November 4, 2003 - by Victor Trombettas
NTSB's current thinking
In his opening statement at the Hearings last year, Mr. Robert Benzon, the NTSB's Investigator In Charge (IIC) of Flight 587 (FL587), declared that the tail of FL587 did not separate until "the aerodynamic and internal loads that the vertical stabilizer experienced were significantly above the ultimate loads required by the French and American certification standards". In other words, the tail worked just like it was expected to. What caused such excessive loads on the tail? Mr. Benzon addressed this by stating -- "the rudder movement(s) resulted in most, if not all, of the loads imposed on the vertical stabilizer". The final question then would be -- who or what moved the rudder so aggressively? The answer is that "Investigators have found no indications of any rudder system anomalies" -- which translated means -- the pilot moved the rudder.
U.S.Read's coverage
U.S.Read and the aviation experts contributing to our coverage have doubts and dismay about the NTSB's current thinking and their handling of the investigation. We will be enumerating our myriad concerns in articles to come. As with our first post-hearings update, in this update we will continue examining the NTSB's position that the tail departing the aircraft was the cause of the crash, consider its implications, and we will, for argument's sake, go with the assumption that the tail separated when the NTSB says it did.
NTSB's supporting evidence
For the NTSB, it all boils down to the "Loads Calculations" -- how much bending stress was placed on the tail at the moment they assume it separated from the craft
note to self--I see need here: what evidence supports that tail sep'd when thhey say it did?
Talk about Feb 8th rudder is 9.5 degrees. I think in Feb 8th update NTSB says loads approached? Not exceeded? That's with FEM 1.0 Whereas at Hearings they (loads) were WAY beyond! April 12 NTSB update said Airbus upgraded FEM. It is under this new upgrade perhaps that we have the new loads calculations! Now -- this raises the question about Flight 903 -- when did Airbus know the load factors and what effect did using FEM 1.0 have on those calcs then.
1. it's important to understand this critical point. The NTSB's rather confident statement (benzon's remarks)
I said "the very evidence the NTSB has
presented to establish the load calculations and rudder movements
mentioned above are unsupported by the Flight Data Recorder."
End with: The evidence that the tail fell off in flight, over Jamaica Bay, is indisputable. The evidence that the tail separation was the initiating event, and that it separated when the NTSB says it did, is weak and open to misinterpretation.
We will be exploring these myriad issues in future write-ups.
FROM MY FIRST DRAFT
Wake and Weather and Time Shifts
The NTSB finally released weather data at the Hearings they had
held onto for a year; data that showed higher wind speeds at altitude
than previously known, increasing the probability of wake encounters.
In addition, there were significant changes to the timeline.
The NTSB had also waited an entire year before updating their
timeline. Some of the changes to the timeline were rather
startling and the NTSB has yet to respond to U.S.Read queries
requesting an explanation for how such dramatic changes could
occur. The NTSB's Mr. Lopatkiewicz did respond to U.S.Read
and stated that the previous timeline (the one the NTSB let stand
for the year leading up to the Hearings) was "crude and preliminary".
U.S.Read's response was that the Flight 587 Update pages (on the
NTSB's web site) which discussed the timing of key events in the
flight, never included the words "crude" or "preliminary".
Before we explain some of the timeline changes,
it's important to review some basic facts about the flight recorders.
FL587 was equipped with a Cockpit Voice Recorder (CVR), which
recorded all Pilot communications and cockpit noises onto an analog
tape. This tape had no time code, and can be susceptible
to recording speed changes, which if present, would alter the
apparent timing of events. FL587 was also equipped with
a solid state (or digital) Flight Data Recorder (DFDR).
The DFDR does has a time code and since it is a digital recording
device it is the more reliable of the two recorders when it comes
to tracking the timing of events. It is not susceptible
to the same potential for error as the CVR.
What is unusual about the timeline changes
the NTSB released at the Hearings, is that though the timing of
CVR events were mostly unchanged in relation to each other, there
were significant changes to the timing of events within the DFDR.
This is difficult to understand since the DFDR is a digital recorder
with time codes. For example, in 2001, the NTSB stated the
two wake encounters were 20 seconds apart. At the Hearings,
we learned they were now only 15 seconds apart. The signature
of these events we had been told (and Mr. Benzon re-confirmed
at the Hearings) were two lateral accelerations, both registering
about one tenth the force of gravity (1/10th g force) in a side
to side movement. Those events were associated with a DFDR
time code in November, 2001, called a Subframe Reference Number
(SRN). The NTSB read that time code and were able to say
the two wake encounters were 20 seconds apart. How did those
two time-coded events, within a digital recording device, become
15 seconds apart one year later? No answer from the NTSB.
The NTSB had also let stand for almost a year,
that the 2nd wake encounter occurred 8 seconds before the DFDR
data ended. At the Hearings, we discovered this number had
changed from 8 to 10 seconds. Same question, applies here
as well -- "how do time coded, digitally recorded events,
move around?" To highlight that U.S.Read was not alone
in these interpretations of the November 2001 timeline, we refer
to the Wake Turbulence Study prepared by NASA for the NTSB which
discussed the probabilities of the two wake encounters.
This report, which became a part of the NTSB's Aircraft Performance
Report, was dated October 9, 2002, about 3 weeks before the Hearings,
and 11 months after the crash. On the 4th page of their
report, NASA stated that the second turbulence event occurred
8 seconds before the DFDR data ended. Apparently, the NTSB
did not inform NASA of the timeline change, or they did, and NASA
failed to assimilate the data. This fact alone should not
have a negative impact on the efficacy of NASA's study.
However, it does raise questions about the efficacy of the timeline.
And if the timeline raises questions, then it raises questions
not only of the probability of wake encounters, but more important
issues to the NTSB -- such as calculations analyzing loads on
the tail. NASA also stated that the second turbulence event
occurred 18 seconds before impact. That is also incorrect
-- the new timeline shows the second turbulence event occurred
approximately 23 seconds before impact. This error by NASA
was also based on the previous timeline which had stated that
the DFDR data ended 10 seconds before the CVR stopped recording.
The new timeline showed yet another significant change -- the
DFDR now ended 13.6 seconds before the CVR did. NASA apparently
wasn't told this either. In certain legal venues, these
types of errors by NASA and the NTSB would have a negative impact
on the trustworthiness of their submissions. In New York
City, one incorrectly written number on a parking ticket is grounds
for dismissal. U.S.Read simply strives to undertsand the
basic question -- "how do time-coded, digitally recorded
events shift in time?" The NTSB can complain about
filtered DFDR's (where certain readings of flight control surfaces
like the rudder's position are averaged and sampled only twice
a second -- and this is a valid concern) but we can't understand
what they do with the rest of the DFDR -- the parts that aren't
filtered.
Killer Turbulence?
CONTINUE HERE
It is difficult to understand the NTSB's approach to releasing
factual data. Especially when compared with the speed and
tremendous availability of information from the Shuttle Columbia
investigation(s).
The NTSB also showed data which indicates FL587
might have encountered wake turbulence from the 747 in front of
it. More importantly, the NTSB released a timeline a year
after the crash which was significantly different from the one
they released 11 months earlier. In all their updates
in the year since the crash, they never addressed these changes,
why they were made, nor how such changes could be made.
However, all the Investigators agreed that this turbulence was
barely perceptible in the simulator, and these encounters did
not produce "upsets"-- did not place the aircraft in
an unsafe attitude (in fact, there was no change whatsoever in
the attitude of FL587
at all during either of these alleged wake encounters --
only a slight sensation of a bump.
In summary, the NTSB position currently is
that the Pilot reacted to very mild turbulence by inputting severe
control wheel and rudder inputs, the airplane oscillated sharply
from side to side in response to these inputs, and the stress
placed on the tail during the last of five rudder movements was
so severe that the tail broke off. Once the tail broke off,
the flight was doomed, the plane started an uncontrolled descent
towards Belle Harbor, with the stresses of this violent descent
ripping both engines (along with the pylon structure as Airbus
has designed it) off the wings. The NTSB issued recommendations
to the FAA and the airlines in February 2002 that (a) Pilots should
be trained that side to side rudder movements (rudder reversals),
even at previously thought of "safe speeds", can place
catastrophic loads on the tail, and that (b) as the speed of the
airplane increases the amount of distance the rudder pedals need
to be pressed to achieve maximum deflection of the rudder, decreases.
At the speed FL587 was traveling at the time of the rudder reversals
-- only a 5/8ths inch movement of the rudder pedal would produce
a maximum deflection of the rudder (approximately 10 degrees)
at that speed.
The battle between American Airlines and Airbus then amounted to trying to blame each other for the rudder movements induced by the Pilot. American proclaiming that Airbus never warned them that rudder reversals could exceed the design limits of the tail, and Airbus claiming that American encouraged use of the rudder in their training programs. Airbus tried desperately to find some indications that the Pilot had a history of reversing rudder. The best "history" they could come up with was the uncorroborated account of one Pilot in 1997 who stated he witnessed Sten Molin reversing rudder. It should be noted that not only was this never corroborated by other American Airlines Captains, the Flight Engineers who were present in the cockpit in 1997 during these episodes, do not remember Sten Molin upsetting the aircraft. It should also be noted that dozens of Captains who supervised First Officer Molin described him as an excellent Pilot. If the Pilot had a history of reversing rudder, and if rudder reversals can lead to catastrophes, then why had this Pilot -- with an alleged "history" of rudder reversals -- never before created an upset which led to a near crash experience? The logical conclusion is that there was something very different indeed with FL587, something far beyond the habits of any Pilot, far beyond the issue of rudder reversals.
Flight 903
USAToday reported on May 26th what seemed to be new information ... that in 1997 another American Airlines Airbus A300 flying near Palm Beach (FL903) on its way to Miami experienced a severe upset as a result of stalling. In an effort, to recover from the frightening and violent upset, the crew also "reversed rudder". These reversals created loads that exceeded the design limits of the tail on more than one of these rudder movements. The crew was able to regain control of the airplane. Only after the crash of FL587 in 2001 did Airbus expose the tail of the FL903 plane to ultrasound scans ... which showed what "intense" visual scans (i.e. with the naked eye) could not show ... underlying damage in the right rear composite lug (attachment point to the fuselage--one of six such "lugs"). The tail was removed from that airplane to be replaced with a new one. This information really wasn't "new". It was revealed at the Hearings last year. What this information did show was that Airbus (in 1997) did not reveal to the NTSB ... that rudder reversals created loads on the tail of FL903 which exceeded design limits ... and Airbus did not take steps to use FL903 as a case study in training the airlines about the dangers of rudder reversals. Without a doubt ... since the NTSB believes the pilot-induced rudder reversals on FL587 were most probably the cause of the crash ... and had the industry been made aware of the dangers of rudder reversals in 1997, Sten Molin would never have considered such movements as part of his repertoire, and Flight 587 would not have crashed. Without a doubt ... this was gross negligence on the part of Airbus. Perhaps this is one very large reason why Airbus has recently agreed to a 50-50 sharing of liability with American Airlines. They cannot escape the shame of this now widespread revelation (thanks to the USAToday article). What is interesting is how no one at the NTSB in 1997 seemed to worry at all about the loads that rudder reversals might place on a tail. The NTSB's lack of concern in this area reflected the widespread understanding at the time among not just Pilots, but the world's premiere accident investigative body, that using the rudder (even reversing rudder) was not a suicidal move in an airplane. Clearly, it wasn't just American Airlines who felt they could be aggressive with the controls -- the NTSB didn't have a problem with it.
WHy no mention of limiter exceedance by Benzon? WHy was he able to say no anomalies? NTSB investigators and others familiar with NTSB "Hearings" have told U.S.Read that is because Mr. Benzon knew what Airbus was going to say. The questions the Investigators ask at Hearings are pre-determined and shared with the "witnesses" -- those who will testify. And certainly, Mr. Benzon knew exactly what Airbus was going to say to the very weak and limited questions his staff would put to Airbus.
discuss info given by insider - the big guy from the south