F I N A L R E P O R T

concerning the accident occurred

on 31 March, 1992,
to the BOEING 707 registered 5N-MAS
(Nigeria) - Trans-Air Limited Company.
MINISTRY OF EQUIPMENT, TRANSPORTS AND TOURISM
GENERAL INSPECTORATE FOR CIVIL AVIATION AND METEOROLOGY

FRANCE

MINISTRY OF EQUIPMENT, TRANSPORTS AND TOURISM

Paris, 28 June, 1993

R E P O R T

concerning the accident which occurred on 31 March, 1992

to the Boeing 707 registered 5N-MAS (Nigeria),
operated by the Trans-Air Limited Company
 F O R E W O R D

This report reflects the technical point of view of the French Accident
Investigation Bureau BEA team assembled to investigate the circumstances
and causes of the accident.

In compliance with Annex 13 to the Convention on International Civil

Aviation, the analysis was not carried out and the conclusions and safety
recommendations were not drawn up with an objective to apportion blame
or individual or collective liability. Basically, this technical
investigation aims at preventing further accidents. Consequently, since
doubt must benefit safety, some recommendations that have been suggested
refer to points whose accurate demonstration has not always been possible,
or in some case, are not directly connected with the causes of this
accident.

This report has been drawn up after thorough investigation and, therefore,
is based on knowledge which may notably differ from that prevailing when
the accident took place.

Finally, although the people and the organizations, whose opinions have
been considered as relevant, have been requested to submit their
information in due time, this inquiry has been carried out without using
contradictory proceedings.

Consequently, using this report for other purposes than the prevention
of further accidents might be misleading.

SPECIAL FOREWORD TO ENGLISH EDITION

This report has been translated and published by the French Accident
Investigation Bureau to make its reading easier for English speaking
people. As precise as this translation may be, please refer to the original
text in French.
 C O N T E N T S

Page

SYNOPSIS .................................1

1. FACTUAL INFORMATION

1.1 History of the flight..............2

1.2 Injuries to persons................4
1.3 Damage to aircraft.................4
1.4 Other damage.......................4
1.5 Personnel information..............4
1.6 Aircraft information...............6
1.7 Meteorological information........10
1.8 Aids to navigation................12
1.9 Telecommunications................12
1.10 Istres aerodrome information......13
1.11 Flight recorders..................14
1.12 Wreckage information..............17
1.13 Medical and pathological
information.......................19
1.14 Fire..............................20
1.15 Survival aspects..................20
1.16 Tests and research................20
1.17 Declarations......................29

2. ANALYSIS

2.1 Loss of the two right engines.....31

2.2 Evolution of the inboard midspar
fitting fatigue crack.............32
2.3 Monitoring of the pylon fittings..33
2.4 Right wing fire...................34
2.5 Distress piloting.................36
2.6 Control assistance................38

3. CONCLUSIONS

3.1 Data given by the inquiry.........40

3.2 Causes of the accident............41

4. SAFETY RECOMMENDATIONS...41

Photos and boards

Annexes
 S Y N O P S I S

Date of accident : Aircraft :

Tuesday, 31
March, 1992 Boeing 707-321CH
at 08.40 hrs UTC (°) Registration :5N-MAS

Place of accident : Owner and operator :

Flight over France Trans-Air Ltd

(flight level 330) 5, Bompai Road
Kano
PO Box 2773
Nigeria

Type of flight :

Charter flight
Cargo transport

Occupants :
Captain
First officer Flight
engineer
Cargo supervisor
Maintenance man

Summary :

The cargo aircraft was performing a flight Luxemburg-Kano (Nigeria). As

the aircraft, climbing towards flight level 330, was flying over the Drôme
area, both right engines separated from the wing. The crew succeeded in
controlling the aircraft and landed at the Istres air base, landing gear
and flaps down, with the right wing on fire. The aircraft rolled off the
runway to the left. The 5 crewmembers left the aircraft. The firemen
extinguished the fire.

Consequences :
╔═══════════════════════╤══════════╤═══════╤════════════_
║ Persons │Aircraft │ Cargo │ Third Party║
║ │ │ │ Damage ║
╟──────┬────────┬───────┼──────────┼───────┼────────────╢
║killed│ injured│ unhurt│ │ │ ║
╟──────┼────────┼───────┤ │ │ ║
║ / │ / │ 5 │Destroyed │ Intact│ None ║
║ │ │ │ │ │ ║
╚══════╧════════╧═══════╧══════════╧═══════╧════════════╝

(°) The times given in this report are in Universal Time Coordinated

2
(UTC). Add one hour to obtain the time for France.

1. Factual information

1.1 History of the flight

The aircraft, under an IFR (°) flight plan, was flying from
Luxemburg to Kano (Nigeria), carrying freight. It took off from Luxemburg
aerodrome at 07.14 hrs with the peak load of 150 tonnes (38 tonnes of
freight, 116 000 pounds of fuel).

The crew was composed of three men, the captain, the first
officer, and the flight engineer. Two passengers were on board; a
maintenance man, and a cargo supervisor.

The aircraft, on a heading of 199°, when passing "VILAR" and

the VOR of Martigues, over the Drôme province, was authorized by le Centre
Régional de Navigation Aérienne sud-est : CRNA/SE (South-East Aircraft
Navigation Regional Center), to leave flight level 290 and climb to flight
level 330. This flight section was performed in IMC, in turbulent air.
With the throttles at climb power and automatic pilot engaged, the air-
craft was flying at an indicated air speed (IAS) of 280 kt passing the
flight level 320. It flew over the far south-east of the Drôme area, 20
NM to the west of Sisteron.

At this moment, the crew was experiencing severe turbulence

and heard a "double bang". The aircraft suddenly rolled to the right.
The captain disengaged the automatic pilot and struggled to keep control
by "countering" with the control stick and the rudder pedals.

The continuous fire warning system sounded. According to the

visual warning, this corresponded to a fire on engine n°4. A short time
later, a visual warning lit up to report a fire on engine n°3. The crew
noted that the throttles of these engines had moved forwards on their
own.

The cockpit noise level was extremely significant dominated

by the engine fire warning that the flight engineer could not switch off
despite the fact he repeatedly pressed the cap on the panel. Another warning
system sounded at the same time to indicate the cabin depressurization
and continued for most of the flight and until the landing (intermittent
warning horn).

The cockpit voice recorder (CVR), as well as the

crewmewbers'additional information enabled identification of the
essential actions respectively executed in this emergency situation by
the captain, the first officer, the flight engineer and both passengers.
It should be noted that these actions ended in the successful landing
at a diversion field.

(°) : Instrument Flight Rules

2
 The captain was worried about the origin of the "fire" warning.
The first officer annouced that engine n°4 (right outboard) "had separated
from the wing" and immediately sent out the distress call "MAYDAY MAYDAY".
A short time later, he specified that, in fact, both right engines "had
gone".

The flight engineer suggested lightening the aircraft by fuel

dumping. The captain immediately agreed.

While the first officer was in charge of radio communications

and determining the nature of the aircraft's damage, the captain, who
was struggling at the flight controls, asked for the meteorological
conditions in Marseilles and ordered the gear extension. Then, a descent
towards Marseilles was initiated.

The flight engineer, helped by the maintenance man, extended

the gear according to the emergency drill and continued with fuel dumping.

The first officer checked that the emergency drill recommended

in case of engine separation was in progress and, still being in charge
of the ATC communications, attempted to obtain the meteorological
conditions in Marseilles. At the captain's request, the first officer
specified to air traffic control that they were capable of only limited
manoeuvring.

The first officer noticed "an airfield ahead", and asked for
its identification. This airfield proved to be the Istres military field.
Then, he asked about the length of the runway (4000 meters) and quickly
got from Marseilles air traffic control the landing clearance.

He asked for a left hand circuit so as to land on runway 15

(downwind runway 33). The Istres controller immediately agreed.

By listening to the cockpit voice recorder, it was apparent

how difficult it was for the captain to complete the last turn before
alignment. The first officer encouraged him by repeating six times "left
turn".

During this last turn, the controller informed the crew that
the aircraft was on fire.

The landing took place slightly to the left of the centreline,

the aircraft touching down on the runway at 190 kt. The first officer
and the flight engineer helped the captain during this phase. The first
officer held the left engines throttles. The captain specified that there
were "no hydraulic brakes!", and thus resorted to the "emergency brake
system". The left main gear tyres burst. The flight engineer selected
maximum reverse power on engine n°2.

The aircraft, after a 2,300-meter-ground roll, went out off

the left side of the runway and stopped 250 meters further on, heading

3
approximately 90° from the runway axis. The firemen estinguished the fire
with their high-capacity fire vehicles (fire brigade: SSIS).

The crewmembers evacuated the aircraft through the cockpit side

window panels with the help of escape ropes. Both passengers went out
through the left front door.

The crewmembers only realized that the right wing was on fire
when the aircraft landed and stopped. In particular, it appeared that
the first officer had not heard the remark of the controller.

The landing took place at 08.35 hrs, that is to say approximately

24 minutes after the loss of the two right engines.

1.2 Injuries of persons

None

1.3 Damage to aircraft

The aircraft was severely damaged by the right wing fire (the
wing was practically destroyed, the right rear part of the fuselage was
damaged by the fire, see the photos in the Annex).

1.4 Other damage

The cargo (spares parts for the oil industry, and medecines)
was undamaged.

The engines, lost during the flight, fell over a rocky desert
area, without causing particular damage.

1.5 Personnel information

1.5.1 Captain

. Male, aged fifty-seven, Swedish.

Aircraft ratings :

. Airline transport pilot licence n°D3506218555

issued on 22 December, 1966 by the Swedish
Aviation Authority.
. Equivalent licence issued on 19 March, 1992 by
the Ministry of Nigerian Civil Aviation,
valid to September 1992.

Qualifications :

. B707 qualification issued in March 1982, as

4
 a Captain.
. Other qualifications : DC6 , L188, S210,
B737.

Experience :

. Approximately 26 000 flight hours.

. On B707 : 7 100 hours.
. During the last 30 days : 33.10 hrs + 4 hours on
a B707 simulator .

Engaged by the Trans-Air Service Ltd on 16 March, 1992.

1.5.2 First officer

Male, aged forty-four years, British.

Aircraft ratings :

. Airline transport pilot licence issued by the

United Kingdom Civil Aviation Authority, C.A.A.
. Equivalent licence n° CP 2134839/A issued on
19 March, 1992, by the Ministry of Nigerian
Civil Aviation, valid to 18 September, 1992.

Qualification :

. B707 qualification issued in March 1982 , as

a First officer.

Experience :

. 14 000 flight hours.

. On B707 : 4 500 hours.
. Since 1 January, 1992 : 60 hours.

Engaged by the Trans-Air Service Ltd on 16 March, 1992.

1.5.3 Flight engineer

Male, aged fifty-five, British.

Aircraft ratings :

. Flight engineer licence n° 1090 issued on 7 Oc

tober, 1988, by the United Kingdom Civil Avia-
tion Authority, valid to 18 September, 1992.
. Equivalent licence issued by the Ministry of
Nigerian Civil Aviation, valid to 31 March, 1993.

Qualification :

. B707 qualification.

Experience :

5
 . 18 000 flight hours, all on B707.
. During the last 30 days : 56 hours.

Engaged by the Trans-Air Service Ltd on 16 March, 1992.

1.5.4 Maintenance man

Male, aged thirty-six, Nigerian.

Not a flight engineer, on board as a supplementary

crewmember. Good experience of B707.

Engaged by the Trans-Air Service Ltd on 2 March, 1992.

1.5.5 Cargo supervisor

Male, aged twenty-seven, Icelandic.

Engaged by the Trans-Air Service Ltd on 16 March, 1992.

1.6 Aircraft information

1.6.1 Airframe

Manufacturer : Boeing Aircraft Corporation.

Type : B707-321CH.
Serial number : 18718.
Delivered new on April 1964.
Registration : 5N-MAS.
Nigerian Registration Certificate n°772.
Airworthiness Certificate n°772, issued on 3
February, 1992, valid to 2 February 1993.
Flight hours : 60 895.
Number of operating cycles : 17 907.

1.6.2 Engines

Manufacturer : Pratt and Whitney.

Type : JT3D-3B.
Engines hours : see the following table.

╔═════════╤═════════╤═════════╤═════════_
║ N°1 │ N°2 │ N°3 │ N°4 ║
╔═══════════════_─────────┼─────────┼─────────┼─────────╢
║ Serial Number │ 644426 │ 645572 │ 645468 │ 643387 ║
╟───────────────┼─────────┼─────────┼─────────┼─────────╢
║ Flight hours │ 39281 │ 59728 │ 35702 │ 43629 ║
╚═══════════════╧═════════╧═════════╧═════════╧═════════╝

1.6.3 B707 concise technical description

6
 1.6.3.1 Flight controls

Roll and pitch

The elevators are fitted with servo-tabs as well as balance

panels. The tabs are manually operated by the flight controls : the
aerodynamic loading generated by the tab produces elevator deflection.

Rudder

The rudder is operated by an hydraulic servo- control. In case

of hydraulic power system loss, the tab enables rudder deflection with
a reduced movement.

1.6.3.2 Hydraulic system

The hydraulic power system includes three different systems

:

Utility system

Utility system pressure is generated by two-engine-driven

pumps, one on engine n°2 and one on engine n°3. This system is used for
:
. landing gear operation
. nose wheel steering
. brakes normal circuit
. flaps
. outboard spoilers

With an inboard engine inoperative, it is possible to isolate

the corresponding hydraulic pump with an hydraulic supply shutoff valve
using a switch on the flight engineer's sidewall panel (it should be noted
that the hydraulic pump is also shut off by operating the "engine fire
shutoff handle").

Two auxiliary circuits, pressurised by electrically operated

pumps : the auxiliary system

These auxiliary circuits, interdependent (the circuit n°1 can

supply pressure for the circuit n°2), are used for :
. rudder
. inboard spoilers

1.6.3.3 Braking

Normal hydraulic braking is operated by the utility hydraulic

system, which includes in particular a specific brake accumulator. It
can also be operated by the auxiliary hydraulic system.

Pneumatic brake system is operated, in case of total hydraulic

system loss, by a pneumatic system (with an air bottle) controlled from

7
the left pilot's panel. In case of emergency, there is no differential
braking or antiskid protection available.

1.6.3.4 Rapid fuel dump system

Two retracting fuel dump chutes (one per wing) are operated
from the flight engineer's sidewall panel.

1.6.3.5 Pressurization

The aircraft is pressurized with the air generated by

turbocompressors driven by the engines n°2,3 and 4 gearboxes, or by bleed
air from the engines. The air then goes through the air conditioning units.

An intermittent horn sounds if the cabin altitude is up to 10

000 feet +/- 250. A press-button on the flight engineer's upper panel
enables cancellation of this warning.

1.6.3.6 Electrical power

Very briefly, the electrical power is composed of:

Primary electrical power

Triphase AC power supplied by four 115/200 V 400 Hz

engine-driven generators.

Secondary electrical power

. 28 V DC power supplied by four transformer

rectifier units and an emergency battery
. 28 V/400 AC power supplied by a series of
autotransformers.

Power is distributed by bus bars. In order to maintain the

availability of the priority circuit, called "essential bus", it is
possible to select any generator from the flight engineer's sidewall panel
to supply the "essential bus".

In normal operation, the "essential bus" is connected to engine

n°3 generator. The order of connection priority defined in the emergency
drill is 3, 4, 1, 2.

1.6.4 Maintenance

This brief chronological account describes the aircraft

maintenance during the last ten years :

. From January 1982 to July 1985, the aircraft

(registered G-BFZF) was stored at Lasham (En-
gland) under the control of Dan-Air Services,
Engineering Division, Lasham Airfield, Nr Alton
Hampshire.

. From September 1985 to February 1986, the Dan

Air Company carried out maintenance work so
as to overhaul the aircraft : "A" ,"B",
"C,2C,3C,4C,5C,D" checks, according to the Boeing

8
 maintenance program.

These checks happened when the aircraft had accumulated 52 558

flight hours and 15 877 landings. The Civil Aviation Authority issued
the Certificate of Airworthiness on 19 February, 1986. The new registration
was
G-BNGH.

. An important modification was made on 2 May,

1986 : "B707 quiet nacelle installation" ( engine
cowling modification achieving better sound
isolation).

This modification is described as follows in the English

documents : "modified engine cowling in accordance with Shannon
Engineering Master Drawing List (MDL) number JS1102001, revision B dated
22 February, 1985 or later FAA approved Revision, STC SA 2699NM and
FAA-PMA". In the text, this modification is called "HUSH KIT".

. From June 1989, the maintenance was entrusted to

the British Company Modern Jet Support Centre
(MJSC) at Manston, Kent, England.

. The main works operated are :

- June 1989, A,B,EQ8 checks
- October 1989, A,B,EQ9 checks
- January 1990, B check
- May 1990, C check
- June 1990, left wing minor repair
- October 1990, engines n°1,3 and 4 change at the owner's
request
- January 1991, B check
- October 1991, C check
- January 1992, B check
- February 1992, A and B checks.

. The aircraft performed no flights from 24 October, 1990

to 3 March, 1992. It received a Certificat of release to service on 4
March, 1992, with the registration "5N-MAS", from Southend (see
paragraph 1.16.4 further on).

Following a burst of the left main landing gear tyres following

a ferry flight between Manston and Southend, on 6 March, 1992, induced
by an "antiskid" braking system failure, the repair as well as an "A"
check were carried out by Heavylift Ltd at Southend, on 8 and 9 March,
1992.

1.6.5 Successive owners and operators

The following table, established from the UK register, gives

a list of the aircraft successive owners, from 1978 to 1992, and specifies
the registration changes.

╔══════════════╤════════════════════╤═══════════════════════_
║ Registration │ Period │ Owner ║
╟──────────────┼────────────────────┼───────────────────────╢
║ │from 09/78 to 01/83 │SCIMITAR AIRLINES Ltd ║
║ G-BFZF ├────────────────────┼───────────────────────╢
║ │from 01/83 to 01/86 │ GREYHOUND EQUIPMENT ║

9
╟──────────────┼────────────────────┤ FINANCING Ltd ║
║ │from 02/86 to 05/86 │ ║
║ ├────────────────────┼───────────────────────╢
║ G-BNGH │from 05/86 to 12/91 │ TRADEWINDS AIRWAYS Ltd║
║ ├────────────────────┼───────────────────────╢
║ │from 12/91 to 02/92 │ ║
╟──────────────┼────────────────────┤ TRANS-AIR SERVICE Ltd ║
║ 5N-MAS │from 02/92 to 03/92 │ ║
╚══════════════╧════════════════════╧═══════════════════════╜

1.7 Meteorological information

1.7.1 General situation

1.7.1.1 Surface information

At 06 00 hrs, a deep depression centred above England, covered

Western Europe. An active cold front was lying from the north of France
to the Rhone Corridor and extended towards the Balearic Islands and the
Straits of Gibraltar. Over Provence and the Alps, the unsettled weather
was characterized by thunder-showers. Hail falls were observed here and
there.

1.7.1.2 Altitude information

At 06 00 hrs, at Fl 300 (300 hPa), there was a low pressure

area in the west of Britain with a trough extended towards the Iberian
Peninsula.

In the front of this trough, a 90 kt south southwest jet stream

was lying from the south of Spain to the Balearic Islands,the Rhone Valley
and Luxemburg.

1.7.2 Meteorological conditions along the

route

The reconstruction of the meteorological situation was achieved

for the period between 08 00hrs and 08.35 hrs, corresponding to the
overflight period from the central Rhone Valley to Istres where the landing
took place.

The cloud layer was composed of 8/8 Ac-As topped by Cs and Ci

up to Fl 250-300. Locally, Cb, embedded in the mass, developed up to Fl
330.

Rain and hail clouds, as well as snow clouds in altitude were

mentioned over the Drôme area and the upper Provence.

The visibility was nil in the cloud layer. Below the cloud base,
around 800 and 1 200 m, the visibility was greater than ten kilometres
but was locally reduced to 1 500-3 000 m because of the showers.

Winds and temperatures at altitude were :

. Fl 320 : 200-210°/70 kt, -57°C,
. Fl 300 : 200-210°/70 to 80 kt, -53°C,
. Fl 180 : 180 to 200°/ 50 to 65 kt, -25°C.

10
 The turbulence, moderate to severe, decreased in the atmosphere
lower layers.

1.7.3 Particular conditions during the two

right engines departing

The studies of the RADAR images and the METEOSAT satellite showed
two Cb over the Séderon - Mont Ventoux -Nyons area, triangle over which
5N-MAS was flying. These Cb were embeddedin the cloudy mass and rose up
to above Fl 330.

In addition, the flight was performed on the right edge of the

90 kt jet stream.

These two simultaneous particular meteorological conditions

generated severe turbulence, confirmed by the crewmembers.

1.7.4 Approach and landing conditions

The conditions were better from Durance and Arles. At Fl 90,

the flight was conducted beneath the Ac-As layer. The Istres approach
was flown at 3 000 feet at the level of a little developed Sc and Cu peak.
The breakdown was achieved at QNH 988 hPa.

The following conditions were transmitted to the crewmembers

by the Marseilles Approach controller :

. At 08.22 hrs : Visibility : 6 km

Cloud : 2/8 Sc 500 feet, 3/8 Cb
1 600 feet and 3/8 Cu 2 000 feet.

. At 08.23 hrs : Visibility : 6 km

Significant weather : rain,
Cloud : 2/8 Sc 500 feet, 3/8 Cb
1 600 feet and 3/8 Cu 2 000 feet.

The final approach over Istres, towards the threshold of runway

15, was performed VFR, with a 320°/10 kt surface wind as the controller
had advised, then a 320°/8 kt one.

The Istres meteorological information as far as the final

approach and landing are concerned was :

. At 08.30 hrs : Wind : 310°/8 kt, gusts at 10 kt,

Visibility : 12 km
Significant weather : none (the
rain had stopped at 08.06 hrs)
Cloud : 3/8 Sc 2 600 feet, 4/8 Sc
3 300 feet, 7/8 Ac 9 000 feet,
Temperature : 9,7°C, dew point :
7,6°C,

QNH : 988 hPa

QFE : 985 hPa.

. At 08.36 hrs : Wind : 320°/8 kt, gusts at 12 kt,

Visibility : 15 km

11
 Cloud : 2/8 Sc 2 600 feet, 5/8 Sc
3 000 feet, 7/8 Ac 9 000 feet,
Temperature : 9,2°C, dew point :
6,1°C
QNH : 988 hPa
QFE : 985 hPa

1.8 Aids to navigation

No fault in the radio aids and in the ATC communications used

had been reported by the crewmembers or was known by the control
organisations' involved.

1.9 Telecommunications

The aircraft was successively in communication with

the control organisations mentioned below. The table specifies the periods
of time during which transcriptions were made.

These transcriptions are given in Annex 1.

╔═════════════╤════════════╤═══════════════════════════════_
║ Control │ Frequency │ Transcriptions ║
╟─────────────┼────────────┼───────────────────────────────╢
║ CRNA/SE │ 126,7 MHz │ 08.09 hrs to 08.16 hrs ║
║ │ │ ║
║ Marseilles │ │ ║
║ Control │ 123,9 │ 08.16 hrs to 08.19.30 hrs ║
║ │ │ ║
║ Marseilles │ │ ║
║ Approach │ 120,2 │ 08.19.30 hrs to 08.31 hrs ║
║ │ │ ║
║ Istres │ │ ║
║ Tower │ 123,6 │ 08.31 hrs to 08.37 hrs ║
╚═════════════╧════════════╧═══════════════════════════════╝

The main part of the dialogue relating to this service is

summed up below.

CRNA/SE

- 08.11.41 hrs - B 707, flight designator KABO671, sends

the first of six MAYDAY messages,

- 08.12.24 hrs - the controller : "I have not no longer radar

contact Sir, I am sorry squawk again 7172",

- 08.13.08 hrs - the controller : "...squawk 7700 Sir 7700",

- 08.14.13 hrs - KABO671 : "we need emergency landing both engines

missing right wing emergency radar",

Marseilles control

- 08.16.09 hrs - KABO671 sends the message MAYDAY,

12
-............ - KABO671 advises the nature of its problems:
"two engines missing two engines missing structural
request straight in landing",

- 08.16.40 hrs - KABO671 : "priority MAYDAY",

- 08.17.52 hrs - KABO671 :"give me the weather for Marseilles...",

- 08.18.40 hrs - KABO671 : "MAYDAY MAYDAY MAYDAY MAYDAY MAYDAY

MAYDAY", 672 request weather",

Marseilles Approach

- 08.19.25 hrs - KABO671 : "Marseilles MAYDAY, Marseilles MAYDAY ..."

-............ - "Marignane UT7209 ...."

- 08.20.46 hrs - the controller : "UT7209 climb radar level 140",

-............ - UTA7209 : "radar level 140 7209",

- 08.21.14 hrs - KABO671 : "This is MAYDAY MAYDAY KABO ..."

- 08.21.50 hrs - KABO671 : "what's the weather, please weather

for KABO KABO request weather",

- 08.22.35 hrs - KABO671 : "what's the weather PALMA, weather PALMA ?",

- from 08.23.21 hrs to 08.26.38 hrs, the controller gives the wea- ther
conditions in Marseilles,

- 08.30.53 hrs - hand-off to Istres.

Istres tower

- 08.30.56 hrs - the controller : "671, this is Istres Istres I

receive you fives Istres 33 runway 33 in use
QFE 985 you are cleared for a down wind arrival Sir",

- 08.31.34 hrs - KABO671 : "we make a left hand pattern",

- 08.33.28 hrs - the controller : "671 you have fire on board (twice)
I confirm fire on board",

- 08.34.40 hrs - the controller : "you are runway 15 the wind 320 you
are on axis you are on axis".

1.10 Istres aerodrome information

The Istres military air base is open to Civil Air Traffic

provided that a previous authorization has been given. The airbase has
a 3 685 meter long paved runway, QFU 15, oriented 155°-335°, with a stopway
that permits in total a 4 000 meter long landing distance available,
with a width of 60 meters.

The aerodrome control service was in operation and its

equipment was in good working order.

13
 The aircraft was in communication with the tower controller,
whose office is in the visual tower cab. This controller has no radar
screen, the radar display being operated in the approach room.

This airbase has a major fire brigade (SSIS) since there are
Air Force Boeing C 135 aerial tankers, as well as the Flying Test Centre
based there.

1.11 Flight recorders

In compliance with the regulation in force, the aircraft was

equipped with two crash protected recorders :

. a Plessey PV 1584A Digital Flight Data Recorder,

SN CH 2333,
. a Fairchild A100A Cockpit Voice Recorder, SN
25027.

These two recorders were intact.

1.11.1 DFDR read-out

The reading of this recorder should reveal the following

parameters : the heading, the roll rate, the IAS, the radio
altitude, as well as the EPR indication for each engine.

It was not possible to reveal the IAS, since the corresponding

sensor was out of order. Annex 3 gives the history of the other parameters.

It appears that :

. before the recorded time "3 200 seconds", the

operation of the engines was normal. In particu- lar,
the four power curves of the engines practi
cally merged,
. around the recorded time "3 140 seconds", the
engine power increased from EPR 1,7 to EPR 2,0
on the four engines,
. at the recorded time "3 200 seconds", EPR3 and
EPR4 sensors gave no indication, while, a short
time later, the power of the engines n°1 and 2
increased.

Consequently, from the recorded time "3 200", the engines n°3
and n°4 had separated from the right wing.

1.11.2 CVR read-out

The CVR was read with the assistance of a pilot and a flight
engineer both belonging to the Flight Control Organisation, qualified
on B 707, so as to identify accurately, in particular, the different warning
horns.

The transcript of the conversations is in Annex 4.

The main communication between the three crewmembers, the first

14
officer and the control organisation after the loss of the two engines
until the landing is reported further on.
In the following text, the abbreviations used are .
captain : Captn
. first officer : F/O
. flight engineer : Fle

Captn : He asks"...? Fire ?"

We can hear sounds indicating physical efforts.

F/O : He reports that "number 4 engine has left the wing!". He sends
out the distress call MAYDAY MAYDAY, twice and asks for radar assistance
so as to execute an emergency landing.

F/O : He reports that they "have lost both engines".

He checks that the captain is heading south as the CRNA/SE has
ordered.

Captn : "I'm trying!"

Fle : He suggests to lighten the aircraft by dumping fuel through the

dump valves, which the captain immediately accepts.

F/O : He advises air traffic control of the nature of the aircraft's damage
and repeats his request for an emergency landing.
He questions the flight engineer about the execution of the
"emergency operating procedure" check-list : "You have cut engines ?"
He transfers to Marseilles.

Captn : The sounds indicating the physical effort to handle the controls
gets more and more intense.
During the descent, he asks for the weather in Marseilles.

F/O : He advises the nature of the aircraft's damage and, once more, sends
out the distress call MAYDAY.
He asks for the weather in Marseilles.
He orders the gear extension.

Fle : He extends the gear using the emergency method, helped by the
maintenance man (the corresponding efforts can be heard by listening to
the CVR as well as the aural signal).

F/O : Bothered by the radio traffic between Marseilles and two aircraft
(Mike Victor, Delta November Whisky), once more he sends out distress
calls MAYDAY and asks for the weather.

Captn : He querries about the defuelling through the dump valves.

He specifies that he has "limited manoeuvring".

F/O : He advises Marseilles Approach that he has limited manoeuvring,

and asks for the weather in Marseilles.

Captn : He considers a landing towards Palma.

F/O : He finally gets the meteorological conditions in Marseilles and

answers the questions of Marseilles Approach (How many passengers do you
have on board ? How much fuel do you have on board ?)

15
 He mentions he can see the airfield.
He say again "this is emergency landing... emergency... full
emergency".
He mentions he can see an airfield ahead and asks what that airfield
is. Marseilles Approach advises him it is a military airfield.

Captn : He considers that the runway is too short.

F/O : He asks for the length of the runway and gets the answer (4 000
meters).

Captn : "OK"

F/O : He advises Marseilles Approach he can see this airfield and that
he can make a left pattern for landing. He gets from Marseilles Approach
the frequency of this military airfield (Istres).
He contacts the airfield. Istres tower acknowledges receipt, and
advises that the runway in use is runway 33 then gives the QFE.
He asks for a left hand pattern.
The controller agrees.

Captn : "OK ! What is the wind ?"

F/O : "What is the wind ?" He gets the information (wind 330°/10 kt, gusting
14 kt).
He asks the question : "Do you have some radar ?" The controller
answers there is no radar and that he has no visual contact with the
aircraft.
He advises the captain to keep a minimum speed of 200 kt.
The controller repeats he has no visual on the aircraft, and asks
the question : "Do you see my runway ?"

Captn : "Negative, no !"

F/O : (at the same moment) "We are just coming out of the top of the clouds.
We come to the west of the field. We see your runway. We are turning on
to the west, 3 000 feet". He addresses the captain : "We turn
left to land".

By listening to the CVR, It is apparent how difficult it was for

the captain to achieve this last turn before alignment. The first officer
encourages him by repeating six times "left turn".

The Istres controller asks for confirmation of the gear extension.

F/O : "Yes".

At this moment, the controller specifies that the aircraft is on

fire : "...You have fire on board, (twice), I confirm fire on board" (it
is 08.33.28 hrs).

F/O : "...?..."

Then, listening to the CVR becomes more and more difficult. Different
aural warnings keep on sounding. In addition, power cuts - and thus
recording cuts - occur and lead to the loss of several sentences in the
dialogue.

16
 Afterwards, the CVR again becomes audible. The controller specifies
: "Good descent, you are on axis".

F/O : "Roger".

Fle : He annouces the landing touchdown (at 8.35.35 hrs) then "reverse".

Captn : "No".

End of the recording.

1.12 Wreckage information

1.12.1 The aircraft

Most of the damage
affects the right wing. The photo n°1 indicates the extent.

The pylons (as well as the engines) n°3 and 4 have disappeared
without having caused significant damage to the leading edge.

The electric cable loom routing in the leading edge on the pylon
of the right inboard engine was torn, opened, and some electric cables
were bruised and burnt.
They showed marks of short-circuits (see photo n°2). By comparison, the
cable loom routing in the leading edge of the right outboard engine, on
the pylon, did not show marks of short-circuits or of fracture.

The most noticeable damage was that caused by the fire. The
wing skin panels above the pylon of the right inboard engine were distorted,
warped and cracked. The blackish marks went from the leading edge, above
the pylon, and then widened towards the trailing edge (see photos n°3
and 4).

The other main damage was all located on the fuel tank n°4 (tank
located between the ribs WS320 and WS733 and the front and aft spars)
(see drawing of the wing included in the photo annex, diagram 0).

The skin panels of the upper wing over this tank had straight
cracks reaching up to 2,5 meters in lenght (see photo n°4 and diagram
0 in the Annex), the width of the molten and burnt lips being of 12 to
25 millimeters.

These crevices were located exactly above the tank vent pipes
(see photos n°5 and 6).

The trailing edge was totally burnt in the area between both
engines (see photo n°6).

The inboard and outboard flaps had completely disappear,

revealing the burnt operating mechanisms (see photo n°7). The inboard
aileron was severely damaged.

Moreover, the examination of the inboard wing box identified

the marks of an inner explosion on fuel tank n°4. This explosion seemed
to be at the origin of significant deteriorations affecting the wing
stiffeness.

17
 This explosion had caused the displacement of the inner ribs
of this tank (see photo n°8). The wing stiffeness was particularly damaged
on the front and aft spars :

- the lower chord of the front spar was cracked at

the pylon of engine n°4, the width of the
crack reaching up to 50 millimeters.

- the upper chord of the aft spar was cracked over a

length of at least 1,25 meters.

Thus, it appeared that the right wing was severely damaged first
because of a fire and then because of an inner explosion at fuel tank
n°4.

The fire was intense enough so that the flames licked the rear
part of the fuselage, at the back of the trailing edge of the wing. The
corresponding cabin windows were cracked, burnt and some were burst. The
paint of the rear part of the fuselage, scorched by the heat, showed the
structure. The skin of the fuselage was corrugated, at the back of the
right rear emergency exit (see photos n°9 and 10), which was the sign
of significant distortions.

1.12.2 Engines n°3 and 4

1.12.2.1 Engine n°3

Engine n°3 (right inboard) was found in the area of Séderon,

in an uninhabited area, on loose ground, near a forest path. The engine
was laid on the left side, in a place close to the aircraft's ground track.

This engine was severely damaged by the ground impact. It was

completely flattened, according to a nearly diametrical plan, extending
from the points 5 o'clock to 11 o'clock (according to the conventional
marking "pilot place").

The photo n°11 depicts its condition.

The four pylon fittings on the wing (described in the paragraph

1.16.1.1 below) were broken. On one of those fittings, the inboard midspar
fitting, we could clearly see a fracture with a strange appearance, a
fracture which had the shape of a quarter of an ellipse located at the
bore for the attachment pin on the wing (see photo n°12).

1.12.2.2 Engine n°4

From the discovery of engine n°3, investigations undertaken

with helicopter, following the heading steered by the aircraft, enabled
detection engine n°4, 800 meters from the first mentioned, also in an

18
uninhabited area, on the side of a mountain, in an area hard to get to
and very rocky. The engine also lay on the left side. Because of the
nature of the ground, it had bounced after first impact.

The first impact on the ground and the bounce severely damaged
this engine as photos n°13 and 14 show. The rear part separated from the
front part, the fracture being between discs n°2 and 3 of the high pressure
turbine.

The cowls had released from the engine. The air intake had rolled
on the sloping ground. Its general circular shape was conserved except
in the left segment from 6 to 10 o'clock. In this area, the front periphery
of the air intake showed two impacts very marked at 8 and 10 o'clock.
On the first, we could observe mat white marks, parallel to brighter grazing
on the soft material (see photos n°15 and 16).

The rear part of the pylon, around the attachment points to

the wing, was very damaged. Examination of the fractures of the fittings
is mentioned in the paragraph 1.16.1.2 below.

Afterwards, back along the track of the aircraft, a

2,15-meter-long element was collected and identified : this was the rear
part of the cowling of the pylon of engine n°4. This light element might
have been blown away by the wind coming from the south (wind from 200°
to 210° according to the meteorological report).

The track of the aircraft appears on the map in Annex 2, derived

from the radar information (Air Force radar System). The impact area of
both engines also appears on this map.

1.13 Medical and pathological information

Not relevant

1.14 Fire

Alerted 15 minutes before the landing, fire brigade (SSIS) of

the Istres air base intervened to bring the fire under control as soon
as the aircraft stopped.

The rescue vehicles were parked on a side strip on the edge

of the runway and followed the aircraft during the landing phase, that
is to say over approximately 2 500 me-ters. The firemen witnessed the
fire on the right wing, the burst tyres of the left main landing gear
(under the ef-fect of the emergency braking) and had to get round some
debris dropped on the runway.

The fire fighting was operated in three phases :

- fighting the right wing fire, widespread fire,

and a second focus on the right main landing

19
 gear, fire limited but constantly fed by kerosene
leaking from the fuel tank of the right wing,
overhanging the landing gear. This phase lasted
only 3 minutes, from 08.36 hrs to 08.39 hrs.

- cooling of the main right landing gear, for

1.30 hour.

- surveillance of the wreckage, for 7.30 hours,

because of major kerosene leaks coming from
the right wing.

The following fire vehicles were used :

- 2 Multivalent Rapid Intervention Fire Vehicles ,

- 2 Multivalent Extinguither Heavy Vehicles,

- 2 Aerodrome Foam Vehicles, which had used

extinguishing powder (1 000 Kg), foam (2 000 l of
emulsifier liquid), and water (40 000 l).

1.15 Survival aspects

Not relevant

1.16 Tests and research

1.16.1 Expert evaluations of the engine

pylons units n°3 and 4

The wreckage of the two engine/pylon units of the right wing

were carried to the Parisian area, to "Centre d'Essais des Propulseurs"
(C.E.Pr - Propulsion Test Centre) in order to be examined.

1.16.1.1 Expert evaluation of the

engine pylon
fittings n°3

Diagram 1 illustrates (perspective view and sectional view)

the method of attachment of the pylon on the wing, which is composed of
4 fittings :

- one upper front spar fitting,

- two midspar fittings (inboard and outboard)
- one lower spar fitting

The fracture of the inboard midspar fitting (identified as "c"

on diagram 2) reveals an evident crack, which has a shape of a quarter
of an ellipse, roughly centered on the angle of the upper entrance of

20
the bore (see photos n°12).

The drawing on the diagram 2 shows that this fitting broke

exactly across the bore. The attach fitting, which had the shape of an
arc of a circle, was found on the wing. It was this part of the fitting
which has been examined. The photos n°17, 18 and 19, with increasing
magnifying power, give a clear view of the fracture.

The upper part of the attachment fitting broke in a straight

line, from a fatigue "half-moon" which started in the inboard side of
the bore (on the side of the fuselage). Numerous corrosion pits have
facilitated the start of the crack. Beyond the "half-moon", the crack
was bordered by a thin shear lip at 45°.

The "half-moon" measured 11mm by 8mm, which represented

approximately 18% of the broken surface. It represented two different
areas (see photo n°18) :

- the major part of its surface was black, this was

probably the result of a significant oxide
deposit,

- at the end of the cracking, the area was edged by

a clearer band which was approximately one
millimetre across.

The crack propagated from 4 areas of initiation, all located

in the bore. The bore exhibited a quite important zone of corrosion on
the inboard side, especially on two of these areas. The binocular magnifier
examination showed that the two other areas were also located on the
corrosion pits.

After cleaning of the crack, some crack arrest lines appeared

quite clearly (see photo n°19). These lines corresponded very likely to
engine cycles.

The electron beam microscope examination had exposed :

- numerous secondary cracks origins in principal

crack origin region,

- a very important oxide deposit on all the darker

part of the crack, feature which prevented a more
accurate observation of this area,

- some quite clear crack arrest lines at the end of

the crack, in the clear part. There were more or
less 25 to 30 lines which were likely to corres- pond to
the same number of flights. Between those
lines, a finer striation could be observed, but
overall the crack of rest lines were the indica tion of

21
the important strain experienced by the crack.

Beyond the "half-moon", the crack presented the characteristics

of a brittle static fracture on a high tensile steel (steel 4330).

The lower part of the attach fitting broke statically under

the effect of a bending downward stress (see scheme on the diagram2).

The fractures of the other midspar attachment fitting (outboard)

were located forward of the bore of the lug, on the beginning of the two
horizontal tangs which framed the structure of the pylon (ruptures
identified as "b" on the diagram 2). Those ruptures were static, their
similar appearances revealed that they started from the inboard and
propagated in parallel towards the outboard.

The fracture of the upper front mounting point broke out on

the beam connected to the upper surface of the wing. This beam, with an
I-section, broke at two points from the inboard edge of the upper chord
towards the outboard edges. This was a fracture in static bending as the
forward end of the beam was submited to a movement towards the outboard
(ruptures identified as "a", diagram 2).

The lower fitting of the diagonal brace broke up on the bores

of the two yokes of the brace . The examination of the fractures (with
a static nature) also showed that their orientation was from inboard to
outboard (see scheme with the following legend : "lug of the diagonal
brace, diagram 2").

The schemes on the diagram 2 present the different aspects of

the fractures of the fittings. These observations permitted assessment
that engine n°3, free on its left-side, had drifted to its right side.
The remaining fittings broke and the released engine carried on its
movement to the right.

1.16.1.2 Examination of the engine

pylon fittings n°4

Photos n°20 and 21 reveal the important distortions of these

fittings, the forward part of the pylon around them was highly damaged.

None of the fittings presented evidence of fatigue fracture.

These fractures occured on the two midspar fittings axes and the beam
of the overwing fitting. The fracture of the lower spar fitting is located
on the diagonal brace forward end, at the root of the clevis lugs. The
cause of the fracture was a torsion at stress. The upwards movement of
the lug on the inboard side corresponded to a rotation of the engine around
the axis of the brace.

The upper front spar fitting lug was little damaged. The fracture
of this fastener was the result of the shear of the pin at the interfaces
of the double lugs of the attach fitting and the overwing beam, as well
as the partial tearing of the lugs. These fractures were due to a forward
tensile static stress of the engine.

1.16.1.3 Expert evaluation of the

engine air intake n°4

22
 Photos n°15 and 16 of the outboard engine air intake give
prominence to :

- the existence of an impact on the inboard side, a

little bit below the engine axis of rotation,

- matt whiteish marks, parallel to brighter grazing

on the soft material .

The analysis of these whiteish marks had revealed that it was

a paint deposit.

In order to compare, some pieces of white paint from the cowl

of engine n°3 were analysed.

These two analysis revealed that the deposit collected on engine

air intake n°4 and the pieces of paint of the engine n°3 were of the same
nature (polyurethane-based paint).

1.16.1.4 Examination of the engines

The wreckages of engines n°3 and 4 were inspected by a Pratt

and Whitney company expert who also had a copy of the engines powers
recording from the DFDR.

The examination of the engines through the air intakes did not
reveal fractures of the fan blade in operation. On the other hand, the
examination of the visible cowls or casings did not reveal evidence of
instability of the rotating sections nor a rupture of the turbine disc.

On examining the hot sections parts, no trace of metallization

of the turbine stages was noticed.

On the basis of these examinations and the study of the engine

data given by the DFDR, the Pratt and Whitney expert had come to the
conclusion that :

- these engines had run normally until the moment

they separated from the aircraft,

- there is no evidence of an engine failure which

would not have been contained or of an external
fire as long as the engines were on the wing,

- there is no evidence of an engine defective

operation which could have prevented them from
developing power,

- the engine fittings on the pylons were unbroken.

1.16.2 Texts concerning the periodic monito-

ring of the engines fittings

1.16.2.1 Texts in force at the time of

the accident

23
 CASES OF THE INBOARD ENGINES

The 88-24-10 airworthiness directive (AD) (dated 1988 according

to the American coding system) concerns the pylons fittings'periodic
monitoring for engines n°2 and 3.

This AD is part of the "Boeing Supplementary Structural

Inspection Document" (SSID) program, reference 54-A45-02. It is effective
from a 12000 hours/4300 cycles ageing of the airframe and a 1500 hours/600
cycles reccurence frequency. It consists of "carry out of a close visual
inspection for cracks in both midspar fittings in the upper tang root
area and in the exposed surfaces of the lug" (visual inspection method).

This AD 88-24-10 replaced a previous AD (AD-77-09-03) of the

same nature. In fact, these two successive ADs ratify the BOEING Service
Bulletin n°3183, issued in June 1975 and developped over time. This Service
Bulletin describes the midspar fittings periodic visual inspection method
and specifies that it is possible to replace the cracked fittings by
reinforced fittings.

The AD in force is justified by the existence of a repetitive

technical defect. In fact, since 1965, 46 crack cases on the midspar
fittings have been recorded. In 4 extra-cases (the present accident
included), a loss of at least one engine during flight had happened.

Only 3 cases among the 46 cases quoted above concern aircraft

which have undergone the "HUSH KIT" modification for sound insulation
of the engines (the modification procedure began in 1986).

Among the 4 cases of engine loss during flight (the first case
dated May 4, 1977), 2 aircraft were modified for "HUSH KIT".

The Service Bulletin and the AD can be found in Annex 5.

CASES OF THE OUTBOARD ENGINE S

A similar inspection must also be done in order to monitor

periodically the outboard engines fittings, without any existing AD, in
this case. This inspection is not in the scope of the BOEING Service
Bulletin 3183 and is integrated in the SSID program under the 54-A40-02
reference. This periodic inspection (which could be found in Annex 6)
is effective from 19000 cycles and at the frequency of 500 cycles. It
consists in a visual inspection of the two midspar fittings for cracks
on the upper tang level and in the fitting lugs area.

According to the information given by the National

Transportation Safety Board (N.T.S.B), BOEING has recorded 9 cases of
inboard midspar fitting fractures, including the Miami incident
(paragraphe 1.16.5 below). The distribution of the fractures between the
HUSH KIT modified aircraft and the unmodified ones is unknown.

1.16.2.2 Development of the Publica

tions following the present
accident

This development only deals with the case of the inboard engines
pylons fittings.

Following the 5N-MAS accident and another one, which occurred

24
on 25 April, 1992 in Miami, the American Civil Aviation Authorities took
the following measures :

- the N.T.S.B issued a safety recommendation on

25 April, 1992 (n°A92-38) asking the "Federal
Aviation Administration", (F.A.A.) to proceed to
the review of the AD 88-24-10 described above to
increase its efficiency by reducing the time
between inspections and by improving the inspec tion
process.

- the F.A.A. :

. published in the "Federal Register", on 23

September, 1992 a draft amendment 39-83-73, AD
92-19-15 by which it explains the aim of the AD
88-24-10 (Service Bulletin 3183) and calls for
comments of the concerned parties by 23 Novem
ber, 1992.

. following the comments made during this first

consultation, published on 27 January, 1993 a
modification to the AD 92-19-15 in order to
reduce the time between inspections of the
fittings and to remind that a reinforced fitting
is required to replace the cracked fittings.

. then published in the Federal Register of the 4

June, 1993 the AD 93-11-02 which amends the AD
92-19-15 :

. the periodic inspections of the fittings are

maintained,

. in case of crack discoveries, the fittings

are replaced by reinforced fittings. Then,
periodic inspections are no longer neces-
sary.

The AD 93-11-02 concerns, according to the estimations of the

F.A.A., approximately 50 B 707 aicraft all types, modified for HUSH KIT
or not.

These documents are in Annex 5.

1.16.3 The 5N-MAS up-to-date maintenance

The investigation was conducted, at the BEA request, by the

A.A.I.B. (Air Accidents Investigation Branch) which investigated the
Modern Jet Support Centre (MJSC), company in charge of the aircraft mainte-
nance since June 1989.

The investigation results, centred on the accomplishment of

the operations relative to the pylons fittings checks (SSIC 54-A45-02
- inboard engines and SSID 54-A40-02 - outboard engines) are summarized
below.

25
 Inboard engines pylons fittings'inspection

The lastest inspection (according to the SSID 54-A45-02) was

performed on 10 October, 1991 at 60779 hours/
17873 aircraft cycles.

Both works requests and the two corresponding execution reports

are presented in the pages B11 to B14 of Annex 6.

The B11 and B14 pages correspond to the description of the

requested works : visual inspection for cracks in the midspar fittings
(carry out close visual inspection for cracks in both midspar fittings
in the upper tang root area and in the exposed surfaces of the lug).

The B12 and B13 pages are the execution reports for the works
achieved on respectively engines n°2 and 3 (cards E3 0557 and E3 0547).
These cards only report one single correction action : retightening of
the two midspar fittings bolts. On the left engine as on the right one,
the visual inspections did not reveal any defects.

It should be noted that the previous inspection had been

performed by the MJSC on 23 May, 1990 at 59947 hours/17686 cycles, that
is to say only 832 hours/187 cycles before the accident, when this
inspection had to be performed only every 1500 hours/600 cycles. The
A.A.I.B. had been able to prove that the most recent one - on 10 October,
1991 - had been performed, on the owner-seller request, in order to give
the aircraft maximum lifetime before any new maintenance.

To sum up :

- the inspection of the inboard engines pylons

midspar fittings was performed twice within
17 months, at a clearly lower interval to the
interval imposed by the AD (832 hours/187 cycles for 1500
hours/600 cycles asked).

- the last inspection, performed only 116 hours/34

cycles before the accident did not reveal the
cracks.

Outboard engines pylons fittings'inspection

The lastest inspection (according to the SSID A54-A40-02

program), was performed on 21 June, 1991 at 17873 cycles. The B10 page
reproduces the work request (numbered 588) certifying the inspection of
engines n°1 and 4, inspection which did not result in the discovery of
defects.

1.16.4 The 5N-MAS recent utilization

It is noted that the aircraft performed no flights between 24

26
October, 1990 and 3 March, 1992.

Afterwards, between 4 March and 31 March, it performed 116 hours

in 34 flights (that is to say as many cycles), without including the flight
duration relating to the accident.

Note that on 4 March, that is to say the very day of its aerial
activity resumption, the aircraft suffered a heavy landing in HongKong
which caused a bounce and the bursting of two tyres of the left main landing
gear.

1.16.5 The similar recent accidents/incidents

The accident of Miami :

On 25 April, 1992, the Colombian B 707 - 324C, registred

HK-3604-X, cargo aircraft, took off from Miami. During the take off, the
right inboard engine separated from the wing and hit the pylon and the
engine air intake n°4, which however remained hooked up to the wing. The
aircraft landed without any other incident.

The examination of the engine pylons fittings n°3 revealed the

existence of a fatigue crack on the inboard midspar fitting (it is on
the same fitting that a fatigue crack was found for the 5N-MAS . However,
this crack was not located on the same place). Annex 7 presents the
substance of the information asked for by the B.E.A. of the N.T.S.B.,
given the similarities between the two events.

This B 707 had a total of 53257 flight hours and 20399 cycles.
It had also undergone the modification of sound insulation of the engines
(HUSH KIT) and, like the 5N-MAS, was to follow the AD 88-24-10.

According to the report of the N.T.S.B., the fatigue crack could

have initiated from a burn during rectification (in this case, the
rectification is a machining operation realized after chrome plating).

The incident of Miami :

On 2 June, 1992, on the B 707 - 351C, cargo aircraft of the

Cordoba Air Airline, registred N8091J, during inspection before flight,
a defect was discovered on the engine pylon fittings n°4.

Close examination revealed the existence of a double fracture

of the inboard midspar fitting. The upper tang and the lower tang had
broken from the development of fatigue cracks whose causes, according
to the sketchy information we actually have, could present analogies with
those of the Miami accident described previously (rectification crack).

This aircraft had a total of 54175 flight hours and 20651 cycles.
It was modified for HUSH KIT. The last inspection of the midspar fittings,
in the frame of the SSID inspection program (A 54-A40-02), had been
completed in November 1990, at the aircraft life of 50049 hours/19773
cycles.

1.16.6 Fuel circuit examination

27
 The left wing fuel circuit scheme is on diagram 3. As the right
wing fuel circuit was perfectly symmetrical, the indications identified
1 and 2 become respectively 4 and 3.

The shut off valve of the engine n°3 and the two
transfer valves of the engines n°3 and 4 were located in an inboard dry
bay reached by an access panel in the wing skin near engine n°3.

The shut off valve of engine n°4 was located in an outboard dry
bay near engine n°4.

These 4 valves were identical. They were electrically operated

and equipped with an indicator lever whose position on the wing it had
been possible to mark, as their dry bay had not been damaged.

- the shut off valves were in the off-position

(lever in the up position)

- the transfer valves were open (lever in down

position)

These positions corresponded to those of the instrument panel

switches of the fuel circuit in the cabin (shut off valve : off-contact
- transfer valves : on-contact).

The 4 valves were removed and tested in a laboratory. Supplied

with 28 Volts DC, the shut off closed entirely during every manoeuvre.

Therefore, we had to come to the conclusion that the respective

conditions of the 4 valves were in agreement with the instrument panel
and corresponded to the testimony of the flight engineer who indicated
he had manipulated the shut off valves.

In this context, the fuel leakage on the leading edge of engine

n°3 could not have been caused by a closing failure of the shut off valve.
Damage of the pattern following the pylon detachment could be the cause.
The exact location of the leak could not be detected.

1.16.7 The path of the aircraft

The path of the aircraft was determined from the radar fixes
and is in Annex 2.

The legend on the map was derived from the CVR dialogue and
the ATC communications with the concerned control organisations.

1.17 Additional information

1.17.1 On board

After the accident, the crewmembers described how the event

took place. The following indications can be deduced from their respective
actions :

Captain

First of all, the accident was evident as double bang and severe
turbulence. The captain concentrated on the piloting in order to keep

28
control of the aircraft. In his statement, he praises the self-control
and the professionalism of his crew.

First officer

The first officer indicated that, before starting the climb

to flight level 330, the aircraft, at flight level 290, met quite severe
turbulence. The loss of the two engines occurred during this climb.

He also specified that, in Istres, given that the aircraft was

capaple of limited manoeuvring, it was not possible to make a left turn
to come properly to the QFU 33 circle to land, as the controls were fully
held to the left. Therefore, the captain made a big left pattern to land
down wind.

On the other hand, on the ground, the aircraft tended to turn

to the right.

Flight engineer

The flight engineer emphasised the problems caused by the

continuously ringing "engine fire" and "depressurization" warning horns.
He did not manage to switch them off despite his numerous efforts on the
control panel.

He specified the following points :

- the navigation radar which had been working

before the loss of the two engines, failed,

- he closed the right shut off valve of the

pressurization circuit to deal with the problem of the
loss of pressurization induced by the departure of
the two right engines,

- he closed the fuel shut off valves of engines

n°3 and 4,

- he closed the hydraulic pump supply shutoff valve

of engine n°3 and verified on the engineer
instrument panel the state of the hydraulic
services. He confirmed that the rudder power unit
was working normally,

- he dealt with the problem of the electrical power

supply by connecting the essential bus to
the A.C generator n°1 (engine n°1),

- he sent the cargo supervisor to examine the right

wing from the cabin windows (note : the cargo
supervisor specified that the two right engi nes
had really disappeared, the leading edge was
unbroken and that there was a fuel leakage).

- he had some difficulties in executing the

29
 lightening procedures of the aircraft by fuel
dumping. Actually, after he manipulated the dump
valves of fuel tanks n°1 and 4, he noticed
that n°1 fuel tank dumping did not occur,
a fact which increased the problems of assymme- try of
the aircraft. He stopped the operation and
selected the dumping of the centre tank.
Afterwards, he noticed that the breaker of the
dump system of tank n°1 had popped. He
reconnected it and was able to dump symetrically
both tanks,

- in the landing circuit, he lowered the flaps in

accordance with the emergency drill (electrical
energy). This drill was slower than the normal one (in
hydraulic mode), it was not possible to lower the flaps entirely
(38° instead of 50°).

1.17.2 On the ground

Several people on the ground saw the landing. Two testimonies

particularly hold attention.

The first person is a member of the military staff of the air

base. He looked at the B 707, coming from the west and flying over the
runway. He asserted that the aircraft was not on fire at this time.

Afterwards, he had lost sight of the aircraft and saw it again

later, coming from the east and making its last turn. He asserted that
the right wing was on fire at this time, "the fire was small at first
and then it became a huge fireball".

The second person is a member of the base fire brigade (SSIS).

He asserted that, then over the base, "no part of the aircraft was on
fire and he identified under the right side a sort of fog". Then, when
the aircraft turned left to line up, he saw orange flames on the right
wing.

On the other hand, the right wing fire, just before the landing,
was confirmed by the controller who advised it to the crew (see Annexe
1).

2. Analysis

2.1 Loss of the two right engines

The examination of the pylon fittings of engine n°3 identified,

on the one hand, the existence of a fatigue crack on the inboard midspar
fitting mount, and on the other hand, the appearance of the static fractures
of the other three fittings as the engine drifted rightwards.

The examination of the wreckage of engine n°4 and its pylon

indicated that the fractures of the fittings were all of a static type.

Moreover, the examination of the engine air intake n°4 identified

30
the existence of an impact on the inboard side as well as white paint
marks of the same nature of those of the engine case n°3.

Finally, the examination of the engine power data given by the

replay of the DFDR indicated that from the recorded time "3200", the power
sensors of engines n°3 and 4 stopped transmitting indications. So, we
had to come to the conclusion that the two right engines separated almost
at the same moment.

Considering all these elements, here is the final scenario for

the loss of the two engines, (see diagram 4):

- as engine n°3 was running at the climb power

(EPR = 2.0) and was torn off the wing because of
the initial fatigue fracture of its pylon inboard
midspar fitting and the consecutive fracture of
the other fittings, it was propelled to the out
board. The fittings fracture process was probably
assisted by the increased strains induced by the
turbulence reported by the crew.

- after spinning round approximately 3/4, following

its longitudinal axis, it hit engine n°4 on
its air intake and then swung over it.

The violence of the shock induced the tensile fracture of the

forward engine fitting n°4. The midspar fittings broke in turn, then the
aft fitting, the engine eventually separated from the wing by swivelling
around it.

This sequence was consistent with the evidence of the crew (the
"double bang" heard), the position of the engines on the ground in
comparison with the flight path (before the beginning of the turn) and
with the start in right roll attitude of the aircraft induced by the massive
assymmetry of the left engines thrust only.

On the other hand, it was essential to compare the present

accident with the Miami accident which occurred on 25 April, 1992.

2.2 Development of the inboard mispar fitting

fatigue crack of engine n°3

In the paragraph 1.16.1.1, it appears that the examined fatigue

"half-moon" includes two different areas:

- the major part of its surface is black, result of

significant oxide deposit,

- at the end of the cracking, the area is edged by

a clearer band which is approximately one
millimetre across. This difference of colouring is
certainly linked with a prolonged storage of the aircraft,
corresponding to the limit of the two clear and dark areas.

These observations permit precise assesment of development of

the crack with time :

31
 - birth of the crack at a date located noticeably
before the prolonged storage of the aircraft
(that is to say before 24 October 1990),

- the crack progressed slowly until the cessation of the

flights. Then, during the ground period this area became
oxidized.

- at the resumption of the flights, the crack

continued to propagate, but the crack had no time
to become oxidized. This part corresponds to the
clear area.

On the other hand, the electron microscope examination of the

clear area (that is to say at the end of the cracking), enabled to count
25 to 30 stoppage lines. This number is consistent with the recent activity
of the aircraft (34 flights).

2.3 Monitoring of the pylon fittings

The AD 88-24-10 (SSID 54-A45-02) imposed the checking of the

inboard engine pylons fittings. The maintenance operation SSID 50-A40-02
imposes the checking of the outboard engines.

The last checking of the inboard engines was carried-out on

10 October, 1991, the life of the aircraft was 60779 flying hours/17873
cycles.

The penultimate checking dated back to May 23, 1990, the aircraft
had then 59947 flying hours/17686 cycles.

It is convenient to note that these two checkings are separated

by only 832 hours/187 cycles whereas the recommended frequency is 1500
hours/600 cycles.

However, according to the metallurgic examination it appeared

that on 10 October, 1991, date of the latest AD 88-24-10 execution, the
crack existed, its propagation surface was corresponding then to the black
oxidised area: consequently, the recommended check appeared to be ineffec-
tive because it did not allow detection the existence of the crack.

On the other hand, even though the date of the "birth" of the
crack is unknown, it was probable that the crack was already existed during
the penultimate check, on 23 May, 1990. It is then possible that the
penultimate check had also not allowed detection of the crack.

So, it seems that the inspection method recommended by the AD

is imperfect because of a difficult operating method.

It consists of a visual inspection of the visible parts of the

attachment fittings. Nevertheless, in the case of the accident, the crack
is located on the bore. Consequently, it is masked by the inboard clevis
lug of the wing and escaped a visual inspection. In these conditions,
only dismantling of the pylon from the wing could have permitted its
detection.

32
 In the case of the Miami accident, the crack was also located
under the inboard clevis lug and was also hidden in the same way and cannot
be visually detected.

Besides, we observe that :

- the pylon midspar fittings on the wing have,

since entry into service of the B 707, shown
weaknesses which have resulted in about fifty
fatigue cracks,

- the midspar fittings structural weakness concerns

the inboard engines first of all, but also, and to
a lesser extent, the outboard engines,

There are good grounds for saying that this phenomenon is

connected with the life of the aircraft. In fact, we must note that "5N-MAS"
and the two other aircraft in Miami had a total of more than 50000 flying
hours and more than 17000 cycles. On the other hand, the HUSH KIT modifi-
cation might have had a negative effect on the resistance of these fittings
(see Annex 5, the F.A.A. commentaries in the successive AD).

2.4 Right wing fire

The two witnesses on the ground asserted that, when over the
base, the wing when the aircraft was coming from the West was not on fire.
One of them was sure that he saw, under the right side, "a kind of fog".

This fact corroborares the testimony of a passenger who went

to verify the condition of the right wing and noticed a fuel leak.

It should be noticed that the fire marks on the upper surface

of the wing went from the leading edge of the pylon of engine n°3 and
widened towards the trailing edge whereas there was no fire mark on engine
n°4. The fuel leak which was at the origin of the fire, was consequently
located on this place.

It is reasonable to think that as a consequence of the closing

of the two fuel shutoff valves of engines n°3 and 4, the fuel leak of
the engine n°4 stopped whereas the leak of engine n°3 persisted as a
consequence of a system damage induced at the moment of the pylon tearing
away.

During all of the descent at a variable but always greater than

220 kt speed, it is probable that the fuel leak carried on without the
fuel catching fire, as the conditions of ignition (depression of the
upperwing, speed...) were not achieved and the vaporized fuel was not
in contact with the electrical short-circuits of the damaged cabling loom
located on engine n°3 leading edge.

These conditions changed during the last turn in consequence

of the semi-extension of the flaps. The speed reduced (between 220 and
190 kt), the depression on the upper wing and the turbulence increased.
Then, it was possible that under the effect of the electric arcs of the
short-circuits quoted above, the fuel ignited, as the conditions of the
kerosene-air mixture became optimal for burning.

33
 This fire was violent as the condition of the upper wing
demonstrated, particulary at the trailing edge. This intense fire had
destroyed the trailing edge as well as the flaps and left evidence of
overheating on all the right aft part of the fuselage.

The Istres controller advised that the right wing was on fire
at 08.33.28 hrs and the landing touchdown occured at 08.35.35 hrs. Conse-
quently, the right wing fire lasted more than two minutes without the
possibility of being more precise on this point.

It is more difficult to understand the process which provoked

the explosion of tank n°4, an explosion which had seriously damaged the
front and the aft spar of the wing.

The most likely hypothesis seems to be the following:

- the intense fire on the upper wing brought the

soft material skin panel up to a high tempera- ture, diminished
its mechanical resistance, a fact which explained the
crazing, the blisters in all the area blackened by the fire and more
particularly in the nearest fuel leak .

- as a result of the difference of pressure between

the outboard and the inboard of the wing, the
overheated soft material plate broke, inducing
the noted crevices .

- by the crevices created this way, and given the

kerosene vapors of the breather system, the fire
progressed little by little and provoked the
explosion of tank n°4. The drawing, diagram 0,
which detailed the damage of the right wing
showed the position of this tank. It was the
most exposed tank because it was located in
its most inboard part, on the level of engine
n°3.

The explosion in the wing probably occurred during flight, a

short time before the landing if we consider the fireball mentioned by
a witness.

2.5 Distress piloting

Listening to the CVR and the testimonies collected have

permitted outline of the behaviour of the crew in this distress situation.

The captain

He maintained the handling of the aircraft. Immediately after

the loss of the two right engines, he went to the essential and tried
to keep control by counteracting on the controls. The rudder power unit
worked normally, the aileron and the pitch controls were intact.

34
However, the CVR indicated an intense physical effort. Therefore, it seemed
that this major effort in order to "maintain" the aircraft was the
consequence of the imbalance due to the weight between the two wings and
of the assymetric thrust of the two left engines.

During the descent, the captain limited thrust by lowering the

flaps as late as possible.

At the same time, he gave brief orders and took quickly the
required decisions. Thus, he agreed straight away with the flight
engineer's proposition of lightening the aircraft by defuelling through
the dump valves. He was concerned about the weather conditions and asked
the first officer to report that the aircraft had limited manoeuvring.

Then, he decided to land on the Istres airfield, fortunately

located on the track and whose runway is very long. The landing approach
and the final turn were performed by left turns, the conventional rule,
which gave preference to the turn performed on the working engines rather
than on the broken engines, was respected.

The first officer

First he just noticed the loss of the outboard engine. He was

in charge of the radio traffic and guided the captain to the intended
runway thanks to the headings given by the ATC control and asked the flight
engineer for the execution of the "emergency operating procedure"
checklist.

Finally, when he saw the Istres runway, he asked for information

on the nature of the ground and the length of the runway. Conscious that
the aircraft had limited manoeuvring, he suggested operating a counter
QFU circuit to the left. He advised the captain to keep a minimum 200
kt speed and helped during the landing by holding the left engines power
handles.

For the completion, he also had time to take a picture, in flight,

of the right wing leading edge where engine n°4 had disappeared ...(Flight
International 4/10 November 1992).

The flight engineer

He successfully completed the following actions :

He resolved the depressurization problem by insolating the right

circuit by closing the shut off valve of the circuit.

He closed the fuel shut off valves of engines n°3 and 4.

He isolated the hydraulic system of engine n°3 by closing the

hydraulic shut off valve.

He resolved the problem of the emergency electrical energy by

connecting the "essential bus" to A.C. generator n°1.

35
 He sent a passenger to check the condition of the right wing
from the passenger cabin windows.

He suggested lightening the aircraft by the rapid fuel dump

in flight system. As the captain agreed, he dumped fuel despite the
difficulties on fuel tank n°1.

Helped by the maintenance man, he extended the gear by the

emergency system and verified its complete lock.

During the last turn, he lowered the flaps in accordance with

the emergency procedure.

Finally, noticing that the captain, because of the normal brake

system failure (hydraulic mode), used the emergency brakes (pneumatic
mode), he selected "reverse" on engine n°2.

General comments

During approximately the last 24 minutes, during which the

manoeuvre was operated, the crew was bothered by several warning horns
that the flight engineer did not manage to switch off.

The crew had no emergency operating procedure corresponding

to the present case. As a result, the crew used the procedure corresponding
to the "Engine fire, severe damage or separation" checklist (card 5).

However that may be, the execution of this emergency operating

procedure proved to be efficient, as the aircraft remained manoeuvering
until the landing and preserving the essential of the necessary emergency
ancillaries :

. the hydraulic rudder booster worked normally in

emergency,
. the normal braking was inoperative, but the
(pneumatic) emergency braking worked,
. the electric power enabled to fuel dumping,
lowering of the flaps in emergency and ensured
the normal operation of the ATC radar beacon system
and communications.

The major workload associated with this exceptional event was

remarkably well shared between the crewmembers and this good organization
led undoubtedly to the success of the manoeuvre.

2.6 Control assistance

Listenings to the ATC communications and the CVR enabled

analysis of the successive phases of the dialogue between the crew of
the distressed aircraft and the controllers.

CRNA/SE

36
 The first two distress calls were timed at 08.11.41 hrs and
08.11.48 hrs. At 08.11.53 hrs, the controller of the CRNA/SE answered
: "say your level and position Sir". He repeated his question at 08.12.20
hrs advising that he did not have radar contact.

Comments The reconstruction of the radar traces performed by

the CRNA/SE confirmed the momentary disappearance of the secondary image
of 5N-MAS between 08.10.50 hrs and 08.12.20 hrs (see Annex 2). Thus, during
1 minute 30 seconds, the controller was not able to identify the aircraft
from the radar. This disappearance of the trace was explained by the fact
that the electrical power supply of the transponder was supplied by the
"essential service bus", connected with engine n°3 at the moment of the
separation of both engines. The transponder no longer transmitted because
it was no longer electrically supplied. It got back to transmit - and
the secondary image reappeared - when the flight engineer connected the
essential bus with engine n°1.

Then, the controller identified the aircraft undoubtedly from

the rereading of the strip. The controller asked the captain to get the
aircraft at flight level 200.

The controller asked the crew to confirm the landing in

Marignane. After the answer ("Anywhere landing immediate"), he requested
the display of the transponder distress code 7700 and ordered the crew
to turn southwards, towards Marignane.

The crew advised the control of the nature of the troubles at

08.14.13 hrs : "We need emergency landing emergency landing both engines
missing right wing emergency landing radar". It should be noted that this
sentence could be understood in two different ways. There is no doubt
that the crewmembers wanted to warn the controller about the detachment
of two engines. In fact, it is highly probable that the controller
understood that the crew wanted to land because of inoperative engines.
The aircraft was transferred to Marseilles Control.

Marseilles Control

As soon as the hand-off was made, the new controller tried to

get pieces of information. At 08.16.22 hrs, he asked : "Ah Roger proceed
direct to Mike Romeo Sierra. What the nature of your problem ?".

The answer of 5N-MAS is more accurate : "Two engines missing

two engines missing structural request straight-in landing". However it
might have been interpreted by the controller as a double engine failure
if the word "structural" was not perceived or understood.

At 08.17.14 hrs, the controller repeated his question, which

seemed to confirm that he had not under-
stood : "How many engines on failure...671 ?". The answer ("Two engines
missing ...") did not remove the ambiguity. At 08.17.51 hrs, KABO 671
asked the meteorological conditions in Marseilles.

The conversation with 5N-MAS was hindered by the radio traffic

of the controller with other aircraft on the same frequency. At 08.18.40
hrs, KABO 671 specified once more the reality of its distress call :"MAYDAY

37
MAYDAY
MAYDAY MAYDAY MAYDAY MAYDAY 671 request weather". Then the aircraft was
transferred to Marseilles Approach, requests for information on weather
at Marseilles not obtained.

Marseilles Approach

As soon as 5N-MAS was transferred to the frequency of Marseilles

Approach, the controller gave the crew the heading 240 and specified that
the runway in service was runway 14. On the same frequency the UTA 7209
flight, just taking off from Marseilles, signalled that it was "climbing
towards 3000 ft".

At 8.20.33 hrs, KABO 671 specified once more the reality of

its distress call :"Roger request all assistance possible we have two
engines broken from the airplane only running on one and two". The
controller answered "Roger I understand" (in fact and always for the same
reasons, he might have not understood the real situation of the aircraft).

Then, KABO 671 mentioned that the aircraft had limited

manoeuvering and specified that they wanted the meteorological conditions
in Marseilles. It is at this very moment that, doubting the good
meteorological conditions in Marseilles, the crew considered to re-route
towards Palma and asked for the corresponding meteorology. The follow-up
of the dialogue revealed that the crew quickly forgot this idea; they
finally got the detailed meteorological conditions in Marseilles,
answered the questions of the controller about the number of people and
the quantity of remaining fuel. Then the controller gave the heading 180°
in order to direct the aircraft towards runway 14 (see path in Annex 2).

The dialogue with 5N-MAS was hindered by the radio traffic of

the controller with other aircraft on the frequency.

The controller got the aircraft turned left to the heading 110.
A short while later, 5N-MAS asked :"KABO we
have an airfield ahead what is that airfield ?". It was the military airbase
of Istres, with its 4000 meters long runway. The crew suggested to land
there ("4000 m I can land there ?"). The controller gave the landing
clearance. The aircraft was transferred to Istres Tower.

This analysis reveals that the various people in contact with

5N-MAS became aware only gradually of how
serious the situation really was and without, moreover, perhaps ever
identifying its real cause. Notably, the

urgent calls and MAYDAY did not suffice to produce this realization. The
service provided was not adequate for the seriousness and urgency of the
plane's distress.

We might note, among other things, that the repeated requests

for frequency changes and the continuation of exchanges with other aircraft

38
increased the crew's workload. When the plane signalled that it was in
distress, requesting permission for an emergency landing, no analysis
was made of the airfield and means of assistance available. Also, when
the plane, whose diminishing ability to manoeuver was known, approached
Marignane to land on 14, take-
offs were continued with no allowance apparently made for this factor.
Lastly, there was no reply to requests for information on weather
conditions until relatively late. And yet a reply, even to hold, would
have spared the crew the impression that their calls were falling on deaf
ears, which added to their nervous tension.

None of the above can be undertaken lightly, particularly in

an emergency situation. The ability to respond properly to such situations,
when they arise, depends on being prepared and trained for them.
Flight-controller personnel, however, does not systematically receive
this type of training.

3.Conclusions

3.1 Data given by the inquiry

The crew had the certificates, the licences and the necessary
qualifications for the operation of the aircraft.

The maintenance operations recommended were performed according

to the instructions in force.

Climbing in turbulent air, following the fracture of the

fittings because of a crack propagation, engine n°3 was propelled outboard
and hit engine n°4, inducing its tearing away.

The fracture of the fittings of the pylon of engine n°3 started

from an inboard midspar fitting fatigue crack. This crack was not detected
during the double execution of the AD which imposed the periodic visual
inspection of the midspar fittings. The visual inspection described by
the AD was not sufficient to permit this detection.

The distressed aircraft control assistance revealed some

defects in the preparation to the handling of emergency situation. However,
in this case, those defects were not at the origin of the accident. There
did not lead neither to worsened consequences.

The crew succeeded in forced landing, right wing on fire, on

the Istres runway.

The efficiency of the fire brigade (SSIS) of Istres avoided

the complete destruction of the aircraft and its freight.

3.2 Causes of the accident

The accident resulted from the fracture of the right inboard

engine pylon fitting, in such conditions that this engine came to hit
and tore away the outboard engine.

The AD, imposing periodic monitoring of the midspar fittings,

proved to be insufficiently efficient.

39
4. Safety recommendations

4.1 The B 707 pylons fittings

This accident brought to light the weakness of the B 707 pylons

fittings. This structural weakness provoked four cases of loss of engines
in flight.

The successive AD in force since 1977 shows that this problem

is not recent. Because they are limited to visual verifications without
dismantling, they failed to detect hidden fatigue cracks, such as those
described in the present report.

On the other hand, it is not impossible that the HUSH KIT

modification is an aggravating factor, increasing the probability of crack
developing.

Consequently, the B.E.A. recommends that :

- in order to ensure the safety of

the flights, the inspection of the current
midspar fittings of the pylons of the engines be
modified in order to enable the detection of the
hidden cracks, or that these fittings be
systematically replaced by reinforced fittings.

Comment : these reinforced fittings should undergo an

examination in order to respond to the new conditions which could be induced
by the sound insulation modification of the engines.

4.2 Control

This accident revealed defects in the management of the

distress. They are linked to the unsual nature of the situation (break
in the routine) and to the sudden nervous tension (stress) due to the
distress situation of the aircraft. If the Air Traffic controllers are
prepared for this kind of situation during their initial training, they
do not have afterwards a specific periodic training.

Consequently, the B.E.A. recommends that :

- the controllers be trained to face up to the distress and

emergency situations, thanks to the theoretical study of possible cases
and the practical corresponding exercises.

40