WT Forums

Pilot Error Blamed for Flight 587 Crash

By LESLIE MILLER, Associated Press Writer

WASHINGTON - The co-pilot of American Airlines Flight 587 caused the November 2001 crash that claimed the lives of 265 people, the staff of the nation's airline safety agency reported Tuesday.

Investigator Robert Benzon of the National Transportation Safety Board staff said the copilot's response to turbulence, just seconds after the Airbus A300-600 plane took off from New York's John F. Kennedy International Airport, was "unnecessary and aggressive."

Benzon also said that investigators found that American Airlines improperly trained its pilots to use the aircraft's rudder while recovering from upsets and said the problem could have been exacerbated by the airline's simulator training.

Benzon also said that the rudder control system on the aircraft is sensitive at higher air speeds, which is potentially hazardous.

The safety board itself was expected to rule later Tuesday on the staff's findings.

On Nov. 12, 2001, First Officer Sten Molin, the co-pilot, moved the plane's rudder back and forth after takeoff, trying to control the climbing aircraft, not realizing he was sealing the grim fate of those on board.

Molin was at the controls when the plane hit turbulence almost immediately after taking off for the Dominican Republic.

"Hang onto it, hang onto it," Capt. Edward States implored.

"Let's go for power, please," Molin said.

A second later came a loud bang, which investigators believe was the tail breaking off. Then came the roar of air rushing against the aircraft and alarms sounding in the cockpit.

"What the hell are we into (inaudible)?" Molin said. "We're stuck in it."

States' last recorded words came five seconds later: "Get out of it! Get out of it!"

Both Airbus Industrie, which manufactured the jetliner, and American Airlines, which trained Molin, agree that if he had taken his foot off the rudder pedal, the tail wouldn't have broken off, the plane wouldn't have plunged into a New York City neighborhood. It was the second deadliest plane crash on U.S. soil.

But Molin didn't know he was putting more pressure on the tail than it could bear. Why he didn't — and who's to blame for that — is the subject of a bitter fight between Airbus and American.

According to investigators, Molin tried to steady the aircraft using pedals that control the rudder, a large flap on a plane's tail. When his initial movement failed, Molin tried again and again. His actions placed enormous stress on the tail.

American, the only U.S. airline to use that type of Airbus plane for passenger service, claims Airbus didn't alert it to the danger of sharp rudder movements until after the crash. The airline also contends the Airbus A300-600 has uniquely sensitive flight controls that can cause more severe rudder movements than the pilot intends.

"Airbus had the ability to truly red-flag the issue," American spokesman Bruce Hicks said.

Airbus says it told American a number of times and in a number of ways that the airline was improperly training pilots about how to use the rudder.

An Airbus spokesman declined to comment on the investigation before the hearing. However, the company has provided the NTSB with a number of documents to support its claim.

For example, a letter dated Aug. 20, 1997, warned American chief pilot Cecil Ewing that rudders should not be moved abruptly to right a jetliner or when a plane is flown at a sharp angle. The letter was signed by representatives from The Boeing Co., the Federal Aviation Administration (news - web sites) and Airbus.

Airbus contends that even people within American Airlines were concerned about how the airline was training its pilots. A letter to Airbus dated May 22, 1997, from American technical pilot David Tribout expressed concern about the airline's then-new training course on advanced maneuvers.

"I am very concerned that one aspect of the course is inaccurate and potentially hazardous," Tribout wrote. His concern: Pilots were being taught that the rudder should be used to control a plane's rolling motion. Hicks countered that Airbus didn't share important safety information about the rudder after a problem with American Flight 903 in May 1997. During that incident, pilots used the rudder to steady an Airbus A300-600 plane on approach to West Palm Beach airport. The plane nearly crashed and one person was seriously injured.

Afterward, Airbus told the NTSB that it included a warning that abrupt rudder movement in some circumstances "can lead to rapid loss of controlled flight," and, in others, could break off the tail.

Hicks said Airbus' comments didn't specifically say the rudder movements on Flight 903 had exposed the tail to so much pressure that it could have been ripped off.

Immediately after the Flight 903 incident, an inspection found no damage to the tail. But five years later, the plane was inspected more closely because of concerns aroused by the crash of Flight 587. Cracks were found and the tail was replaced.

John David, a spokesman for American Airlines' pilots union, said pilots had always thought that they could use rudders to the full extent without hurting the airplane. He also believes Airbus didn't properly communicate what it knew.

American now gives its pilots specialized training on the rudder control system based on information learned during the investigation.

"The power to Destroy the planet, is insignifigant to the power of the Air Force----Mudd Vader

ummm shouldent the PILOT be taking off and landing? And shouldent they be issued a board with a nail in it for occasions like this.

"We sleep safely in our beds because rough men stand ready in the night to visit violence on those who would harm us". George Orwell

Fighting For Justice With Brains Of Steel !
<img src="http://www.fas.org/man/dod-101/sys/ac/atengun2X.GIF" border=0>

_________________
The only time you have too much fuel is when you're on fire.

ummm shouldnt the Verticle stay attached to the aircraft, and a Yaw dampaner been in place like all Modern aircraft?

Pilot only did what was natural and had been trainied to him all the way up to His ATP certification. The pilot was fighting adverse yaw conditions and the Whole Tail section came off.

I will let Will chime in here on Crew Takeoff and Landing procedures.

"The power to Destroy the planet, is insignifigant to the power of the Air Force----Mudd Vader

of course the stab should have stayed in place but it's an Airbus. My point was it seemed like the pilot and co weren't co-operating. And I still think the BMOC should do the takeoffs and landings.

"We sleep safely in our beds because rough men stand ready in the night to visit violence on those who would harm us". George Orwell

Fighting For Justice With Brains Of Steel !
<img src="http://www.fas.org/man/dod-101/sys/ac/atengun2X.GIF" border=0>

_________________
The only time you have too much fuel is when you're on fire.

Funny this came out in the news today. I wrote a safety report on it for Embry-Riddle about 1.5 years ago. Here is some of my report:

<b>Captain John F. Lavelle, an American Airlines pilot who has flown with F/O Molin, described him as a perfectionist with excellent skills. Lavelle observed Molin in 1997, apply “excessive rudder input” on several 727 flights they operated in 1997. When Molin was questioned about this technique, Molin said his action was "per the American Airlines Advanced Aircraft Maneuvering Program (AAMP).</b>

Keep in mind, this was an American Airlines course. Not, an Airbus course. Airbus knew this procedure was being taught:

<b>There was some debate about rudder use at one of the public hearings held by the National Transportation Safety Board (NTSB). Airbus, the aircraft manufacturer, expressed concern for the overemphasis placed on rudder use under the AAMP instruction. </b>

An American Airlines pilot wrote to Airbus:

<b>A300 technical pilot for American Airlines, Captain David Tibout, expressed concern for the course. He wrote a letter to Airbus in May of 1997. He wrote that the course is "potentially hazardous. It states that at higher angles of attack, the rudder becomes the primary roll control. The instructor teaches that in the event of wake turbulence encounter, recovery from the stall teaches that the rudder should be used to control roll."</b>

Basically what it boils down to is that American Airlines was going against manufacturers recomendations.

To add to the cauldron...the aircraft had been previously damaged in the vertical stabilzer area. This is what my research found:

<b>This aircraft was involved in a severe turbulence incident in 1994 on a flight to Puerto Rico. The turbulence was severe enough to cause injuries to 47 passengers. (www.abcnews.com, 19 Nov. 2001) Due to this incident, the vertical stabilizer on Flight 587 had been repaired with metal rivets and additional materials. These types of repairs on composites are frequently unreliable, especially for joints and attachments involving primary or, load bearing structures. </b>

With American Airlines training and the repair that was done; this aircraft was doomed. It just needed the right pilot in the seat.

I also understand that is routine to have the Co-pilot at the controls during take-off and landings. This is for training.

Go Ugly Early !
No body's ugly after 2:00 A.M.

_________________
Go Ugly Early!
Nobody is ugly after 2 A.M.

mudd not being a pilot please explain why they said the pilot should have not used the rudder and should have had his feet on the floor for most of the flight.

I thought the rudder was a primary flight control and the pilots feet were always on the rudder. When ever you watch discovery channel the are teaching pilots in small planes to keep feet on rudder and to use the rudder constantly as part of the control system.

Is it just that the airbus has a sophisticated computer control system.

When something breaks in the air, I am inclined to criticize the design criteria or look at the specific interpretation of the criteria. I think this incident will cause a revision to the Federal Aviation Regs regarding the calculation of design loads for vertical fin/rudder. We can all agree that through some combination of aircraft characteristics and pilot input, the fin saw loads it was not designed to resist. Is there is something unique in the yaw control system that allows pilots to apply full rudder against the sideslip while at maximum sideslip angle? This is not a normal design condition. If I am not mistaken, there are three types of design rudder manuevers. They are:

The rudder kick (while at zero sideslip the rudder is deflected suddenly)

maximun sideslip/maximum rudder (rudder defelected to produce max sideslip, including some dynamic overshoot of steady state maximum)

maximum sideslip rudder reverse (rudder is suddenly deflected to decrease sideslip while at maximum sideslip).

The case that is most like this incident is the rudder reverse. In olden days with only muscle powered controls, the rudder reverse case could not get as bad as today. By the time the rudder had reversed direction, the fin force had reduced the sideslip substantially, so the total load was naturally limited. Modern, heavy jets apparently respond slower to the yaw moments produced by fin load and the control systems have much more capability, in terms of rate of control surface deflection and the peak force achieved before actuator "stall". Artificial stability requirements tend to demand high power and rate from actuators so maybe that is related. Pilots have more power and rate at their disposal and this invites them to over control and get into pilot induced occilations that end up overshooting the allowable sideslip/rudder combinations.


I think there must be conflicting information out there, Hogmender, because I remember a story in Aviation Week that said that the repair doublers on the fin of the Flt 587 aircraft were done at the factory before the plane was delivered. So, they were unrelated to the Flt 903 overload incident. Also, this repair would have needed confirmation of its capability from experts, so I am confused over the statement "These types of repairs on composites are frequently unreliable, especially for joints and attachments involving primary or, load bearing structures.". If that is true, the entire European system is fouled-up and we should not be riding on Airbuses.

It is repeated in many sources that the AA pilots practice that the rudder is the primary yaw control at high angles of attack. I don't know, but this would be consistent with using the swept wing dihedral effect to roll the airplane by inducing sideslip with the rudder. Is this common? I like the idea of all the axes having their own control surfaces. Strong cross coupling, like adverse yaw from aileron deflection, seems to get you in trouble. This is a job for the flight dynamics/controls people. They could come up with some control laws to prevent overload. Of course, you need a fly-by-wire system to implement it properly. The F-22 has a feature in the flight control software that recognizes extreme control movements, like stabilators going from stop to stop, and takes over for a time to smooth over the situation and avoid pilot induced occilations. Forgive me, it may be an oversimplification to describe it that way, but it seems to work.


THE RAMPTOR ENGINEERING TEAM <img src=icon_smile_big.gif border=0 align=middle>
"Who cares if it works? Does it look good on the ramp?"

_________________
????

I thought it rather humorous that the pilot was blamed for using the rudder. Generally, when I'm flying, I tend to hope that parts don't just fall off the airplane when I try to use them...

ATTACK!!!!!!!

<BLOCKQUOTE id=quote><font size=1 face="Verdana, Arial, Helvetica" id=quote>quote:<hr height=1 noshade id=quote>
mudd not being a pilot please explain why they said the pilot should have not used the rudder and should have had his feet on the floor for most of the flight.
<hr height=1 noshade id=quote></BLOCKQUOTE id=quote></font id=quote><font face="Verdana, Arial, Helvetica" size=2 id=quote>

This is very much a matter of personal preference, technique, aircraft type, the situation, and habit patterns.

I have flown light Cessnas and Pipers, transport category airplanes, A-10s, and lots of stuff in between. I ALWAYS fly with my feet on the rudder pedals, more from habit back when I flew taildraggers than anything else.

In a transport category jet, often you are not sending control inputs directly to the control surface, but through a series of systems that use hydraulic pressure to deflect the surface. Jets fly at higher airspeeds, which mean higher control pressures, and not enough strength in a pilot's legs to overcome those forces without some help.

In many jets, a yaw damper or dampers automatically cancel out any yaw, thereby allowing the pilot(s) to put their feet flat on the floor and not worry about yaw - the airplane takes care of that for them.

Conversely, in a light propeller-driven aircraft, there are a lot of yaw forces affecting the airplane - moreso than on a jet. The prop induces varying yaw moments depending on engine power, number of engines, engine location, aircraft pitch attitude, etc.

The two extremes would be flying a light taildragger like an Aeronca Champ or Piper Cub versus flying a Boeing 767. If you don't use rudder in the taildraggers you're not going to get off the ground; if you do use it once you are airborne in the 767 (with all systems nominal and under normal flight conditions) it is very likely you'll actually be working against the aircraft (and the passengers won't like it).

<BLOCKQUOTE id=quote><font size=1 face="Verdana, Arial, Helvetica" id=quote>quote:<hr height=1 noshade id=quote>
I thought the rudder was a primary flight control and the pilots feet were always on the rudder. When ever you watch discovery channel the are teaching pilots in small planes to keep feet on rudder and to use the rudder constantly as part of the control system.
<hr height=1 noshade id=quote></BLOCKQUOTE id=quote></font id=quote><font face="Verdana, Arial, Helvetica" size=2 id=quote>

All of the above is true. Just like people have different personalities, airplanes act differently and so rudder may or may not be important to flying it...personally, as I said, I may let the autopilot have the stick for a while but my feet are always on the pedals, no matter what I'm flying.

<BLOCKQUOTE id=quote><font size=1 face="Verdana, Arial, Helvetica" id=quote>quote:<hr height=1 noshade id=quote>
Is it just that the airbus has a sophisticated computer control system.
<hr height=1 noshade id=quote></BLOCKQUOTE id=quote></font id=quote><font face="Verdana, Arial, Helvetica" size=2 id=quote>

If Kerry and Bush were airplanes, Kerry would be a Airbus and Bush would be a Boeing.

I don't think even the Airbus software logic knows what the hell it wants to do half the time. There's some great footage of an early A319 or 320 going into a forest after the pilot did a low fly-by for spectators and commanded a go-around - only the airplane didn't want to go-around, it wanted to land instead.

Anytime a computer can veto what I want to do with an airplane (with certain exceptions, such as G-limiting on fighters), I say get the friggin' computer out of the loop.

Hope this answers some questions!

ATTACK!

<BLOCKQUOTE id=quote><font size=1 face="Verdana, Arial, Helvetica" id=quote>quote:<hr height=1 noshade id=quote>
I thought it rather humorous that the pilot was blamed for using the rudder. Generally, when I'm flying, I tend to hope that parts don't just fall off the airplane when I try to use them...
<hr height=1 noshade id=quote></BLOCKQUOTE id=quote></font id=quote><font face="Verdana, Arial, Helvetica" size=2 id=quote>

Couldn't have said it better myself.

Airbus = <spit> french <spit> piece of shit

If it ain't Boeing, I'm not going!

ATTACK!

<BLOCKQUOTE id=quote><font size=1 face="Verdana, Arial, Helvetica" id=quote>quote:<hr height=1 noshade id=quote>If it ain't Boeing, I'm not going!<hr height=1 noshade id=quote></BLOCKQUOTE id=quote></font id=quote><font face="Verdana, Arial, Helvetica" size=2 id=quote><img src=newicons/anim_lol.gif border=0 align=middle>

Love it! 30MM jr will have a new mantra!

"Live every day like it's the last, 'cause one day you're gonna be right!" Ray Charles (6/10/04 was the day)

_________________
\"Those who hammer their guns into plows
will plow for those who do not.\"
- Thomas Jefferson

Yep, the guy was just doing what he was trained to do. The only way he would have any idea that the rudder is stressed for one input and not one then back deflected the opposite way would be if he had background in flight test or aircraft certification. Of course the info is out there and available to the public. Who would have thought that certification would only account for stresses induced under a one time yaw doublet, not a reverse input. Boeing and the airlines (most companies)came out right after this and let the pilots know in a memo the certification and flight test profile for transport catagory aircraft. We still dont train for it in the sim. The FAA and the airlines pretty much bank on the rare occurence that it might never happen again. Remember, safety costs. It is expensive to train pilots on anything that is not verbatim required by the FAA. Anything out of the ordinary and not canned is RARELY trained due to limited simulator availability and once again money. Also when you train a guy for something out of the ordinary there is the chance he will fuck it up. If this happens then you have to RE-TRAIN him and put it on his record of a failure. The airline training is so expected and canned that you know and are expecting EVERY Emergency or NON-Normal. I really could go on all night but I'm just back from a Honolulu turn and my employer just filed for bankruptcy. I'm done!
BTW to the dude that posted above about the AC or the Captain making every takeoff and landing. You couldn't be more wrong. Thats another topic altogether.
Goodnight!
Will Whiteside

a10stress, chances are, there is conflicting info out there. Keep in mind I wrote the paper about 4 months after the crash. Try a search on abcnews.com for 19 Nov 2001. That article I used might still be there.

Another cool reference that I used is here: http://www.aviationnow.com/media/jpg/587dfdr2.jpg

It shows exactly what he did to the rudders. Study it closely. At 9:15:59, the rudder sensor goes off of the chart. I'm pretty sure that's when it separated. He used them quite violently. If Airbus had some concerns over rudder use, I think this was bound to happen.

Go Ugly Early !
No body's ugly after 2:00 A.M.

Edited by - hogmender on Oct 27 2004 10:20 PM

_________________
Go Ugly Early!
Nobody is ugly after 2 A.M.

I agree with the contention that the pilot overcontrolled this A300 and ended up breaking the fin off. What I don't understand is whether this manuever was predictable, given the properties of the airplane and control system. Like I said, if it happened in the air while following reasonable proceedures, operating inside the approved envelope, then the incident must be totally understood. Also, the design criteria for airliners must be revised to protect future airliner designs, and proceedures for flying existing airliners must be changed. Although incompetent or negligent piloting can break bits off, I'm not convinced that this guy was just an incompetent. I'm uncomfortable with blaming a person who can't defend himself, and I think we can make sure this does not happen again.

I'd like to make another observation. Typical designs using composite materials produce component strengths twice as high as metal ones, that is, they will break at loads twice as large as a metal part designed to the same in-flight loads. It is not because the composites have magical properties. It is because we are not allowed to take advantage of "typical" strength. In order to achieve the confidence that the metal designs have, we discount the strength of typical material about 50%, a higher safety factor, if you will. That discount is taking into account temperature and humidity, so called "effects of defects", and unique composite material quality control concerns. The fact that a composite material tail broke in the air might mean that this fin got three times the design load on it. Yikes! That is very scary.

THE RAMPTOR ENGINEERING TEAM <img src=icon_smile_big.gif border=0 align=middle>
"Who cares if it works? Does it look good on the ramp?"

_________________
????

<BLOCKQUOTE id=quote><font size=1 face="Verdana, Arial, Helvetica" id=quote>quote:<hr height=1 noshade id=quote>
I have a feeling that this wouldn't even have been allowed in an A320 and up nor any Boeing aircraft. The A300 is just an old POS Scarebus, the new ones are acutally quite good aricraft even if they are built in some part in France.

<hr height=1 noshade id=quote></BLOCKQUOTE id=quote></font id=quote><font face="Verdana, Arial, Helvetica" size=2 id=quote>

If they miss calculated the the vertical fin design condition on the A300, why should I think they did a better job on the A320? Why should I take the word of the Europeans when they certify any airplane as airworthy? Can U.S. designs get themselves into trouble like this, with important bits falling off? Are all modern airliners vulnerable? Are there other errors in loads calculation lurking, the horizontal stabs, wings etc., etc.? We carry a safety factor of 50% higher strength than the predicted in-flight loads. This strength level is proven by ground test. In this case, the maneuver they found themselves in exceeded the predicted by >2. I hope it was pilot error, otherwise we're in a heap o' trouble.

THE RAMPTOR ENGINEERING TEAM <img src=icon_smile_big.gif border=0 align=middle>
"Who cares if it works? Does it look good on the ramp?"

_________________
????

This is not Pilot Error. This was nothing more than a Tail Shearing in flight, IF you look at the Gain Loads, In order for a pilot to get that type of momentum (parabola) going he would have had to be jumping up and down on the rudder. This aircraft departed flight in less than 1 Minute with a 20 second Pause before the Occilations began, Their is also Roll Coupling involved with a condition that Leans into a bank with a Rudder input pause for 20 secs. This aircraft continued the roll as the yaw was fluttering.

Ive spoken with sevral Friends flying Airbus Aircraft. From their descriptions. The gains exhibited in that Flight Exceed what they could simulate in the REal World and in the Simulator. Their has been a big stir about this for a while. I would expect ot see this get worse in the news.

"The power to Destroy the planet, is insignifigant to the power of the Air Force----Mudd Vader

Stress, I'm aware of the 50% load limit. The Investigating Officer In Charge of the crash, Robert Benzon, said the tail broke off at only 29%. Could it be because of the repair of earlier damage on the flight to Puerto Rico?

Go Ugly Early !
No body's ugly after 2:00 A.M.

_________________
Go Ugly Early!
Nobody is ugly after 2 A.M.

<BLOCKQUOTE id=quote><font size=1 face="Verdana, Arial, Helvetica" id=quote>quote:<hr height=1 noshade id=quote>
Stress, I'm aware of the 50% load limit. The Investigating Officer In Charge of the crash, Robert Benzon, said the tail broke off at only 29%. Could it be because of the repair of earlier damage on the flight to Puerto Rico?
<hr height=1 noshade id=quote></BLOCKQUOTE id=quote></font id=quote><font face="Verdana, Arial, Helvetica" size=2 id=quote>

Sometimes definitions help discussion.

<b>Limit Load</b> is the maximum load expected in flight. It is calculated by reliable established proceedures so that design can proceed. It is eventually confirmed by in-flight measurements on the flight-test aircraft while performing the actual design maneuver.

<b>Ultimate Load</b> is the required strength level. It is the limit load multiplied by a safety factor, usually 1.5, or 50% higher than limit. Strength at ultimate load is usually demonstrated on a ground test article. Sometimes it can be shown by analysis.

I believe the 29% number Hogmender mentions must refer to an estimated load 29% above ultimate, or 1.5 x 1.29 = 1.93 times larger than limit. Two possible reasons for this overload are bad pilot technique or an underestimated design load (too low).

Back in 2002 the NTSB detailed what potential causes they were investigating. Here they are with some lame commentary from me:

1) The adequacy of the certification standards for transport category aircraft.
<i>I guess they were satisfied with the way the design loads were calculated and that the selected maneuvers that were critical to the fin were correct. Maybe I would agree with them if I had all the facts. Finding a problem here would have serious implications for the structural integrity of all airliners. This is not a preferred path.</i>

2) The structural requirements and integrity of the vertical stablizer and rudder.
<i> They are also satisfied with the demonstration of strength for the entire fin/rudder design and the data supporting the basic material properties, any repairs performed etc.</i>

3) The operational status of the rudder system at the time of the accident.
<i>All the data they have says the rudder system performed as expected and no malfunctions can be identified.</i>

4) The adequacy of pilot training.
<i>This was a major bone of contention, along with...</i>

5) The possible role of pilot actions in the accident.
<i>Bingo. This is their selection for the primary cause. They are convinced that an alternate pilot technique would have saved this airplane. Since the successful pilot technique might require weeks of simulator time to work out (if the simulator has any fidelity in this out of control scenario), and some practice to master, I would rather make the fin stronger instead of depending on the pilot to have "the right stuff". Since all these airliners are flying around now, it is somewhat of a dilemma. You make the call.</i>


THE RAMPTOR ENGINEERING TEAM <img src=icon_smile_big.gif border=0 align=middle>
"Who cares if it works? Does it look good on the ramp?"

_________________
????