TAM A320 crash at Congonhas, Brazil
Join Date: Jul 2007
Location: UK
Age: 59
Posts: 43
Likes: 0
Received 0 Likes
on
0 Posts
To Dani
Dani,
Ref your response 1997 to my post 1982.
Of course they would if the investigation establishes this as a causal factor.
Well, I wouldn't bother applying for any jobs in accident investigation or safety/certification just yet if I were you. Here are a couple of rules:
- One can rule out possible failure modes on the basis of evidence only. Otherwise the best you can do is assign relative probabilities (based on past failures/fault trees/circumstantial evidence/engineering judgement etc..) to the different possibilites and maybe end up with a probable cause.
- If a failure mode has not yet occurred, this has no effect on the probability of its happening (a suprisingly common misconception).
My personal feeling from the evidence made public so far is that pilot error/HF is more likely than a linkage failure (based on previous incidents (not limited to Airbus)). The official inquiry must address the possible failures which I identified in post 1982. Hopefully adequate evidence has survived to give an unequivocal answer.
Human factors also apply to maintenance personnel. Split pins have a low (but non-zero) failure rate if they are correctly fitted. Their efficacy is much diminished if they are omitted.
I couldn't agree more.
Ref your response 1997 to my post 1982.
If there had been a technical/mechanical problem then Airbus/TAM/CIAA would have at least done some checks on other Airbusses, issued technical recommendations and bulletins...
I think we can easily rule out any other reason than the obvious one. ... I think we can safely argue that the pilots left the TL in CLB detent...
- One can rule out possible failure modes on the basis of evidence only. Otherwise the best you can do is assign relative probabilities (based on past failures/fault trees/circumstantial evidence/engineering judgement etc..) to the different possibilites and maybe end up with a probable cause.
- If a failure mode has not yet occurred, this has no effect on the probability of its happening (a suprisingly common misconception).
My personal feeling from the evidence made public so far is that pilot error/HF is more likely than a linkage failure (based on previous incidents (not limited to Airbus)). The official inquiry must address the possible failures which I identified in post 1982. Hopefully adequate evidence has survived to give an unequivocal answer.
Human factors also apply to maintenance personnel. Split pins have a low (but non-zero) failure rate if they are correctly fitted. Their efficacy is much diminished if they are omitted.
Back to the facts please...
Join Date: Jul 2007
Location: UK
Age: 59
Posts: 43
Likes: 0
Received 0 Likes
on
0 Posts
To FlyingNewbie10:
I don't know if an Audio Spectrum Analysis of a CVR ever had so much importance as it apparently has now.
Are there any alternative CVR transcripts available following the latest political interference?
Guest
Posts: n/a
Are there any alternative CVR transcripts available following the latest political interference?
One thing I would like to figure out is the reason of that right TL deviation at idle stop (-3 degrees instead of zero) at chart 9 of the FDR, in a context of mechanical/electronical problem.
Sun worshipper
Join Date: Dec 2001
Location: Paris
Posts: 494
Likes: 0
Received 0 Likes
on
0 Posts
Tahiti-Faa overrun
Join Date: Jan 2005
Location: France
Posts: 2,315
Likes: 0
Received 0 Likes
on
0 Posts
Lemurian,
Thanks! I've saved the report.
"Enjoy" may not be quite the right word....
But I'll see if I can extract something relevant to the current subject (I'm bilingual English-French, so no language problem).
Thanks! I've saved the report.
"Enjoy" may not be quite the right word....
But I'll see if I can extract something relevant to the current subject (I'm bilingual English-French, so no language problem).
Sun worshipper
Join Date: Dec 2001
Location: Paris
Posts: 494
Likes: 0
Received 0 Likes
on
0 Posts
ChristiaanJ,
I took a quick look at the Camair CDG incident you gave us a link to, and I came across the comment made by the BEA saying that this overrun had some similarities with another two incidents, namely F-GITA at Faa and China Air B-165 at Kai Tak.
That one is also worth a careful read as, from the accident analysis,
China Air B-165 HKG
What I find quite interesting is that this report destroys all the blurb we've seen on this thread about the need for moving throttles (because of *tactile feed back*, *visibility* (knowing that 744 throttles are a damn sight bigger / taller than 320 T/Ls).
Finally, going back to the argument someone had with bsieker on occurrences'statistics, I'd say that, considering the number of sectors flown by a long -hauler, the statistics do not seem to be in favour of the M solution, do they ?
Of course I have a very simple mind
Best regards
That one is also worth a careful read as, from the accident analysis,
- they landed normally in difficult rain / windshear conditions
- T/Ls retarded to *idle*, speedbrakes extended and autobrakes kicked in.
- Somehow #1 T/L was pushed 8 degrees forward of idle, the other T/Ls in a lesser extent
- this action on #1 T/L disconnected the autobrake and caused the speedbrakes to retract (with the corresponding lever movement on the pedestal)
The consequences of advancing the levers instead of selecting
reverse thrust were serious. Residual forward thrust
increased, wheel braking stopped and the speedbrakes were
retracted. Neither pilot noticed these very important
changes. They did not look at the speedbrake lever and they
may not have felt the effect of wheel brakes because brake
pressure was being phased in. - When eventually, reverse thrust was reapplied, the speedbrakes extended again
- The airplane overrun the runway at a speed of some 30 knots.
No casualties.
China Air B-165 HKG
What I find quite interesting is that this report destroys all the blurb we've seen on this thread about the need for moving throttles (because of *tactile feed back*, *visibility* (knowing that 744 throttles are a damn sight bigger / taller than 320 T/Ls).
Finally, going back to the argument someone had with bsieker on occurrences'statistics, I'd say that, considering the number of sectors flown by a long -hauler, the statistics do not seem to be in favour of the M solution, do they ?
Of course I have a very simple mind
Best regards
Last edited by Lemurian; 5th Sep 2007 at 23:46. Reason: Comment added
The Reverend
Join Date: Oct 1999
Location: Sydney,NSW,Australia
Posts: 2,020
Likes: 0
Received 0 Likes
on
0 Posts
Not trying to blow my own trumpet as a flight engineer but had there been a flight engineer present, neither AF or Air China would have finished up in the drink. SOP on the classic 747s was for the flight engineer to hold the engineer's spur thrust levers against the stops after selection of reverse thrust which prevented any forward movement of the thrust levers.
Sorry, I can't seem to be able to enlarge this image.
Sorry, I can't seem to be able to enlarge this image.
Last edited by HotDog; 6th Sep 2007 at 02:52.
Join Date: Sep 2000
Location: Bielefeld, Germany
Posts: 955
Likes: 0
Received 0 Likes
on
0 Posts
Originally Posted by flyingnewbie10
I don't know if an Audio Spectrum Analysis of a CVR ever had so much importance as it apparently has now.
PBL
Join Date: Sep 2000
Location: Bielefeld, Germany
Posts: 955
Likes: 0
Received 0 Likes
on
0 Posts
I have been thinking about some different ways to gain insight into this moving/non-moving thrust lever debate, as well as the somewhat lesser "leave the TOGA option" debate.
My intuition is that with sufficiently-trained operators, nothing decides the issue definitively. However, I am looking for general reasons why this may be so.
There is at least one such phenomenon for the "leave the TOGA option" debate. It is associated with the name of John Buridan, who was rector of the Sorbonne in 1328 and 1340, but apparently he never said it (see Anthony Kenny, Medieval Philosophy, A New History of Western Philosophy, Vol. 2, Oxford U.P. 2005, p96). If you take an ass and place it between two equally attractive bales of hay, it will be unable to decide between them and (so the theory goes) starve to death. The argument can obviously be generalised to two unequal bales of hay: there is a point in between them at which the ass will inevitably be undecided.
There is a theoretical formulation of this for (mathematical) metric spaces, due to Lamport and Palais, from the 1970's. There is, for every binary (generalise: discrete) decision function on a sufficiently continuous decision space a point of undefinition: where there is no defined decision. Lamport wrote this up in a more digestible version for non-mathematicians in a paper on the Buridan Problem in the 1980's. Both may be found on his WWW site.
The original application was to digital arbiters. Say, flip-flops, or even bits. Electrical current comes in, in a suitably continuous fashion, and you want the thing to go to 0 or 1. Can't always happen, say Lamport and Palais. Yes, but the chances of it not happening are vanishingly small, say "practical" people. Turns out not to be so. Turns out that chip producers have to do some work to figure out the "metastable behavior" (the new word for this indecision region) of certain HW to try to minimise the indecision behavior. Indeed, the phenomenon was discovered independently at about the same time by that eminently practical man Charles Molnar, who built the first desktop computer. Neither Lamport nor Molnar could get their papers published ("can't happen" said the referees. So much for the worth of peer review).
So, after all these words, Point 1: the "Buridan" indecision phenomenon is established. Now for its application.
For every discrete decision problem one constructs from continuous input, there will be an indecision region. It doesn't matter whether we are talking hardware, software or wetware. For every wonderful array of hinting devices, interlocks, moving/non-moving thrust levers, and so on, there is going to be a region in which one's decision criteria for continue the landing/go around don't yield a clear-cut decision.
If you change the technology, you just change the indecision region, you don't eliminate the phenomenon. If you push the decision somewhere else (say, 10 seconds before TD), you don't eliminate the phenomenon. If you wiggle the thrust levers, or leave them still, you don't eliminate.. etc.
Can I put this in one phrase? The "Buridan" indecision phenomenon persists through all technologies and designs.
One name? The Metastability Principle. One Mnemonic? Buridan's ass or yours!
This has a number of corollaries. One is to pull the teeth of the argument "yes, but if they had only had blue thrust levers instead of pink ones, this accident wouldn't have happened". Maybe true, but if you make the change you have just put the problem somewhere else, where some other flight will discover it which wouldn't have if you had left them pink.
There are more obviously larger areas of indecision. See the accident to a Tower Air B747 on RTO from JFK, December 20 1995,
https://www.flightsafety.org/ap/ap_mar97.pdf
http://www.ntsb.gov/Publictn/A_Acc2.htm
in which forward thrust was also applied during the RTO.
I am tempted, but too lazy this morning, to try to apply the Metastability Principle to the moving/non-moving thrust lever argument. I had some inkling from memory that there were examples of mismatched thrust levers on landing even on moving-lever airplanes, but I was simply too lazy to look them up or get on the phone to those who know (thank you, Lemurian!)
PBL
My intuition is that with sufficiently-trained operators, nothing decides the issue definitively. However, I am looking for general reasons why this may be so.
There is at least one such phenomenon for the "leave the TOGA option" debate. It is associated with the name of John Buridan, who was rector of the Sorbonne in 1328 and 1340, but apparently he never said it (see Anthony Kenny, Medieval Philosophy, A New History of Western Philosophy, Vol. 2, Oxford U.P. 2005, p96). If you take an ass and place it between two equally attractive bales of hay, it will be unable to decide between them and (so the theory goes) starve to death. The argument can obviously be generalised to two unequal bales of hay: there is a point in between them at which the ass will inevitably be undecided.
There is a theoretical formulation of this for (mathematical) metric spaces, due to Lamport and Palais, from the 1970's. There is, for every binary (generalise: discrete) decision function on a sufficiently continuous decision space a point of undefinition: where there is no defined decision. Lamport wrote this up in a more digestible version for non-mathematicians in a paper on the Buridan Problem in the 1980's. Both may be found on his WWW site.
The original application was to digital arbiters. Say, flip-flops, or even bits. Electrical current comes in, in a suitably continuous fashion, and you want the thing to go to 0 or 1. Can't always happen, say Lamport and Palais. Yes, but the chances of it not happening are vanishingly small, say "practical" people. Turns out not to be so. Turns out that chip producers have to do some work to figure out the "metastable behavior" (the new word for this indecision region) of certain HW to try to minimise the indecision behavior. Indeed, the phenomenon was discovered independently at about the same time by that eminently practical man Charles Molnar, who built the first desktop computer. Neither Lamport nor Molnar could get their papers published ("can't happen" said the referees. So much for the worth of peer review).
So, after all these words, Point 1: the "Buridan" indecision phenomenon is established. Now for its application.
For every discrete decision problem one constructs from continuous input, there will be an indecision region. It doesn't matter whether we are talking hardware, software or wetware. For every wonderful array of hinting devices, interlocks, moving/non-moving thrust levers, and so on, there is going to be a region in which one's decision criteria for continue the landing/go around don't yield a clear-cut decision.
If you change the technology, you just change the indecision region, you don't eliminate the phenomenon. If you push the decision somewhere else (say, 10 seconds before TD), you don't eliminate the phenomenon. If you wiggle the thrust levers, or leave them still, you don't eliminate.. etc.
Can I put this in one phrase? The "Buridan" indecision phenomenon persists through all technologies and designs.
One name? The Metastability Principle. One Mnemonic? Buridan's ass or yours!
This has a number of corollaries. One is to pull the teeth of the argument "yes, but if they had only had blue thrust levers instead of pink ones, this accident wouldn't have happened". Maybe true, but if you make the change you have just put the problem somewhere else, where some other flight will discover it which wouldn't have if you had left them pink.
There are more obviously larger areas of indecision. See the accident to a Tower Air B747 on RTO from JFK, December 20 1995,
https://www.flightsafety.org/ap/ap_mar97.pdf
http://www.ntsb.gov/Publictn/A_Acc2.htm
in which forward thrust was also applied during the RTO.
I am tempted, but too lazy this morning, to try to apply the Metastability Principle to the moving/non-moving thrust lever argument. I had some inkling from memory that there were examples of mismatched thrust levers on landing even on moving-lever airplanes, but I was simply too lazy to look them up or get on the phone to those who know (thank you, Lemurian!)
PBL
Join Date: Jul 2002
Location: Confusio Helvetica
Posts: 311
Likes: 0
Received 0 Likes
on
0 Posts
Buridan was rector of the University of Paris, not the College de la Sorbonne (which was part of the University). In fact, Buridan taught at the College de Navarre.
Buridan's Ass applies specifically to beings without a rational soul. Humans are supposed to be able to reflect on the equal demand presented by the sensitive appetite to two equally attractive options, and determine that any decision is better than indecision.
Buridan's Ass applies specifically to beings without a rational soul. Humans are supposed to be able to reflect on the equal demand presented by the sensitive appetite to two equally attractive options, and determine that any decision is better than indecision.
Buridan's Ass applies specifically to beings without a rational soul. Humans are supposed to be able to reflect on the equal demand presented by the sensitive appetite to two equally attractive options, and determine that any decision is better than indecision.
Join Date: Sep 2000
Location: Bielefeld, Germany
Posts: 955
Likes: 0
Received 0 Likes
on
0 Posts
Originally Posted by DingerX
Buridan was rector of the University of Paris, not the College de la Sorbonne (which was part of the University). In fact, Buridan taught at the College de Navarre.
PBL
Last edited by PBL; 6th Sep 2007 at 11:29. Reason: I am convinced by DingerX
Join Date: Sep 2000
Location: Bielefeld, Germany
Posts: 955
Likes: 0
Received 0 Likes
on
0 Posts
Originally Posted by DingerX
Buridan's Ass applies specifically to beings without a rational soul. Humans are supposed to be able to reflect on the equal demand presented by the sensitive appetite to two equally attractive options, and determine that any decision is better than indecision.
Originally Posted by atakacs
Never in my wildest dream would I have thought to read such a statement in this forum
PBL
Last edited by PBL; 6th Sep 2007 at 10:01. Reason: Fonts
Join Date: Aug 2007
Location: Brazil
Age: 71
Posts: 131
Likes: 0
Received 0 Likes
on
0 Posts
one day at the airport
The other day I was at the airport and bumped into some friends, airline pilots. We went for a coffee and an interesting "discussion" came up. At the "table" we could see two Boeing pilots, on their mid 50's, three Airbus pilots (one on his late 40's, the other two on their early 30's) and me (a mere helicopter pilot and a light twin airplane sunday driver). No academics there (sorry PBL).
The moving/non moving TLs issue came up, but didn't last long. The old timers said they don't mind if they move or not, as long as they do what they are supposed to - control thrust. One Airbus pilot (the older one) said he missed flying Boeings and the other two A pilots said they enjoy flying Airbuses very much, and they said non moving TLs makes more sense for them. One even said he would prefer big switches instead of TLs. Makes sense, he is from the computer games generation...
Of course, about this TAM accident, the TL was the most discussed thing. It was not a surprise to me when all three Arbus pilots mentioned problems with TLs, not only here in Brazil but also around the world. These problems were one of TLs not working at all. Two cases during taxi, one TL went dead (really dead, not "normal" dead...), one case reducing from TO power to CLB. I asked for more detail, they said they will e-mail me. When I get those, will post.
The Boeing pilots said that weard things happen with Bs too.
One interesting thing happened on that "round table" discussion, none of the pilots there were into discussing systems, A x B, etc... They all talked about maintenance, and Murphy's law was cited. All participants of that "coffee break" didn't rule out HF as probable cause, but since they know equipment fail all the time also (A or B), they are eager to see what investigators come up with. But not to put a blame, but to learn from it.
And the meeting ended with that old joke about future of aviation, when copilots will be replaced by a German shepherd dog...
Regards,
Rob
The moving/non moving TLs issue came up, but didn't last long. The old timers said they don't mind if they move or not, as long as they do what they are supposed to - control thrust. One Airbus pilot (the older one) said he missed flying Boeings and the other two A pilots said they enjoy flying Airbuses very much, and they said non moving TLs makes more sense for them. One even said he would prefer big switches instead of TLs. Makes sense, he is from the computer games generation...
Of course, about this TAM accident, the TL was the most discussed thing. It was not a surprise to me when all three Arbus pilots mentioned problems with TLs, not only here in Brazil but also around the world. These problems were one of TLs not working at all. Two cases during taxi, one TL went dead (really dead, not "normal" dead...), one case reducing from TO power to CLB. I asked for more detail, they said they will e-mail me. When I get those, will post.
The Boeing pilots said that weard things happen with Bs too.
One interesting thing happened on that "round table" discussion, none of the pilots there were into discussing systems, A x B, etc... They all talked about maintenance, and Murphy's law was cited. All participants of that "coffee break" didn't rule out HF as probable cause, but since they know equipment fail all the time also (A or B), they are eager to see what investigators come up with. But not to put a blame, but to learn from it.
And the meeting ended with that old joke about future of aviation, when copilots will be replaced by a German shepherd dog...
Regards,
Rob
Last edited by Rob21; 6th Sep 2007 at 17:17. Reason: typo
Join Date: Aug 2005
Location: fl
Posts: 2,525
Likes: 0
Received 0 Likes
on
0 Posts
Rob21, After the German Shepherd bites you for touching something he doesn't think you should, he doesn't help with the situation you are in so I think we will have copilots for a very long time, hopefully.
Join Date: Oct 2005
Location: USA
Posts: 21
Likes: 0
Received 0 Likes
on
0 Posts
Southwest went off the end of a snowy runway at Chicago Midway and killed a car passenger because deployment of reversers was significantly delayed over the assumed deployment time in their landing-distance calculation.
Sun worshipper
Join Date: Dec 2001
Location: Paris
Posts: 494
Likes: 0
Received 0 Likes
on
0 Posts
Could an F/E have saved the day ?
HotDog
Maybe, but one cannot generalise as in the Camair incident -a 742 -, the flight engineer participated in the confusion in the cockpit, with the consequences that we now know :
(my translation)
In this instance, we see the often observed synchronised drops in performance of all the members of a crew in heavy stress situations.
Not trying to blow my own trumpet as a flight engineer but had there been a flight engineer present, neither AF or Air China would have finished up in the drink. SOP on the classic 747s was for the flight engineer to hold the engineer's spur thrust levers against the stops after selection of reverse thrust which prevented any forward movement of the thrust levers.
...the reason for the thrust increase of #1 engine remains difficult to explain...
That the F/E, by leaning over the pedestal in order to reach the speedbrake handle might have inadvertently displaced #1 throttle towards full forward thrust cannot be discarded...
A lack of coordination between the captain and the F/E, as well as an absence of cross monitoring with the F/O might have contributed in the irreversibility of the situation...
That the F/E, by leaning over the pedestal in order to reach the speedbrake handle might have inadvertently displaced #1 throttle towards full forward thrust cannot be discarded...
A lack of coordination between the captain and the F/E, as well as an absence of cross monitoring with the F/O might have contributed in the irreversibility of the situation...
In this instance, we see the often observed synchronised drops in performance of all the members of a crew in heavy stress situations.
I support PPRuNe
Join Date: Jul 2007
Location: Belo Horizonte, Brazil
Posts: 162
Likes: 0
Received 0 Likes
on
0 Posts
From a Brazilian Pilot Site
Interesting site from a talented Brazilain Jet Pilot. Mostly in Portuguese Brazilian, but some segments in English.
Interesting point he makes is that the right engine suddenly accelerated in a rapid pace before the front and the left wheels had touched the ground. This sudden acceleration reached its peak in 3-5 seconds raising the EPR of the right engine to 1,260.
He also has some considerations for the mechanics of the TLs and what could have gone wrong although not probrable.
Here:
http://aviationtroubleshooting.*************/
Interesting point he makes is that the right engine suddenly accelerated in a rapid pace before the front and the left wheels had touched the ground. This sudden acceleration reached its peak in 3-5 seconds raising the EPR of the right engine to 1,260.
He also has some considerations for the mechanics of the TLs and what could have gone wrong although not probrable.
Here:
http://aviationtroubleshooting.*************/
Join Date: Jun 2001
Location: Surrounded by aluminum, and the great outdoors
Posts: 3,780
Likes: 0
Received 0 Likes
on
0 Posts
As a neophyte 320 "pilot" followed the MEL procedure watched the "momentery" non ops rev engine accel as it takes a bit for it to "realize" the rev did not actually deploy then it returned to "idle" as per the "ops procedure"....all speculation aside...my experience was as per MEL "o" procedure....maybe theirs was not the case.....**** happens