Reply #364: Another voice of sanity [View All]

Subj: Complete version of Article
Date: 6/20/2003 9:59:58 AM Central Standard Time
From: reesf@kih.net
To: tbieter@aol.com
Sent from the Internet



(This is the complete version - it contains the footnotes and bio.)
------------------------------------
Duluth, MN June 19, 2003

A pilot critiques the Wellstone crash

By William H. Rees
Reader Weekly

The discussion about the cause of the Wellstone plane crash has got
completely out of hand. The Reader published five or six articles by Prof.
James Fetzer about it and as a result conspiracy theorists of various
stripes slander each other in scores of emails and a lawsuit is threatened.
None of this should have happened. Fetzer's series should have ended with
his first article, or before.

Fetzer began by wrapping himself in the Scientific Method and then promptly
violated it.

Fetzer wrote in his fifth paragraph, "One of the most basic principles of
scientific reasoning. . .the requirement of total evidence, insists that, in
the search for truth, all the evidence whose truth or falsity or presence or
absence make a difference must be taken into account." I agree completely,
and that is why I write. I'm a retired Air Force pilot with 5,500 flying
hours, mostly in jet fighter types and quite a bit of it in weather. I see
things in the reports of the Wellstone crash that should end this debate
once and for all.

Fetzer's conspiracy theory depends absolutely on neither pilot error or
weather being the cause of the accident. If either cannot be written off
then his whole conspiracy theory has no reason for existing. He wrote them
both off by his own judgement, although he is neither an experienced pilot
or a weather expert. He cited "evidence" that is merely his own uninformed
opinion of mostly hearsay, ignored or trashed the opinions of people better
qualified than he, and then proceeded to his conspiracy. He is no
scientist.

Fetzer wrote, "The key to understanding the crash appears to be the complete
cessation of communication between the pilots and the control tower...It
(the accident cause) had to be something that caused a loss of communication
as well as of control." The second part of this proposition is partly true.
The first part is simply silly.

Of relatively small importance but worth noting is the fact that there never
was any communication between the pilots and the control tower. The
Eveleth-Virginia Municipal airport has no control tower. Planes arriving
under instrument flight plans are under the control of Duluth Approach
Control. Planes entering traffic there announce their presence to each
other over the "Unicom" radio channel. It cannot be said that there was a
"cessation" of communications that never existed.

Fetzer is much concerned that the pilots did not report their problem on the
radio. He thinks this is very suspicious, and indeed it does need
explaining. The explanation is that pilots very rarely tell anybody when
their plane goes out of control.

When a pilot gets in trouble with aircraft control and thinks it may be
possible to pull out of it he has much more important things to do than push
a mike button and tell somebody about it, and if he does pull out of it he
may very well not want the world to know anything happened (I've been there,
done that.). "Loss of Control" means certain death unless control is
regained in time, and thus leads instantly to attempts to regain control and
absolutely nothing else is of the slightest importance whatever until
control is regained and some composure regained. Once those happy events
occur, the next step is to do whatever is necessary to continue the flight
to an acceptable conclusion, followed often by figuring out what to tell
anybody about it, if anything at all. This can all happen in a very few
seconds, but it is only at this point that radio transmissions are likely to
be made.

Once a pilot realizes for sure he's going to die in five seconds or so it is
quite likely his mind focuses on things other than telling somebody about
it. Typically, the cockpit voice recorder captures "Oh s--t!" on the
intercom and nothing else.

So much for "cessation of communications." Now to loss of control.

Aircraft accidents are almost always the result of a combination of
unexpected events or factors. Had any one of them been different the
accident would not have occurred. And, it is always fully-qualified pilots
who have accidents, for otherwise they would not have been flying the plane.
Needless to say, no pilot ever has more than one fatal accident.

From the news accounts it is apparent that both pilots had serious
weaknesses. "(The captain) just seemed real slow. Always hitting the wrong
things, saying wrong things... .forgetful and made random errors. . ..had a
bad feeling about him.... a smart guy, intelligent, but he was lacking
something,"<1> and two copilots had to take the controls from him for
failing to maintain altitude in weather.<2> The captain made mistakes of a
type that indicates a disorderly thinking process. He was weak and he knew
it and if the word got to management his flying career would abruptly end.
He let his copilots do most of the flying, <1, 2, 3> which is good to give
them experience, but also reduces the odds of them seeing him make mistakes
and thus may actually have been a cover-up. He had a confidence problem.

"(The co-pilot had) 701 hours of flying time...tended to be 'fixated' during
his approaches and 'airspeed and torque (engine power) would...get too low.'
..had a habit of keeping both hands on the flight controls and had to be
reminded to keep one hand on the throttles to control airspeed...had trouble
grasping airplane system knowledge."<2> He clearly tended to have "tunnel
vision," a narrow awareness of what was going on with the plane and around
him, not on top of the situation but rather a reactor to it.

I've known pilots like each of them, and covered for some of them, and
limited the missions for others until they gained or regained proficiency.
As a flight examiner I've given check rides to pilots I already knew were
weak, and passed them because they did what the check required of them.
Check rides are objective and cannot test every combination of
possibilities. They are like photographs, a slice of time, not like any
other, and weak pilots can pass them when they're having a good day. All
pilots have weaknesses, and bad days, which is one reason for having two
pilots in a cockpit. A proficient captain can back up a weak copilot while
he gains experience and competence, and a good copilot can keep a weak or
fading captain out of trouble until he retires or flunks a check ride. But
two weak pilots together in the same plane are an accident looking for a
place to happen.

Weather often creates good places for accidents to happen. Fetzer wrote
that "the weather was perfectly fine" for the flight. Perhaps he means that
weather did not by itself cause the crash, which is true, but weather is
almost never the sole cause of a crash. It is the things pilots DO about
the weather that causes weather-factor crashes. Weather certainly did
affect this flight.

Weather dominated the captain's thinking before and during the flight and
determined the rules of the flight and required an instrument approach at
destination. The weather was forecast to be above legal landing minimums,
for if not the captain would have had no decision to make. He could have
simply canceled the flight. But he had to make a decision and he didn't
like it. The weather was for him marginal but since it was legal it was
difficult to justify not going. He couldn't back out gracefully. He took a
chance to avoid having to explain why he wouldn't fly to an airport other
pilots were flying to. Perhaps he didn't realize his copilot's weaknesses.
The flight would be routine, provided nothing unexpected happened.

Something unexpected happened. On the descent they may have picked up a
little ice on the wings, not enough to see but enough to disturb the airflow
a little and raise the stall speed a few knots. They may have pulled the
power way back and let the plane coast smoothly from cruise altitude down to
approach altitude, then leveled off and forgot to push the power up. This
is easy to do and easy to miss when both pilots are a little disorganized
and not functioning as a team. They were going off the final approach
course, which would have required the pilot's attention to correct back, and
perhaps make him think there was a crosswind and try to mentally calculate
how much correction to apply, and thus reduce his attention to aircraft
control. By now they were a little behind the airplane, a little task
saturated, and they knew it. If the speed was low, and/or the wings a
little iced, and a correction toward course was made roughly or jerkily, the
plane could easily have stalled, or at least a stall warning may have
sounded which would be totally unexpected and thus alarming. They were
still in the weather at the time, and probably they seldom practiced stall
recoveries purely on instruments, possibly never in actual weather.

To the above add a probable complication. A Star Tribune article of March
27th reported that an NTSB chart (probably based on FAA radar) showed the
plane "flying at a speed of 170 knots 5 miles from the runway.<4> That's
pretty fast for that distance, and consistent with coasting down from
altitude with the power back and the landing gear still up, possibly to get
though an icing level quickly. The published minimum altitude at five miles
is 2900 feet, over 1,500 feet above field elevation. They had to get rid of
about 40 knots and at least 1,000 feet of altitude and lower the landing
gear and stabilize the power and airspeed and get below the clouds and be
pretty well lined up with the runway in the next four miles in order to be
able to land. They were considerably behind events, behind the airplane as
pilots would say, and a lot of activity would be compressed into those four
miles. Too much?

The chart showed the plane at 129 knots less than three miles from the
runway, <4> which was about right. The final approach should have been
stabilized at least by then; gear down, flaps set, power and airspeed
stabilized, on course, rate of descent stable. But a mile later the chart
showed the plane at only 76 knots and still in the clouds at just above 400
feet above the ground. <4> It's impossible to know from the published story
the accuracy and scan rate of the radar that provided this information, but
assuming the numbers are close two things stand out. One, there are three
published instrument approaches (depending on the navigational aid used) for
this runway and the lowest minimum ceiling for any of them is 400 feet - the
weather was actually at or below published minimums. The captain had been
briefed before takeoff that the ceiling was at 900 feet and probably
expected to break out of the clouds three or four miles from the runway.
That they did not may well have upset his planning and further damaged his
composure. Two, the decrease of airspeed between three and two miles was
drastic. One possible reason was that the power was pulled way back, and
then the landing gear was dropped (as in "Oops! Forgot something!"), leaving
them with low power and high drag and thus a very rapid bleed-off of
airspeed and changing control pressures and just too many things changing
all at once.

Now add one more unexpected factor, the immediate reaction to a poorly
understood emergency by two weak and apprehensive pilots. They may well
have BOTH tried to recover the aircraft, at the same time, in conflicting
ways, and they had no time to sort it out. At low speed any abrupt or
uncoordinated maneuver would cause a stall. It may have taken as little as
5 seconds from their first inkling of serious trouble to enter an
unrecoverable situation.

All of these possibilities lead to the one most likely thing, stall. Fetzer
wrote that "actual tests with King Air Al00s have shown that they do not
stall out until air speed fall below 70 knots...(the stall warning alarm)
triggers off at 85 knots." Aircraft stall is not nearly so simple.

There is no such thing as a "stall speed", except at a specific combination
of weight, configuration and maneuver. Stall is a function of angle of
attack, the angle at which the airflow strikes the wing, and above a certain
angle of attack the airflow over the top of the wing becomes turbulent and
lift is lost. Wings are designed so that the critical angle of attack
varies a little along the wing, e.g. stall may begin near the wing root
causing mild turbulence the pilot can feel and recognize as a warning, while
the outer part of the wing is still flying and aileron control is still
effective.

In perfectly straight and level flight (called "1 G flight") at a specific
weight there is in fact a stall speed (published in the flight manual), a
knot above and there's no stall, a knot below and there's "burble" the pilot
can feel, and X knots below that the whole wing stalls. Stall
characteristics vary with the wing design, but commonly there is a nose
drop, some times accompanied with the drop of one wing (and the amount and
balance of power on the plane at the time may influence all that). In a
turn the inside wing is a tiny bit slower and it will stall first and drop,
increasing the bank and causing the nose to drop even more.

To get out of a stall the pilot must reduce angle of attack, usually by
lowering the nose (towards the ground). He will almost certainly also add
power but some altitude loss is inevitable. The deeper the stall the longer
to recover and the greater the loss of altitude and pilot composure. He
must return to level flight which means he must stop the descent, which
means he must have more lift than he needed for level flight, which means he
must have more speed than he had at the time of the stall or he will exceed
the critical angle of attack and stall again. This is called a secondary
stall and it is quite common when a pilot is in a bit of a panic. Flying
requires thinking ahead and staying ahead, and thinking under pressure, and
composure.

Given a specific stall speed in 1 G flight at a given weight, there is a
different stall speed (but always the same critical angle of attack) at all
other conditions of flight. In a 60 degree banked level turn the plane must
pull 2 Gs, which means the wing must generate twice as much lift as in 1 G
flight, which means the speed must be high enough to generate double the
lift without exceeding the critical angle of attack. If the speed is less
and the pilot attempts a sudden 2 G turn, or pull out, the wing will exceed
the critical angle of attack and stall. The pilot can then immediately
release some back-pressure on the elevator control and reduce the angle of
attack and thus usually stop the stall, but this interrupts the maneuver.
If I recall correctly, a 30 degree bank level turn requires 1.14 Gs, and
thus 1.14 times the lift and some higher speed than 1 G level flight. Any
amount of pull-up from that point will increase the stall speed and any
amount of push-over will reduce it.

Stall warning devices usually measure angle of attack rather than speed, so
it is misleading to say that they function at a given speed. Rapidly or
roughly increasing the angle of attack may result in going through the stall
wamings and immediately entering what is called "high-speed" or
"accelerated" stall.

So, does all this tell us what caused the Wellstone crash? No it does not.
The precise causes will probably never be known. The pilots probably didn't
realize them all by the moment of their death. All it does is tell us that
pilot error was quite possible, perhaps highly probable, and thus there is
no need or justification for turning to conspiracy theories.

The question does arise, may be asked in court, "Shouldn't somebody have
prevented these two pilots from flying together?" That would depend on how
pilot supervision is commonly done in commercial aviation. In a military
fighter squadron supervision is tight, flight commanders regularly fly with
their pilots and know their weaknesses and adjust their challenges to their
ability. In bomber or transport units that use formed crews the same people
fly with each other most of the time and weaknesses are known and dealt
with. In units and commercial operations that do not use formed crews
trouble-spotting and supervision is much more difficult. Suppose a captain
makes five dumb mistakes in a month but each with a different copilot. None
see a pattern, and everybody makes a dumb mistake now and then, so nobody
reports a problem. The captain knows of course, and he should get some more
training or other help, or quit. But if he lacks integrity and/or common
sense and needs the job he may just try to fake it and hang on, and he may
get away with it for a long time. Short of using formed crews, pilots and
copilots who always fly together, I don't know any way to avoid the problem.

In flying as in all other serious endeavor there is no substitute for
personal integrity.

--------------

Footnotes:

<1> "More criticism of Wellstone pilots," Minneapolis Star Tribune 3/6/03 by
Tony Kennedy and Greg Gordon.

<2> "Wellstone's pilot balked at flying on morning of crash," Star Tribune
2/22/03 by Kennedy & Gordon.

<3> "Pilot in command of Wellstone flight often let copilot fly," Star
Tribune 12/22/02 by Kennedy & Paul McEnroe.

<4> "Wellstone plane was flying too slowly, investigators say," Star Tribune
3/27/03 by Tony Kennedy.

The author is a retired Lt. Col. in the Air Force, formerly Director of
Operations and Training for the 23d Air Division Headquarters at Duluth Air
Base and an active jet pilot there. He flew all-weather fighter interceptor
jets in several assignments and combat as a forward air controller over the
Ho Chi Minh Trail in Laos during the Vietnam War.