| On
December 22, 1996, an Airborne Express DC-8 N827AX
with 6 crewmembers on board (3 flight crew and
3 maintenance/avionics technicians) crashed in
mountainous terrain in the vicinity of Narrows,
Virginia. The crash was the result of the crew's
failure to fully recover from a stall that they
had intentionally initiated as part of a Functional
Evaluation Flight (FEF), which was required after
modifications had been performed on the aircraft.
Although this accident involved a modern jet airliner,
there are valuable lessons to be learned for pilots
of any aircraft.
The crew of this DC-8, N827AX, departed from Piedmont
Triad airport in Greensboro, North Carolina late
in the afternoon of December 22. They had planned
to depart earlier, but maintenance delayed their
takeoff time. ATC assigned them a block altitude
of 13,000 to 15,000 feet mean sea level for the
checks they needed to perform. About 13 minutes
before the crew initiated the stall, they had
reported flying in and out of some clouds and
had reported some ice build-up, but cockpit voice
recordings indicated that they were clear of icing
conditions before the stall was initiated. The
extent of any ice build up on the aircraft at
the time of the stall is unclear. It was also
dark outside when the crew initiated the stall,
so outside visual attitude references were limited.
Guidance on FEF profiles at the time recommended
that maneuvers such as stall series be performed
during the daylight hours.
After other checks had been completed, the crew
decided to initiate the stall series. The plan
was to record the airspeed at which the stick
shaker activated and the airspeed of the first
stall indication. At that point, they would recover
from the stall. They expected a stick shaker at
128 knots and first indication of the stall at
122 knots. The approach to stall was uneventful,
as the pilot slowed the aircraft down at about
1 knot per second. The PF (pilot flying) noted
"some buffet" at 151 knots at 18.08:06,
and the crew commented that this was early for
buffet (Whether this was due to some ice build-up
or the fact that the control surfaces had been
re-rigged during maintenance is unclear). At 18.08:09,
the sound of rattling was heard on the cockpit
voice recorder. At 18.08:11 the flight engineer
stated "that's a stall right there…ain't
no [stick] shaker." The PF called "set
max power" at 18.08:13. Perhaps confusing
to the crew was the fact not only that the stall
occurred at a higher than expected airspeed, but
also that the stick shaker failed to activate.
For the next 8 seconds, the PF continued to hold
the nose up, maintaining a relatively constant
pitch attitude. Popping sounds were also heard
coming from one or more of the engines, and engine
indications of surging were present, which can
happen when airflow into the engine intakes is
at an excessive angle. Airspeed continued to decay,
and the aircraft began descending as the stall
progressed. At 18.08:30 the PNF (pilot not flying)
stated "You can take a little altitude down…"
He was implying to the PF to push forward on the
yoke. But at 18.08:42, he added, "Start bringing
the nose back up." For the next 56 seconds,
the DC-8 continued descending and began a series
of roll reversals. At times, the PF did move the
control column forward somewhat, but the data
recorder indicated several instances where the
control column was full aft, which corresponded
with un-commanded aircraft downward pitches. The
crew failed to recover from the stall and impacted
terrain at approximately 3400 feet mean sea level
in a 52-degree, left wing low and 26-degree nose-down
attitude.
Complicating successful recovery was the lack
of outside visual references. Further, once the
aircraft began to descend, it entered IMC conditions
and remained there until shortly before impact.
Most civilian aircraft do not have any on-board
angle-of-attack visual reference, despite the
fact that most military aircraft have such references.
Angle-of-attack indicators and their associated
equipment are not complicated devices, and many
official agencies have recommended repeatedly
that they be installed in airliners. It is possible
that, without adequate visual references of pitch
attitude or angle-of-attack, and with an inoperative
stick shaker, that the crew of this DC-8 may not
have realized in the descent that the aircraft
was still stalled. It appeared as though their
priority may have shifted from a stall recovery
to a nose-low unusual attitude recovery as they
retarded the engines toward idle, pulled the yoke
well aft as the nose pitched down, and attempted
to roll wings level at one point with full left
rudder deflection as the aircraft rolled to over
100 degrees of right bank.
It is worth noting the stall characteristics of
an aircraft such as the DC-8, and how that differs
from the stall characteristics of other airplanes.
Most straight wing aircraft have favorable stall
characteristics. Their wings stall at the root
first providing ample warning that the stall is
progressing (fuselage buffet as the turbulent
air flows off the wing roots and past the fuselage).
They have fairly good lateral stability since,
at least in the early stages of a stall, the outer
portions of the wing are un-stalled, and the ailerons
are somewhat effective. A modern swept wing airliner
such as the DC-8 has somewhat different stall
characteristics. Since aft-swept wings tend to
stall at the tips first, the Center-of-lift may
move ahead of the Center-of-gravity (CG) at the
stall causing a pitch-up moment. The aircraft
may begin descending in a nose up attitude unless
positive forward pressure on the yoke is applied
(Straight-wing aircraft will usually pitch down
on their own accord when the stall occurs, as
long as they are within aft CG limits.). That
is why swept-wing aircraft have stick shakers
that give an artificial warning of impending wing
stall. Some even have stick pushers to force the
aircraft to a lower angle-of-attack before the
stall progresses too far. Many types of aircraft
will tend to roll or yaw if recovery from the
stall is delayed, and swept wing aircraft are
particularly prone to becoming laterally unstable
as the stall progresses. (Of note is the fact
that the simulator that this crew had recently
trained in did not exhibit lateral instability
if held in a stall). To further complicate the
problem, those aircraft with engines mounted underneath
the wings can cause a further pitch up, since
the engines' thrust lines are below the aircraft's
CG. The pitch-up associated with adding power
can cause the stall to worsen, if the yoke is
not moved forward to counter this tendency, or
if emphasis is not placed on lowering angle-of-attack
first with forward pressure on the yoke. Swept-wing
aircraft do not normally have the luxury of engine
or propeller wash over the horizontal tail and
elevator, or stabilator, to aid in pitch control.
In the past, if a pilot encounters an impending
stall in such an aircraft, he has been taught
to hold the pitch attitude and apply maximum power
to minimize altitude loss and to "fly"
out of the stall. The success of this recovery
lies in the fact that a stall has not yet occurred
(the stick shaker will typically activate at an
airspeed 5-10% above the stall speed.). It is
not really a stall recovery that most of us would
use in a typical general aviation aircraft for
instance. It is not really a stall recovery at
all, since a stall has not occurred. Guidance
did exist at the time of this accident both for
the DC-8 and other similar aircraft that recommended
pushing forward on the yoke to lower angle-of-attack
first, then adding maximum power.
The stall characteristics of the DC-8 are relatively
good for a swept wing aircraft. The crew of N827AX
obviously did not anticipate any problems. The
crew noted early buffet, but the flight data recorder
indicated that the actual stall occurred within
a few knots of planned stall speed. Published
guidance on the DC-8 warned that, when approaching
a stall, aircraft buffet does not always precede
the stick shaker, and may occur simultaneously
with the shaker. But on this night, the stick
shaker did not activate at all. The pilot's decision
to hold the DC-8's pitch attitude constant as
maximum power had been applied would have been
uneventful if the stall had not progressed as
far as it did. This leads to an important consideration
involving pitch attitude and angle-of-attack.
The pitch attitude is typically defined as the
angle between an aircraft reference such as its
longitudinal axis and the horizon. Angle-of-attack,
conversely, is the angle between the wing's chord
line and the relative wind, which could be coming
from anywhere. We can approximately see the aircraft's
pitch attitude by referencing the angle that the
wings or nose makes with the horizon, or by referencing
the attitude indicator or other attitude reference.
Certainly if we were to approach the stall from
a steady 1-G deceleration in level flight, critical
angle of attack and our pitch attitude would be
roughly the same. We would "see" the
stalling angle-of-attack by referencing the position
of the wings or aircraft's nose with respect to
the horizon. But anyone who has practiced power-on
stalls extensively knows that higher pitch attitudes
are reached when the stall occurs, even though
the wing always stalls at the same critical angle-of-attack.
And if you have ever encountered a stall in descending
flight or perhaps while recovering from a dive
following a spin recovery, you know that a stall
can occur at "negative" pitch angles,
when the nose is definitely below the horizon.
Consider again the DC-8 mishap. Because the stick
shaker was inoperative, the pilot actually stalled
the wings of the DC-8. Airspeed continued to decay,
and, because lift decreased at the stall, the
aircraft began descending. In the diagram in Figure
1, the aircraft on the left is at some angle-of-attack
either below or at critical angle-of-attack. Since
the flight path is level (as indicated by the
relative airflow parallel to the horizon), the
angle-of-attack and pitch angle are approximately
the same (any typically small difference between
the two would lie in the aircraft reference used
to define pitch angle, which may not be the chord-line
of the wing as it is for angle-of-attack, and
in any built in angle-of-incidence of the wing
as it is attached to the aircraft). The same aircraft
depicted on the right side of the diagram has
now entered a descent. Note that the relative
wind is coming up at the aircraft from below.
The pitch attitude of the aircraft has not changed,
but the angle-of-attack has increased in the descent
due to the upward flow of the relative wind. In
the pilot's desire to hold the DC-8 at constant
pitch attitude he unwittingly allowed the angle-of-attack
to increase further into the stall as the aircraft
began a descent. This would not be readily apparent
by noting the pitch attitude of the aircraft (and
references were limited anyway), but there were
other cues that the crew was not successfully
recovering from the stall. They included:
|
