At 1919:48 local time on a cool summer’s evening, American Airlines Flight 96 lifted off from runway zero three right (03R) at Detroit’s Metropolitan Airport. The date was 11 June 1972. Flight 96 was a DC-10 (Series 10) scheduled service from Detroit to Buffalo and was lightly loaded with only 56 passengers and 11 crew. The aircraft had originally departed from Los Angeles and its final destination was La Guardia in New York City.
Detroit, Michigan, lies on the northwestern flank of Lake Erie, while Buffalo, New York, lies at its most eastern point, with the US/Canada border running lengthwise along the centre of the lake. As the American Airlines DC-10 turned due east under ‘departure’ radar control, the aircraft (registration N103AA) continued on its climb towards the Canadian town of Windsor in Ontario, with flaps and slats being selected up in sequence.
The visibility was only 1 1/2 miles in the still, hazy conditions and the cloud base was 4,500 feet. Just under 3 minutes after take-off the DC-10 entered cloud and at the same moment was cleared for further climb to flight level two one zero (21,000 feet). First Officer (F/O) Peter Paige-Whitney was at the controls with the commander, Captain Bryce McCormick operating the radio. Second Officer (S/O) Clayton Burke occupied the flight engineer’s station. Flight 96 was running about one hour late because of a number of delays, the most notable being the problem in closing the rear cargo door. Difficulties with the door had occurred on a number of occasions and it had taken 18 minutes to close it in Los Angeles. In Detroit it had taken only 5 minutes to shut the door but the ramp service agent had to use his knee to force down the locking handle. He was sufficiently concerned to mention it to an American Airlines mechanic who quickly checked the handle and found it safe. On the flight deck the ‘door open’ light extinguished confirming the door locked.
Departure control now instructed Flight 96 to contact Cleveland Centre on 126.4 MHz. Although the short flight routed mostly through Canadian airspace, US controllers maintained contact. With the after take-off check complete and the aircraft climbing safely at 250 knots, the first officer engaged the autopilot. Captain McCormick called Cleveland Centre climbing through 7,000 feet and was instructed to select radar transponder code 1100. The DC-10 was further cleared to flight level 230. At 10,000 feet the copilot set the vertical speed control to 1,000 feet/minute climb to reduce the angle for acceleration from 250 knots to the normal cruise speed of 340 knots. Passing about 11,500 feet, directly over the town of Windsor with the speed increasing through 260 knots, N103AA broke from cloud cover into a clear evening on top. A Boeing 747 could be seen high in the sky above.
“There goes a big one up there”, remarked Captain McCormick.
The flight crew began to relax and to take the view. Suddenly without warning, a resounding thud echoed from the rear of the aircraft. Someone yelled, ‘Oh shit’. The rudder pedals ‘just exploded’ and smacked to the full left rudder position. The captain was resting his feet on the pedals and his right leg was thrown back against the seat with extreme force. F/O Paige-Whitney hit his head on the back of the seat. At the same moment the three thrust levers snapped back to flight idle with the number two (tail) engine throttle hitting the stop with a loud crack. A great rush of air swept past the flight crew throwing dirt and grit into their faces and blinding them with the dust. The stinging effect was like a fire cracker going off below their noses. The captain’s headset was knocked from position and as the aircraft jerked the autopilot automatically disconnected. The captain’s first thought was that the windshield had failed, but when his eyes cleared he could see it still in place. Disbelievingly he stretched his hand out to touch the window.
‘What the hell was it, I wonder’, called Captain McCormick.
One of the crew replied with a long whistle. The fire warning bell rang for number two engine and the cabin altitude warning horn sounded indicating the cabin air pressure had reduced to an altitude equivalent to over 10,000 feet. The autopilot red ‘disconnect’ light flashed and a red failure flag appeared on the airspeed indicator. The captain then thought the radar dome might have gone. With the aircraft nose cover missing erratic airspeed indications could be expected. The first officer, by habit, still had his hands on the controls, but the aircraft banked slowly to the right out of control and the nose dropped sharply.
‘We’ve hit something’, said the second officer.
The co-pilot had been looking out and had not seen anything of another aircraft so thought it more likely to be disintegration of number two engine. That would explain the fire warning and the rudder problem, although the fire indication subsequently proved false. Whatever the cause of the trouble there was no doubting the severity of the problem. In normal circumstances, with the cabin pressurisation warning horn sounding, the crew would have been tempted to commence an emergency descent, but the captain was reluctant to force the aircraft into a steep dive until the damage could be assessed. Flying at around 12,000 feet, few breathing problems would be experienced, although strictly speaking flight crew should wear oxygen masks above 10,000 feet.
In the first few seconds after the loud bang the confusion in the cabin was just as great. At the rear of the aircraft Stewardess Sandra McConnell sat next to the right-hand emergency exit chatting to Stewardess Beatrice Copland who was positioned on the opposite side. Just forward of the girls was a small lounge bar normally used for cocktail service. Since there were very few passengers and the sector was short it was considered not worth offering the facility and no passengers were seated in the area. Both stewardesses had undone their seat belts and were thrown from their seats with the ‘explosion’. Stewardess McConnell was thrown against the bar portion and landed by the edge of a gaping hole in the bar floor. She could feel herself slipping into it as the floor around crumbled into the gap. Stewardess Copland found herself lying in the hole and she could see into the cargo compartment. The bar unit was lying nearby and her head and one foot were trapped by the debris which had fallen from the ceiling. She called out for help.
Stewardess Carol McGhee usually worked in the cocktail lounge, but since her services were not required she sat strapped in by the front exit for longer than normal. Suddenly she heard a noise like a ‘frump’ and saw the escape hatch from the downstairs galley shoot up from the floor and strike a passenger on the head. The flight deck door burst open and out flew the crew’s hats. A dusty rush of air gushed from the flight deck and the cabin filled with ‘cool fog’. Stewardess Carol Stevens was also still strapped in her seat when she heard a ‘vroom’ kind of noise. The cabin fogged and air rushed from the front to rear of the aircraft. When the fog cleared she could see ceiling panels hanging down from the rear of the economy section and the bar lounge floor caved in. By the collapsed bar unit she could see Stewardess Copland pinned down the hole and could hear her cries for help. Stevens tried to call for support on the cabin interphone but without success, so ran forward to summon assistance from the other crew members.
In the service centre at the forward end of the aircraft, Head Stewardess Cydya Smith was already preparing coffee for the cabin service. Suddenly the galley lift door blew open and a smoky ‘substance’ billowed out. She was thrown off balance but somehow managed to hold herself upright. She noticed ceiling panels drop and experienced a feeling of weightlessness. Immediately she thought of depressurisation and quickly checked to see if oxygen masks had dropped in the cabin, but they remained in position. She had no difficulty in breathing. At cabin altitudes above 10,000 feet the warning horn on the flight deck sounds but it is only at cabin altitudes above 14,000 feet that passenger oxygen masks drop automatically. She rushed forward to the flight deck, noticing on the way the captain’s hat lying on the floor, and asked, ‘Is everything all right up here?’
‘No’, called the captain.
The co-pilot turned and shook his head. ‘You go back to the cabin.’
Captain McCormick retrieved his headset from the back of the seat and quickly radioed Cleveland Centre declaring an emergency. There was little information he could give as they didn’t know what had happened except that they had a serious problem. The co-pilot still handled the control column, but by now relinquished flying to the captain who wanted to ‘feel out’ the controls himself.
‘I think it’s going to fly’, assured the first officer.
The speed fell to 220 knots as the DC-10 descended towards the cloud. If they could stay in the clear until the damage was assessed it would make life easier. The captain pulled back on the controls but soon ran out of elevator to arrest the descent. He pushed the thrust levers forward: the wing- mounted engines responded but number two tail engine remained in flight idle. The number two throttle could be moved backwards and forwards quite easily and it was obviously not attached to anything. The power from the number one and three (wing) engines effectively pitched the nose up and the aircraft managed to maintain 12,000 feet. By good fortune McCormick had practiced in the simulator, flying the DC-10 on engine power only, assuming total loss of flying controls with hydraulic failure. The DC-10’s engine position permit relatively effective control when employed in this manner and Captain McCormick became quit adept at the task. He was now able to put his experimental endeavours to good use.
On returning to the cabin, Head Stewardess Cydya Smith comforted the passengers on the public address (PA) system and instructed them to stay in their seats, not to smoke, and to remain calm. As she finished speaking Stewardess Stevens raised the alarm that one of the girls was trapped at the back and required assistance. A male passenger offered to help and the three made their way rearwards, amazed at the destruction they saw as they approached the tail.
With the aircraft now under control, the captain also lifted the PA handset to reassure the passengers. They had experienced a major problem, he explained, which was now resolved but would necessitate a return to Detroit. He would keep them informed of progress. The interphone from the cabin chimed on the flight deck. One of the stewardesses, unaware of the help being summoned for the trapped girl called to say that Stewardess Copland required rescuing from a ‘hole in the floor’ and could someone lend a hand. Since the immediate flying problems seemed to be resolved, the flight engineer left his seat to help. Stewardess Copland managed on her own to wrench her head free from the debris trapping her in the gaping hole, then; leaving her shoe caught in the wreckage, pulled her foot out and was able to escape. She climbed out of the area, over a rear bulkhead, and on to one of the lounge seats where she was helped to her feet by the arriving rescuers. There was no sign of the other girl and Smith called out her name several times. As if from nowhere Stewardess McConnell appeared out of the debris and was helped into the cabin. In spite of the ordeal neither of the girls was hurt.
Second Officer Burke had only just left the flight deck and was still looking for his hat when the interphone sounded again. The two girls were now safe, he was informed, and his assistance was no longer required so he returned to his seat. In the cabin passengers near the tail were moved forward from the extensively damaged rear section while others were comforted and nursed by the attendants.
With the situation apparently under control Head Stewardess Smith returned to the flight deck to report to the captain. There were no serious injuries amongst the passengers and crew but there was a large hole in the floor and in the aft left-hand side of the fuselage. Captain McCormick instructed her to prepare for an emergency landing, and she returned to gather the other attendants in the service centre for a briefing.
Captain McCormick was still having control problems but was managing with difficulty to keep the stricken aircraft flying. The DC-10 could only bank about 15 degrees in turns. F/O Paige-Whitney now operated the radio and requested emergency services for their arrival. Radar control was going to position the aircraft on a 20 mile final approach and cleared Flight 96 for descent. The best rate of descent that could be achieved was 200 feet/minute, which was painfully slow. Radar control continued to vector the aircraft towards Detroit while the crew completed their checks. In the cabin the attendants prepared for an emergency landing and evacuation although it was not yet known if such an eventuality was likely. As a precaution, shoes and loose articles were collected, the brace position was demonstrated and the passengers were briefed on escape routes and the use of the slides. When all was ready one of the stewardesses went forward for confirmation that the escape slides would be used after landing. She popped he head through the flight deck door. ‘Do you guys have a problem up here?’ she asked. In the very tense atmosphere, with the captain struggling to maintain control, the question seemed quite ridiculous and the crew roared with laughter. ‘Yes, we have a problem’, they called.
The captain said he would activate the evacuation signal if necessary but they weren’t sure yet what was going to happen. Captain McCormick then spoke to the passengers again; in as calm a voice as possible, apologising for the inconvenience and reassuring them that the aircraft was under control and that everything possible would be done for them on the return to Detroit.
The approach was begun at 20 miles out with 160 knots indicated airspeed and a rate of descent of 600/700 feet/ minute. Attempts at reducing speed resulted in unacceptable sink rates. In order to keep the aircraft aligned with the runway the entire approach was conducted with the nose pointing 5-10 degrees to the right. The landing gear was successfully lowered and the remaining flaps extended just before touch-down. The DC-10 landed flat and fast 1,900 feet from the threshold. The landing was smooth, but once on the ground ‘all hell broke loose’. Almost immediately Flight 96 veered off the runway and ploughed through the grass. The co-pilot took control of the reverse thrust levers, pulling maximum power on the number one engine on the left and cancelling number three. This eased the aircraft left, back on to the hard tarmac of the runway, and the machine stopped about 1,000 feet from the end. After the rough landing the captain felt an emergency evacuation might be prudent and activated the alarm. Fortunately, no further danger ensued and all the passengers quickly disembarked down the chutes. In the end only a few minor injuries were sustained. But for good fortune and the skill of the crew, the result could have been much worse.
The next morning National Transportation Safety Board (NTSB) accident investigators arrived on the scene and easily diagnosed the cause of the accident. Scuff marks on the rear cargo door’s securing mechanism clearly showed that the door had not been properly locked, in spite of indications to the contrary. As Flight 96 had climbed out of Detroit, the cabin had been pressurised as usual to allow relatively normal breathing for the passengers and crew. The fuselage and the doors are designed to prevent the pressurised air from bursting outwards into the rarefied atmosphere. Any failure of door, panel or window, however, results in air exhausting rapidly with force. In this case loads of up to five tons had been placed on the partially locked cargo door. Eventually the latches had sprung under the load and the door had blown open, causing explosive decompression of the aircraft. The door had been blown off by the airflow, damaging the left tailplane in the process. The pressurised cargo hold air had immediately exhausted to the atmosphere via the gaping hole and the cabin air pressure had placed an undue load on the floor. With insufficient venting in the cabin floor area, the floor had simultaneously collapsed, trapping the bar unit in the open cargo door exit. Through the beams of the cabin floor ran control cables, hydraulic pipes and wiring of which a number had been severed or jammed, resulting in control difficulties. Fortunately sufficient control systems had remained intact to sustain stable flight. It had been a close-run event and swift and decisive action would be required to prevent a recurrence..
The ‘Windsor incident’, as the near disaster became known, prompted an inquiry by American Airlines, the NTSB, the Federal Aviation Administration (FAA) and McDonnell Douglas into the locking difficulties of the rear cargo door. A division of General Dynamics, Convair, the manufacturer of the door, was also involved. The FAA’s Western Regional Office, one of 11 such establishments in the US, was responsible for regulating California’s aviation industry. Since the state contained the biggest concentration of aviation related manufacturing talent in the world, the regional office’s administration duties were a formidable undertaking. The two great rivals, Douglas and Lockheed, both resided within its jurisdiction. Head of the FAA’s Western Regional office was Arvin Basnight, a career public servant whose ability was equal to the task.
Shortly after the Windsor incident, on 13 June, Dick Sliff, Basnight’s head of aircraft engineering, contacted Douglas at Long Beach regarding the cargo door problems. The company’s attitude was less than helpful, and it was only after some agitation that information was made available. The documents examined by Sliff made interesting reading and revealed that there had been about 100 previous reports of difficulties in closing the door. Some even more damning documents, however, were not revealed. In particular, one written by Dan Applegate, an engineer of Convair, the cargo door designers, expressed extreme concern with safety aspects of the door. Had Sliff reviewed this information the story might have been different. In the normal course of events, records received by the manufacturing company from airlines are passed to the FAA for correlation in the form of Maintenance Reliability Reports. In this case, McDonnell Douglas, ever wary of adverse publicity for an aircraft in a highly competitive market, had not honoured the arrangement (although there was no legal requirement for them to do so) and the FAA was quite unaware of the troubles. Before the Windsor incident, however, the company had already attempted to take matters in hand and had issued Service Bulletins to the four airlines operating DC-10s (American, Continental, National, and United) recommending rewiring of the door’s electrical actuators. It was hoped that by increasing the power of the actuators with the use of heavier gauge wire, the latches would be driven more fully home. Rewiring DC-10 cargo doors was still in progress at the time of the incident but Captain McCormick’s aircraft had not been modified.
To those involved at the FAA office the proposed remedy did not seem adequate. If the DC-10 was to continue flying over the busy summer which lay ahead some interim measure was required until a more effective solution could be devised. The problem was that the actuators were not driving the latch linkage to the over-centre position (see diagram) and it was frequently necessary to shut the door manually using a hand crank. With the linkages correctly placed, closure of the outside door handle slid a lock pin in place to secure the system. With over-centre not achieved a restraining flange prevented engagement of the locking pin thus interrupting closure of the door handle. Before Flight 96 had departed from Detroit, the door had been only partially secured and the locking pin had jammed on the restraining flange. The ramp service agent had then forced the handle home with his knee, distorting the locking pin rods in the process. This had given the false impression that the door had not only closed but had locked, in spite of the fact that the locking pins were not in position. In flight, as the interior pressurised, the enormous force on the latches had been transmitted via the incorrectly positioned latching mechanism to the actuator bolts which had sheared under the strain with the inevitable results.
One simple solution seemed to be provision of a one-inch diameter peep-hole of toughened glass through which a locking pin could be checked in position. It would, of course, require a certain amount of dedication from the individual closing the door to check properly the pin’s position, especially on a cold, wet and windy night. It was not the complete answer, but something immediate was necessary to help prevent a recurrence of the incident over Windsor. The peep-hole requirement would also be backed by the full force of federal law to ensure airline’s compliance. The documents employed to enforce such an order is known as the Airworthiness Directive (AD) and is issued only in matters concerning safety. The release of the AD, unfortunately for McDonnell Douglas would make public the circumstances, and being an airworthiness requirement the costs of the modification would have to be borne by the manufacturer. Early on the morning of 16 June, the draft of the AD was telexed to the FAA headquarters in Washington for approval, but it soon became apparent that consent was not to be forthcoming. Before 0900 hours, Basnight received a telephone call from Jackson McGowan, the president of the Douglas Divison of McDonnell Douglas. The previous evening McGowan had been in conversation on the phone with Jack Shaffer, the head of the FAA in Washington DC, and was now relaying hisinformation to the Western Regional Office. After some discussion, McGowan said, he and Shaffer had agreed that ‘the corrective measures could be undertaken as a product of a gentleman’s agreement, thereby not requiring the issuance of an FAA Airworthiness Directive’. The staff at the local FAA office was somewhat taken aback by the turn of events, especially receiving the details from Douglas who had been less than helpful. Much telephoning ensued between California and Washington DC, but it soon became clear that an AD was not to be issued and that Basnight and his staff was being by-passed in discussion with the airlines. A telephone conference was arranged between Douglas, the FAA in Washington DC, and the four airlines to agree on proposals for modifying the cargo door. Agreement was reached to continue with the wiring program and to placard the doors with a warning for the ground staff not to use force in excess of 500 lb when closing the door handle. How this figure was to be gauged was not explained. It was a totally inadequate response to a very dangerous situation.
The FAA plays contradictory roles of somewhat incongruous dimensions being both the watchdog and promoter of aviation in the USA. In the America of President Nixon, during the years leading to the disgrace at Watergate, little doubt was left to the government’s attitude. The advancement of trade, commerce and industry was of prime importance, and the small affair of a defective cargo door was not going to be allowed to rock the boat, or for that matter the DC-10 McDonnell Douglas had some very powerful friends at high places. The federal organisations are usually headed by political appointees nominated by the White House, and the FAA then and now is no exception. Nixon’s intentions, however, of weakening the independence of the federal agencies by creating organisations subservient to his personal command were being effectively implemented and were beginning to take their toll. The head of the FAA, John Shaffer, a man of little commercial experience, was a political appointee. The decline suffered by the FAA under Shaffer, a basically honest man but overprotective of the industry, was far-reaching and took man years to resolve. The board of the NTSB consisted of five men all of whom were political appointees. Working under the NTSB chairman, John Reed was Ernest Weiss, a very able administrator but known to be an active Democrat. Plans were laid out to oust Weiss from his position. Head of the NTSB’s Bureau of Air Safety was another astute gentleman, Charles (Chuck) Miller, a renowned safety expert and distinguished engineer. His contribution to air safety was impressive. In the early days of the Boeing 747, turbine blade failures resulting in engine breakdowns were becoming embarrassing with some aircraft losing more than one engine on the same flight, but little was being done by the FAA to solve the problem. In October 1970, Miller, frustrated by the FAA’s inaction, vented his impatience by publicly criticising the Authority. He pressed for measures to resolve the ‘potentially catastrophic’ situation. By the end of the year the FAA relented and issued an AD against the problem engines, but only one requiring regular inspection of the turbine blades. It was several months before the blade manufacturers resolved the situation, by which time another 16 incidents had occurred. The fact that the potential disaster predicted by Miller had not materialised only served to encourage the FAA in their policy of seeking voluntary responses. By such practices it was intended to resolve aviation’s problems quietly and to maintain the good image of the industry. But would it be enough?
In March 1971, under pressure from the White House, the NTSB chairman John Reed demoted Weiss and replaced him with another political appointee, Richard Spears, whose qualifications were only the minimum for the job. It was inevitable that Spears and Miller, now a thorn deep in FAA’s side would clash, and by the summer of 1972, at the height of the DC-10 cargo door problems, the hostility between the two men broke into the open. At a time when Miller’s expertise was most needed, his attention was diverted by systematic interference in the affairs of the Bureau of Aviation Safety, and in particular Spear’s campaign against the Accident Prevention Branch. Miller was not going to be allowed publicly to criticise the FAA and get away with it*
* By April 1973 Miller found himself facing Spear’s allegations of incompetence which were lodged with the NTSB, and in August the Board sat in judgement. The verdict was improve or be fired. Miller counter-attacked by taking the matter to a higher authority and later an inquiry was convened by the Senate Commerce Committee. Before it could be resolved, however, Miller became ill with heart problems and a long stretch of periodic sick leave ensued. In December 1974 he took early retirement from the NTSB on the grounds of ill health. Fortunately he fully recovered and took up lecturing in air safety.
On 6 July 1972 the NTSB formally presented its recommendation to the FAA: the door should be rendered ‘physically impossible’ to close incorrectly, modifications should be made to the floor and vents and the floor strength should be increased. The response at the time from the FAA and McDonnell Douglas was generally to ignore the proposals. Service Bulletins (SB) 52-27 regarding the wiring improvements was followed over July and August by two more notices. SB 52-35 recommended the installation of the peep- hole with a diagram by the door frame indicating safe and unsafe positions. The blue paper on which the bulletin was raised indicated its relevance to safety. SB 52-37, routinely printed on white paper, recommended the installation of a support plate to prevent distortion of the locking pin rods, and also extension of the locking pin travel by one quarter of an inch to make jamming of the pins more obvious. Douglas’s response, on the surface, seemed to be reasonable, and the proposed modification at least adequate for the immediate future. Had the recommendations been fully implemented the door locking problems may have been solved, but the lack of urgency with which they were delivered left a lot to be desired. Without an AD to enforce the issue, the four airlines, all US, flying the 39 DC-10s in service at the time, were slow to apply the modifications. Three months later, by October 1972, only five of the total fleet had been modified. By the end of the year 18 had still not been altered, and one was still operating without a support plate
In the summer of 1972, at height of the DC-10 cargo door controversy, Lockheed with its Tristar, and McDonnell Douglas with its two versions of the DC-10 (Series 10 and Series 30), were each looking overseas for further orders. Both were simultaneously wooing Turkish Airlines who seemed a good prospect for sales, and the outcome could influence deals with other airlines. In the end McDonnell Douglas won the order and three DC-10-10 aircraft were duly delivered to Turkey in December 1972. In the meantime Nixon’s first full term of office came to an end and he was successfully re-elected. In the course of such events, tradition demands that all senior political appointees offer their resignation and the head of the FAA, John Shaffer, acted accordingly. No one was more surprised than Shaffer when his resignation was accepted! The following year he was replaced by Alexander Butterfield who had been formerly involved with internal security at the White House. Before Butterfield’s arrival, however, and perhaps influenced by Shaffer’s imminent departure, the FAA began reviewing their attitude regarding possible floor damage on wide-bodied jets as a result of explosive decompression. The Dutch equivalent of the FAA, the RLD, had been involved since the Royal Dutch Airlines (KLM), decision to buy the DC-10 and had been expressing concern for some time. In September 1972, RLD representatives met with the FAA and McDonnell Douglas to discuss the issue. Douglas was quick to point out that the DC-10 floor strength met the FAA requirements, but the Dutch countered, in the light of the Windsor incident the regulations must be considered inadequate. This view was supported by the NTSB research. The outcome of the talks was inconclusive with the FAA supporting McDonnell Douglas and agreeing to differ with RLD.
By February 1973 the FAA was beginning to admit the error of its ways and was now urging the big three aircraft manufacturers to consider strengthening their floors, increasing venting, and re-routing essential control lines away from the floor. Both Boeing and Lockheed, perhaps understandably, were indignant at being drawn into the argument surrounding another maker’s aircraft. McDonnell Doulas insisted that present regulations were satisfactory. In June 1973 the FAA asked its regional officers to obtain technical details about big jet floors, but with surprising unwillingness on the part of the Western Regional Office, it was not until February 1974 that the request was followed through. The reaction of all three manufacturers was predictable as each baulked at the thought of the expense involved in a detailed study. The Dutch had by now reluctantly certificated the DC-10, and the FAA’s rather belated attempts at action were not well received. On 25 February 1974, McDonnell Douglas replied stating that if the FAA insisted on a study then the government should bear the cost.
Less than a week later, on 2 March, the English and French rugby football teams faced each other in the Parc de Prince in Paris for what proved to be a fast and hard fought game. The result was a 12-12 draw. The estimated 30,000 English fans in attendance swelled the visitors to the city and placed an extra strain on the transport facilities between Paris and London. To make matters worse British European Airways (BEA) ground engineers at Heathrow had called a short strike in support of a pay claim against the company and all their European services were grounded. On the day following the rugby match, Sunday 3 March, the chaotic situation at Paris’s Orly Airport was compounded by other travellers arriving from elsewhere in Europe, placing themselves one step nearer home in an attempt to return to London during the weekend. The result was a gigantic headache for BEA staff at Orly as they tried desperately to find seats on other airlines to get their passengers to London.
Turkish Airlines (Turk Havana Yollari -THY) Flight 981 from Turkey to Britain that Sunday was an ideal contender for the beleaguered BEA staff. The aircraft was a DC-10 (Series 10) of 345-seat capacity, routing Ankara-Istanbul-Paris-London, and in excess of 200 seats were calculated as being available on the final Paris-London sector.
The Turkish DC-10 registration TC-JAV, landed on schedule at just after 1100 hours local time (1002 GMT) with 162 passengers on board. The aircraft parked at stand A2 at the west satellite of the Orly-Sud air terminal and 50 passengers disembarked while those in transit remained on board for the one hour stop over. Scheduled departure was local mid-day. Turkish Airlines had a few staff on hand to supervise operations in Paris but much of the ramp area work was subcontracted to Samor Co. Aircraft loading was one of Samor’s responsibilities and its personnel were instructed in DC-10 cargo door closing techniques. After switching on the power, a button in a recessed control panel adjacent to the door frame was to be pressed to power the actuators. With the door shut the operator was to continue depressing the button for a further 10 sec to ensure correct positioning of the latches.
The external handle was then to be placed flush with the door to engage the locking pins and to close a small vent door (see diagram). This was stated as being the indication that the door was safe. A warning was also issued on the use of any force in closing the door handle. A final check of the locking pin position through the peep-hole was not part of the duties of Samor’s staff but was the responsibility of Turkish Airlines. The resident THY ground engineer at Paris, Osman Zeytin, was in Istanbul on a course, and in his absence Flight 981’s transit was supervised by another ground engineer, Engin Ucok, who was to continue with the flight to London.
To facilitate procedures in Paris, THY had distributed the baggage and mail conveniently throughout the cargo holds. All the load intended for London had been placed in the forward cargo compartment which was not opened. Passengers boarding at Orly had their bags placed in the central compartment, while the aft cargo compartment contained only baggage and mail destined for Paris. It was completely emptied. Nothing was loaded in the rear hold and the door was closed at 1035 GMT by one of Samor’s staff, Mahommed Mahmoudi. He was a 39- year old Algerian expatriate who could speak two languages fluently, Arabic and French, and could read and write in them both. He could not read English, however, so could not understand the placards by the door indicating safe and unsafe positions of the locking pins.He had on a number of occasions seen Zeytin, the THY resident ground engineer, place his eye at the peep-hole but he was not aware of its function.
Mahmoudi had correctly followed the door closing procedure and nothing had given him cause for concern.THY had experienced numerous difficulties with the opening and closing of TC-JAV’s rear cargo door and a number of times it had to be closed using the hand crank, but on this occasion it shut without any undue effort. The final check of the safe condition of the door should have been conducted by one of the THY staff, but neither Zeytin’s replacement, the ground engineer Ucok, nor the DC-10’s flight engineer, Erhan Ozer, checked the peep-hole. Had they done so they would have seen that the locking pins were not in place. The latches had not been powered home fully and the latch linkage had not been driven to the safe over-centre position.
As in the Windsor incident all indications were that the door was closed and locked. History was about to repeat itself. But in the Windsor incident the handle had been forced in position with a knee while Mahmoudi in Paris had placed the handle flush with ease. In fact, it had been almost too easy.
In the Douglas factory at Long Beach, THY’s DC-10, TC-JAV, was designated the code Ship 29 during its construction. It was completed in the summer of 1972 and delivered to THY at the end of the year. The paper work for Ship 29 clearly stated that the current Service Bulletins had been implemented but, because of an oversight, the requirements of SB 52-37(extension of the locking pin travel and fitting of the locking pin rod support plate) had not been done. After delivery to THY, an adjustment to the locking pin travel was made but the aircraft was still awaiting fitment of the support plate. By a gross blunder, however, alteration of the locking pin travel was incorrectly applied. Instead of the travel being extended to seat the pin properly when locking the door, or to make obvious the jamming of the door handle with improper setting of the latch linkage, the locking pin travel was actually decreased. Even in the locked position the pins were hardly effective (see diagram).
McDonnell Douglas had calculated that with the locking pin travel extended as required, a force of 21 5 lb was needed to close the handle with the latch linkage incorrectly set, and with the support plate also fitted a force of 430 lb (beyond human strength) was necessary. With the locking pin travel decreased as on TC-JAV, a force of only 13 lb was required to place the door handle flush with the door still unlocked. The incorrect adjustment also affected the lock limit warning switch which illuminated a ‘door open’ light on the flight deck to alert the crew of incorrect positioning of the lock pins. The mis-rigging resulted in the lock limit micro switch failing to extinguish the flight deck ‘door open’ warning light, even when the door was locked. At some stage the lock limit switch striker had been extended by the addition of extra shims to permit a more positive contact (see diagram). The result of the tampering was that the flight deck ‘door open’ light now extinguished even when the door was still unlocked.
To all intent and purposes, therefore, TC-JAV’s rear cargo door was closed and locked. Only a visual inspection through the one-inch peep-hole would have indicated the true unsafe condition, and in absence of the resident ground engineer that was overlooked.
From the chaos of the terminal building passengers began to board Flight 981 and take their seats. In all 216 people were boarded to join those waiting on the aircraft giving a passenger figure of 334. In the confusion 10 or so seats remained unallocated. The crew of 11 (three flight crew and eight cabin staff) plus the ground engineer travelling on board, brought the grand total to 346. The boarding of so many extra passengers in such difficult circumstances resulted in the inevitable delay, and it was not until 1111 hours GMT that THY first contacted Orly ‘preflight’ on 120.5MHz for departure instructions. Flight 981 was assigned departure route 18 from runway 08–routing via Tournan intersection, Coulommiers and Montdidier, with an initial climb to flight level 40- and was instructed to select radar transponder code 2355 on take-off. Departure route 18 tracked aircraft east and then north to avoid overflying Paris. THY then changed to ‘ground frequency’ 121.7 MHz at 1114 hours and after completion of boarding and engine start was cleared at 1124 hours to taxi to runway 08. The conditions were fine with a light wind, some patchy cloud, a temperature of 6 C and good visibility. On the flight deck Captain Nejat Berkoz and his crew of F/O Oral Ulusman and F/E Erhan Ozer completed the before take-off checks and, approaching the runway, changed to ‘tower’ on 118.7 MHz. Flight 981 was cleared to line up and take- off, and at 1130:30 the DC-10 weighing 163 tonnes lifted off from the runway. One and a half minutes later ‘departure’ was contacted on 127.75 MHz and clearance was received for further climb to flight level 60. The after take- off check was completed and the seat belt sign switched off, but most passengers preferred to remain seated with their lap straps fastened. The autopilot was engaged. TC-JAV reported level at 60 and was instructed to contact Paris (North) on 131.35 MHz. On contact with the area controller, Flight 981 was cleared at 1136 hours for further climb to flight level 230 and instructed to turn left to Montdidier. Five routine communications passed between Flight 981 and Paris (North) in the course of the turn and at 1138 hours the DC-10 stabilised on a heading of 346 degrees, climbing through flight level 90 at a speed of 300 knots. The aircraft interior continued to pressurize as the DC-10 climbed into the rarefied atmosphere, placing a force of almost five tons on the unlocked rear cargo door.
At 1139:56 Flight 981 passed 11,500 feet climbing over the village of Saint-Pathus at a rate of climb of 2,200 feet/ minute and a speed of 300 knots. Suddenly the door latching mechanism could no longer withstand the strain and the door burst open and tore from the side of the fuselage. The heavily laden rear section of the cabin floor collapsed completely with the force of the explosive decompression and the last two rows of triple seat units above the door, with six passengers and parts of the aircraft, were ejected. The cabin fogged and dust swirled in the rush of air. On the flight deck the crew were taken completely by surprise. The throttles snapped closed and the autopilot disconnected. Almost immediately the DC-10 banked to the left and the nose pitched down violently.
‘Oops. aw, aw,’ someone exclaimed.
The co-pilot grabbed the controls as Flight 981 dived onwards the ground and the cabin pressurisation warning horn sounded.
1140:05, Captain Berkoz: ‘What happened?’
F/O Ulusman: ‘The fuselage has burst.’
1140:07, Captain Berkoz: ‘Are you sure?’
The control cables and hydraulic lines running rearwards below the floor were torn from their tracks, or jammed, and elevator and stabiliser controls were lost. The rudder seized at an angle of 10 degrees to the left. There was now no means by which the crew, in spite of their efforts, could regain sufficient control of the aircraft. As the nose dropped the speed began to build up.
1140:12, Captain Berkoz: ‘Bring it up, pull her nose up.’
At this moment, 1140:13, a radio transmission was received from Flight 981 either by chance or intention, by the radar controller at Paris(North). He could hear a heavy background noise with Turkish being spoken and the noise of the cabin pressurisation warning horn sounding. As the crew fought to regain control the transmission continued.
F/O Ulusman:’I can’t bring it up- she doesn’t respond.’
The aircraft continued to accelerate towards the ground. By 1140:18 the nose was pitched down 20 degrees with the speed increasing through 362 knots.
1140:19, F/E Ozer: ‘Nothing is left.’
At 1140:21, with the DC-10 descending into the denser atmosphere, the pressurisation warning horn ceased. Passing 7,200 feet the speed increased to 400 knots, and the aircraft continued banking in a left turn as it raced earthwards.
F/O Ulusman: ‘Seven thousand feet.’
A second or so later the over speed warning sounded as the speed edged beyond the ‘never exceed’ speed. At the same moment the Paris (North) controller noticed TC-JAV’s flight label disappear from the secondary radar scope. On the primary radar screen a thin silver of light representing the ejected parts detached itself from the echo and remained stationary while the DC-10’s trace curved to the west.
1140:28, Captain Berkoz: ‘Hydraulics?’
F/O Ulusman: ‘We have lost it . . . Oops, oops.’
At 1141:31 the nose down pitch began progressively to decrease and the speed stabilised at 430 knots. The radio transmission being received on the ground abruptly ended at 1140:41 as the controller continued to monitor the DC-10’s progress.
1140:50, Captain Berkoz: ‘It looks like we are going to hit the ground.’
1140:52, Captain Berkoz: ‘Speed.’
The overspend warning continued to sound throughout, but fainter than at the beginning.
1140:57, Captain Berkoz: ‘Oops.’
The DC-10’s angle of descent stabilised at a shallow angle but the machine continued at great speed towards the ground. TC-JAV was beyond help. At 1141:04 and 1141:06 further short radio transmissions were received by the ground controller from the DC-10. These were the last to be heard. Flight 981 struck the trees of the Ermenonville forest at 1141:08 at a speed of 430 knots in almost level flight, but with the wings banked 17 degrees to the left. At 1141:31 the DC-10 impacted in a rugged valley at a place known as Bosquet De Dammartin, 37 kilometres (20 miles) northeast of Paris, only 77seconds after the door burst open. The aircraft ‘cut a swath through the forest some 700 metres long by 100 metres wide’, and literally disintegrated as it struck. All 346 on board perished. With the force of the impact TC-JAV exploded into a million tiny pieces and scattered over a wide area. There was virtually no fire. Nothing had a chance to burn.
In the control centre the radar controller watched helplessly as the echo disappeared from his screen. Many times he tried in vain to contact Flight 981 and with no reply he raised the alarm. Soon a large scale rescue operation was put into action and the first rescuers arrived quickly on the scene at 1215 hours. There was little they could do except start to clean up the mess. Experts who witnessed the devastation stated that they had never seen an aircraft disintegrate so completely over such a wide area.
Within 22 minutes of the disaster, even before the firs rescuers arrived at the site, the BBC broadcast news of the crash. It was the accident the world had feared: the first involving a fully laden wide body jet since their introduction into service only four years earlier. The worst previous had killed 176 when a Boeing 707 crashed on landing at Kano, Nigeria, returning with Muslims from a pilgrimage to Mecca in January 1973. The THY DC-10 catastrophe, with all 346 on board killed, was at the time the worst disaster in civil aviation’s history.
Kind courtesy of: Air Disasters by Stanley Stewart Published by Ian Allen England 1986 & Wikipedia.org