern is that it appears some 'bus drivers seem to feel that they can't stall the jet, and then the lack of certain "protections" in the various reversion modes comes into play when things go south. That is why I suggested better training both in the sim and in the real airplane. I know involves lottsa $$$, but sheesh. In the early days of my experience as an IP in a full FBW jet, we first let the student get used to the side stick, and then we would go to the limits. We did not have a simulator for 3 years after the jet became operational! And you could not demo the gee limiter in a stupid sim. Habsheim was a great example of poor planning, poor preparation, and poor judgement. Sorry to come down hard on a fellow pilot, but that's my feeling. Even our demo hops for the crowds allowed for some leeway, and the maneuvers were practiced over and over. A degree or two of AoA that the FCS allowed at max command was never a consideration. Back to my peanut gallery...... |
Originally Posted by vilas
(Post 8430213)
... but here the issue that was being raised was did the protection cause the accident.
The question is not about whether the protections "caused the accident" as much as it is whether the slower initial response due to HAP mode made the contact with the trees more substantial than it might have been. At this point I think agreeing to disagree is the best we can manage, because as I said before, there are simply too many variables involved to make a definitive educated guess. @gums - If there are FBW Airbus pilots who believe they can't stall one under any circumstances (and to be honest I'm sceptical of that), then they clearly haven't read the manuals and training materials properly and require being pulled off the line for immediate remedial training! I know there's a school of thought that has the AF447 PF pulling up because he was expecting the protections to keep the aircraft in check, but there's no clear evidence of that. The fact is that there have been studies performed into startle response, which found that severe cases of startle will cause a pilot to instinctively pull up in (IIRC) 80% of cases regardless of whether the aircraft has AoA protection or not. |
DOZY
I am in agreement with you and I also do not believe there pilots out there who do not know that Airbus can be stalled in alternate law. Pilots are getting complacent on automation and loosing their basic scan. Far too many accidents happen because of not monitoring speed on approach. Take SFO for instance no pilot is going to keep quiet when he sees the speed is 15/20 KTS below Vapp. They simply are not looking at the speed because most of the time automation looks after it. AF447 pilot applied memory items of after take off situation otherwise there is no requirement of TOGA power. Unreliable airspeed is only practiced during type rating and never in yearly or six monthly checks. |
Originally Posted by BARKINGMAD
(Post 8427397)
Tin hat is on...........
Originally Posted by Vilas
AF447 pilot applied memory items of after take off situation
|
Barkingmad yep being ex Boeing myself now Airbus for quite a while. I was surprised after AF447 to find out the 330 stall characteristics as all the stalls done in the sim were conventional (buffet then pitch down) no one mentioned it could "deep stall" the bus does require continuous study. They are fascinating however.
|
Originally Posted by HN39
Considering that the cockpit display includes Valphaprot and Valphamax, which speed would you like to add?
This has been discussed extensively earlier in the thread. If there is anything that doesn't add up, please be more specific. How do they or you get identical speed traces for Bechet and Asseline as they adopt so different attitudes ?
Originally Posted by Owain Glyndwr
To analyse the data subsequently one had to annotate and add scales by hand so these are inevitably "untidy" compared to a printed version.
To add to the misery the gridlines on the plotter paper were, IIRC, printed in a delicate shade of pink, which was great if you were reading directly off the plotter paper but which all but disappeared when passed through the sort of copying machine then available. |
Originally Posted by CONF iture
How do they or you get identical speed traces for Bechet and Asseline as they adopt so different attitudes ?
Secondly, one of many explanations is that Asseline flew into a forest and Bechet did not. Asseline's IAS would have been higher that Bechet's if he had not lost 5 kt of headwind in the 5 seconds before he hit the trees: http://i.imgur.com/JeHgmNn.jpg?1 |
Originally Posted by CONF iture
If Vs1g was Valphamax Airbus would not have called Vs1g "stalling speed".
Furthermore, in Bechet's simulator test, at about 6.8 seconds after moving the thrust levers forward, the AoA is 17.5 degrees and the airspeed is 116.5 kIAS. According to the FCOM, Vs1g is 109.5 kt at a weight of 59,000 kg, so 116.5 kt is 1.064 times Vs1g. The loadfactor at an AoA is proportional to airspeed-squared, and should be 1.064*1.064 = 1.13. That value is a perfect match with the loadfactor shown in the graph below, that is derived from the altitude trace (or the radius of the flight path) of the simulated flight. In other words, Bechet's simulation confirms that Vs1g corresponds to alpha-max of 17.5 degrees at 1g. http://i.imgur.com/JiAfjEg.jpg?1 |
Originally Posted by CONF iture
(Post 8432194)
Sorry but there is ample way to do it professionally ... or you can always choose to do it unprofessionally.
Tell me ... is it that same sort of copying machine that swallowed Annexe VII ? The way to check this would be to go and have a look at the original paper copy in the Republic archives, but I wouldn't know how to go about doing that (and as a foreign citizen may not be able to access it anyway). |
Habsheim Report / French Official Journal
Dear Dozy & al.
The Republic's archives will come to you ! The reports published by the BEA on their website merely serve an informative purpose. The "genuine" reports are those published in the paper version of the french "Journal Officiel", a publication that regroups all official french communications, such as newly promulgated laws, minor and major administrative changes or reports of importance. The "Journal Officiel"'s publication relating to Habsheim is available here: Journal Officiel 24 Avril 1990 As anyone can see, annex VI devoted to the CVR transcription is there in its entirety, starting page 29 of the report. A technical error has placed the pages 27 & 28 after page 29 instead of before. But the CVR transcript continues on page 30 and following. I hope that this puts to bed the now mythical "dissapearance of annex VI" and that everyone realises that the inversion (not the omission) of two pages in a document cannot constitute the grounds for accusations of "tampering". |
@Agnostique75:
That's kind of you to provide that link, but that's the same one we've been using since the start of the thread. CONF iture is referring to a different Annexe - number 7 (VII), which is listed on p.23 of the PDF as "Corrélation DFDR-CVR", but appears (on p.36 of the PDF) to actually be a different portion of the DFDR trace. |
Oops ! I shall copy 50 times "Fully engage brain before writing" :\
|
Originally Posted by CONF iture
The cockpit displays only include what is believed to be Valphaprot and Valphamax according to the FAC ... but such characteristic speeds have theoretical values, and knowing that Asseline was playing in that area, it should have been the most natural thing for the BEA to communicate those values in their report.
In the energy plot shown in post #772 it can be seen that Bechet lost 89 ft of total energy after he started to move the thrust levers until 5 seconds later the engines had spooled up sufficiently and the total energy started to increase. Bechet could absorb most of those 89 ft out of his kinetic energy, but if the speed drops below Valphamax it is not possible to arrest an increasing rate of descent. |
What are you guys exactly trying to determine? Apart from reenacting the event...;)
Simply what happens when you pull the stick fully aft and keep it there? Could it be that airbus decided to write that valpha-max may be reached by pulling full aft sidestick, because of the habsheim accident? To me the sidestick input certainly does not look like he wanted to reach valpha max throughout. According to the figure in page 7 he only reached max aft side stick in the last few seconds. With loadfactors changing, engines spooling up, wind changing... the flight control computers might simply not have had enough time to reach valpha max (before crashing). Therefore some lawyer decided to write may be reached. |
Originally Posted by 737Jock
(Post 8435964)
With loadfactors changing, engines spooling up, wind changing... the flight control computers might simply not have had enough time to reach valpha max (before crashing).
Therefore some lawyer decided to write may be reached. I love how there's always this assumption of skullduggery in the case of this event, when the general drift of this thread indicates that there probably wasn't any. For the record, going back to my post #26 - "may be reached" comes from the FCTM for the A330, not the A320. The A320 doc is worded thus: If the PF then pulls the sidestick full aft, a maximum AOA (approximately corresponding to CL Max) is commanded. I think you are right that the important factor was time, but it had more to do with the belated realisation of the danger they were in (as a result of an approach which was in itself belatedly improvised and, for want of a better word, botched) and the late application of TOGA thrust than it did any issue with the computers. |
Valphamax is a minimum airspeed requirement, and the aircraft never dipped below it prior to impact according to HN39's analysis. The maximum aoa was never reached, but the sidestick was only fully aft for about 2 seconds. During those 2 seconds a lot happened and the FCS might not have had enough time to actually provide the commanded maximum aoa and associated valphamax. Certainly considering, thrust spooling up, loadfactor changes and wind change. I would imagine valphamax being reached more easily in circumstances that are more "stable". Manufacturers certainly are concerned by litigation regarding their manuals. The only way to determine would be to find an old FCOM/FCTM. But since we don't have that we will never know. Anyway im not even saying, nor did I imply that the system has a fault that needed to be clarified, it might just be a rewording to emphasize its not a guarentee that valphamax will be reached. But I am sure that it is commanded. Success depends on many factors, some internal and many external. I'm sure those pilots just pulled the stick as they saw the forrest come closer. All the detailed info is nice, but those sidestick inputs do not suggest they wanted to fly at max aoa, evident from not having sidestick full aft. Just somewhere close to it. They had no idea what their exact speed, aoa or vs was. |
Originally Posted by 737Jock
(Post 8436194)
I would imagine valphamax being reached more easily in circumstances that are more "stable".
What seems to have made the response slower than it might have been was the significant deceleration with the stick half-back and thrust still at idle - then staying half-back until a few seconds after TOGA was commanded, because in that scenario HAP mode restricts pitch rate change to maintain stability and trajectory. The test flight being discussed by HN39 (as part of the original investigation) involves a slightly different scenario where the stick was moved full-aft simultaneously with TOGA being applied, thus the aircraft was in pitch command mode, which allowed pitch (and AoA) to increase more rapidly - initially. However what effect this might have had - if any - regarding impact with the trees can only be a matter of speculation, as the initial faster pitch change is arrested until the thrust starts bringing the airspeed back up again. |
But the aircraft was flying *faster* than Valphamax throughout - the EFCS didn't have to "reach" anything, and there was no impediment to the aircraft reaching Alpha Max other than time. They never flew at Valphamax, they never commanded Valphamax until they were about to kiss the trees, evidenced by the lack of full backstick until the last 2 seconds before crash. What seems to have made the response slower than it might have been was the significant deceleration with the stick half-back and thrust still at idle - then staying half-back until a few seconds after TOGA was commanded, because in that scenario HAP mode restricts pitch rate change to maintain stability and trajectory. They only applied a full aft input on the sidestick in the last 2 seconds. Just in those 2 seconds, loadfactors, engine thrust and wind were changing rapidly. This would lead me to simply think the flight control computers did not have enough time. If you want to reach Valphamax, the simple matter of fact is that you have to decelerate significantly. If you do this with idle thrust and then smack on toga thrust, the change of force on the airframe will be much more significant then if you had for example 40% N1 on. Lets not forget the pitch-up effect from underslung engines that needs to be compensated for, which happened to really kick in when the sidestick reached full aft position (habsheim). Combine that with a constantly changing altitude, and you see significant thrust changes and loadfactors. In more stable conditions (bechet) the FCS will have less of a problem reaching Valphamax (on applying full backstick) as the required output is not constantly changing. P.is he clearly had the benefit of time. I dont think it has anything to do with some sort of sluggishness of the flightcontrols that only happens after a prolonged time in HAP, it is just the system managing the changing forces as best as it can. Although its a computer it is still a reactive system. A human certainly could not fly at Valphamax under those circumstances with any accuracy. |
Why would you want to "reach" Valphamax? The intent was to perform a high pitch flypast at relatively low speed, not take it to the absolute limit as far as I was aware.
|
Is that even a serious question? Why do you think he pulled full aft stick in the last seconds?
Maybe he realised he ****** up his original plan and needed max performance to escape from terrain contact? What would you do in an airbus if you are about to hit the ground involuntarily? Maybe you have a read through the QRH and see what they want you to do in case of imminent terrain contact. I haven't met someone who tries so hard to misunderstand another person. |
Originally Posted by BEA Final Report
Synopsis: Résumé de l'accident :
Dans le cadre d'une manifestation aérienne, l'avion effectue un passage au-dessus de la piste 34 R à une hauteur voisine de 30 pieds, moteurs réduits, avec une incidence croissante jusqu'au maximum possible compte tenu du taux de décélération de l'avion. Annexe VI: 12h30'20'' - CDB: (...) on fait un passage à cent pieds, train sorti et là, tu me laisses faire. Je t'amène en alpha max, je débraye l'alpha floor et à ce moment là, si je te dis que c'est dur, tu m'aides et tu tiens les gaz à vario zéro. (...) |
Interesting stuff watching you playing with numbers and graphs ... now how do you justify Bechet, according to your graph, gains 2.5 deg of alpha in 2 sec, reaches 1.1g but still manages to lose altitude ... ?
Load factor increase should reveal a change of path, from level flight to climb ... where is that altitude gain ? In other words, which data are erroneous and/or 'adjusted' ?
|
Originally Posted by CONF iture
In other words, which data are erroneous and/or 'adjusted' ?
But you are constantly changing the subject. The issue was Valphamax, where you disagreed with me, and you have not answered my question: How would you define Valphamax? P.S. RE:: "Load factor increase should reveal a change of path, from level flight to climb ... where is that altitude gain ?" The story begins with a descending path. At t=2.5 seconds the descending path has changed to level flight, then begins to climb. At t=4.75 seconds the altitude has increased 5 ft and is again equal to that at t=0. Between t=0 and t=4.75 the vertical speed changes from approximately -2.5 ft/s to +4 ft/s, i.e. a change of 6.5 ft/s in 4.75 seconds is 1.37 ft/s/s. Divide that by the acceleration of gravity (32.17 ft/s/s), add 1 and you get an average loadfactor of 1.043 in that period. However, allowance must be made for the resolution of the graph in the BEA report. The whole graph is 1550 x 965 pixels. The scale of the radio altitude trace is about 1 foot per pixel, and the pen trace is about 3 pixels wide. |
@HN39
I must say I admire your patience and tolerance - and I wholly agree with your analysis! |
Love all this pulling the stick back etc. on an AB the side stick commands G. So if you hold say 1/2 side stick back you command say 1 3/4 g. The flight controls will then move to give you that G up to their max deflection if required. So permanent back stick will get you to Alpha Max. The only way to stop the continued demand for G is to centre the stick. In theory the aircraft then holds the pitch you were at then. In practice it usually "nods" down AA degree or so.
So if he was holding the stick aft of ctre he would be pitching up until alpha max. (note lateral stick is a roll rate demand) The above is a simplified description I know. :ugh: |
Originally Posted by IcePack
(Post 8441987)
Love all this pulling the stick back etc. on an AB the side stick commands G.
As far as the human pilot is concerned, in almost every scenario one can think of the distinction is so transparent as to be immaterial - i.e. at its simplest, the aircraft will go where you point it, and the tech will do its best to keep the shiny side up. The only reason the modes are being scrutinised so closely in this case is because questions have been asked as to why the aircraft behaved as it did at a level deeper than the original report was required to go. Certainly as far as I can tell with the benefit of extra analysis from those who are experienced in the aero end of things, the EFCS was complying with the control demands to the best of its ability, but with the aircraft in such a low-energy state, neither technology nor manual handling finesse would have been sufficient to defy the laws of physics. |
I say again, " 'bus linedogs", how much more are you learning about your vunderplane from these pages than what you've managed to extract from the official OM/FCOMs?
Somewhere in the murky depths of my memory I recall the great surprise in the 'bus world, following the VA A340 tailscrape accident at Gatwick, that the 'frame would behave differently above 50' radio height than below it. If memory serves me right, that quirk was unknown to the early 'bus drivers until after that accident, but I wait to be corrected? |
Originally Posted by HN39
The scale of the radio altitude trace is about 1 foot per pixel, and the pen trace is about 3 pixels wide.
The story begins with a descending path. Is it new to the BEA as they did not mention such discrepancy ... ? 1.16.1.4.1. Dans le deuxième cas, où le pilote du simulateur était le président de la commission d'enquête, le recoupement de l'évolution de la vitesse en fonction de temps est également excellent, l'altitude du vol simulé étant toutefois toujours très légèrement supérieure à celle du vol réel (+ 8 à 10 pieds). It must be the same clerk that messed up again ... At t=2.5 seconds the descending path has changed to level flight, then begins to climb. At t=4.75 seconds the altitude has increased 5 ft and is again equal to that at t=0. But you are constantly changing the subject. The issue was Valphamax, where you disagreed with me, and you have not answered my question: How would you define Valphamax? "It represents the speed corresponding to the maximum angle of attack that the aircraft can attain in pitch normal law." |
Originally Posted by CONF iture
Not differently that the FCOM :
"It represents the speed corresponding to the maximum angle of attack that the aircraft can attain in pitch normal law." P.S. The special conditions for certification of the A320 (and associated guidance material) define Vs1g as the minimum speed (corrected to 1g and zero thrust) obtained in the following maneuver: - forward center of gravity - idle thrust - normal law except deactivated alpha-floor - decelerate at 1 kt/sec until sidestick reaches aft stop - maintain sidestick on aft stop for 2 seconds or until pitch attitude does not increase further, whichever occurs later - recover |
All times are GMT. The time now is 22:51. |
Copyright © 2024 MH Sub I, LLC dba Internet Brands. All rights reserved. Use of this site indicates your consent to the Terms of Use.