Examination of wreckage by the
Air Accidents Investigation Branch
51. On 3 June 1994 Mr Tony Cable, a Senior Inspector of the
AAIB, flew to the crash site and remained closely involved thereafter
in the investigation until he signed on 5 January 1995 the statement
to the Board which he had prepared. This detailed report, which
ran to some 66 pages and is published in HL Paper 25(i), resulted
not only from Mr Cable's own work at the site and at AAIB Farnborough
but also from opinions expressed by the relevant manufacturers
to whom salvaged components or information had been provided.
52. The aircraft was very severely damaged: 80
per cent of the fuselage structure was appreciably damaged by
ground fire, of which around 20 per cent was destroyed (AAIB statement
para 7.1). It had neither cockpit voice recorder nor accident
data recorder. This presented a formidable task to Mr Cable who
expressed the view to us that "the evidence was remarkably
thin" (QQ 956, 968, 1013). From the final position of the
aircraft, which had by then extensively broken up, he estimated
that at the initial impact the flight path was 20 degrees up relative
to the horizontal, the pitch angle was 30 degrees nose up with
a 5-10 degrees left roll and the ground speed was approximately
147 knots, as disclosed by detailed examination of the cockpit
ground speed and drift indicator (AAIB statement para 6). The
left rudder pedal appeared to have been applied 77 per cent (AAIB
statement para 7.4.3, 7.4.9).
53. The aircraft was fitted with a Racal Avionics
"SuperTANS" Tactical Area Navigation System providing
navigation information from two independent sources. The system
enables a number of way points to be fed into it before a flight.
When flying from the point of departure to the first way point
the screen[20]
shows bearing, distance and "time to go" from the aircraft's
current position to the way point. When the pilot alters the system
from the first way point to the second, the distance and bearing
of the former are replaced on the screen by those of the latter
and so on as way points are progressively changed. Racal confirmed
that the system was performing perfectly at the time of loss of
power and extracted from its memory the information that way point
B had been selected when way point A was 0.81 nautical miles distant,
bearing 018°. The distance from the way point change to the
point of impact was 0.95 nautical miles. The system gave no information
as to height or time at the way point change but had recorded
that at approximately 15-18 seconds before power down the aircraft
was at a height of 468 feet ± 50 feet (Board report para
49). The manufacturers have told us that "18 seconds is likely
to be a better estimate".[21]
The TANS had also recorded that the height above sea level at
impact was 665 ft[22],
whereas in fact it was 810 ft. The investigating board noted this
discrepancy (para 49); they considered it probably due to "the
mechanics of the crash and the developing fireball", but
we know of no evidence to support this. The TANS is not intended
to act as a Flight Recorder or what is colloquially known as a
Black Box, and the information referred to above was achieved
by a somewhat complicated and ingenious method of extraction employed
by Racal.
54. The AAIB considered the engines and controls
and because of the reported FADEC service difficulties investigated
the DECUs in detail. DECU no. 2 remained partially functionable
with deficiencies consistent with impact damage, and with no faults
or exceedances traced in its memory of the last flight. DECU no.
1 had suffered gross fire damage with part of its casing melted
away and severe damage to the interior components whereby its
memories of exceedance and fault listing had been destroyed.
55. A detailed examination of the flight control
mechanical linkages was carried out and the AAIB stated, "No
evidence to suggest a control jam was found, although such a possibility
could not be excluded, given the level of system damage"
(para 7.4.2).
56. Important parts of the hydraulic flight control
systems were housed in a small closet, colloquially known as the
"broom cupboard", at the rear of the cockpit. There
were two control pallets containing respectively 23 and 26 threaded
inserts for component attachment. On 10 May 1994 the thrust balance
spring attachment bracket on the aircraft's thrust/yaw control
pallet had detached; this was due to the somewhat inadequate method
of attaching the inserts to the pallet. This detachment had resulted
in an undemanded flight control movement (UFCM) in the collective
system.[23]
An engineering report of the following day relating to this incident
stated among other things,
"Detachment of the bracket within
the flying control closet during flight could present a serious
flight safety hazard, with the danger of a detached bracket fouling
adjacent flying controls".[24]
57. After the accident the investigators found that both inserts
for the thrust balance spring attachment bracket had detached
as well as most of the other inserts to both pallets. The AAIB
stated, "as an insert could apparently pull out of the pallet
without appreciable distress to the components necessarily resulting,
the possibility that insert(s) had detached prior to the accident
could not be dismissed" (para 7.4.2). In the Flight Control
Summary the AAIB reiterated that "the possibility of control
system jam could not be positively dismissed" and further
stated that "little evidence was available to eliminate the
possibility of pre-impact detachment of any of the pallet components"
(para 7.4.9).
58. The investigators also found a considerable
quantity of very small metallic particles and four fine metal
slivers in the hydraulic system of the lower control actuators
which form part of the flight control system. They concluded that
this contamination had occurred pre-impact, but that it had not
contributed to the accident (AAIB statement para 7.4.4).
59. With a view to ascertaining the manoeuvre
necessary to produce the initial impact conditions the airframe
manufacturers, Boeing, at the request of the Board produced mathematical
models simulating the aircraft's final behaviour from a postulated
range of initial steady flight conditions. In order to do this
they were provided by Mr Cable with what he considered to be the
pitch attitude and flight path angle of the aircraft at initial
impact together with the extensions found in two components of
the flight control system, namely, the differential airspeed hold
(DASH) actuators and the longitudinal cyclic trim actuators (LCTAs),
and other pieces of information derived from the site (Q 950 and
p 86 of HL Paper 25(ii)).
60. The Boeing simulation tried a series of input
conditions suggested by AAIB, but found that close simultaneous
matching of the predicted conditions with the control criteria
was possible in only a few cases. A ready match was found where
initial conditions combined an airspeed of 150 knots[25]
with a rate of climb (ROC) of 1000 ft/min (AAIB statement para
8). In order to achieve a maximum ROC, which far exceeds 1000
ft/min, airspeed requires to be reduced to 80 knots or below.
Boeing then estimated that, some four seconds before the initial
impact, input was applied to the controls to achieve a "cyclic
flare", increasing the ROC and decreasing forward speed by
bringing the nose up.
61. At initial impact Boeing's simulation produced
the following result (AAIB statement para 8):
Airspeed - | 135 kt
|
Normal Acceleration - | 2.2 g
|
Rotor Speed - | 204 rpm (91%)
|
DASH Extension - | 23% |
LCTA Extension - | Virtually Fully Extended
|
Aircraft Pitch Attitude - | 31° Nose Up
|
Aircraft Roll Attitude - | 5° Left
|
Aircraft Yaw Attitude - | 1° Left
|
Flight Path Angle - | 20° above the horizontal
|
Climb Rate - | 4670 ft/min |
Horizontal Distance Travelled - | 822 ft (250 m)
|
Vertical Distance Travelled - | 128 ft
|
Groundspeed - | 158 kt |
15