Twelve minutes after leaving Tokyo on the evening of 12 August 1985, the rear pressure bulkhead of a Boeing 747 gave way. The escaping air blew off most of the vertical stabiliser and severed all four hydraulic systems. The crew were left with an aircraft that had no ailerons, no elevators, no rudder — nothing but four throttles.

They flew it for thirty-two minutes. Five hundred and twenty people died, four survived, and the reason the bulkhead failed was a repair carried out seven years earlier.

A hard landing in 1978

On 2 June 1978, on the same Tokyo–Osaka route, JA8119 bounced on landing at Itami, over-flared, and struck its tail hard on the second touchdown. Twenty-five people were injured. The aft pressure bulkhead — the dome that holds cabin pressure at the back of the fuselage — was cracked.

Boeing sent technicians to repair it. The approved procedure called for a single continuous splice plate secured with three rows of rivets. What was actually fitted was two separate splice plates laid parallel to the crack.

The distinction sounds pedantic. It was not. Cutting the plate that way negated the effectiveness of one row of rivets, reducing the joint’s resistance to fatigue cracking to roughly seventy percent of a correct repair.

Investigators later calculated that the joint would fail after about 11,000 pressurisation cycles. The aircraft flew 12,318 more flights. And because the overlapping plates concealed the defect, Japan Air Lines’ own inspections never found it.

Eighteen twenty-four

Twelve minutes after takeoff, climbing through roughly 23,900 feet over Sagami Bay, the bulkhead let go. The cabin ceiling came down near the rear lavatories. The tailcone containing the auxiliary power unit was blown clear. About five metres of the vertical fin went with it, and with the fin went every hydraulic line in the tail.

Part of that fin was later fished out of the sea.

The flight deck now had a 747 with no working flight controls at all. Ailerons, elevators, rudder and yaw damper were gone together.

Thirty-two minutes on the throttles

A detail worth getting right: Captain Masami Takahama, 49, was not the one flying. He was instructing First Officer Yutaka Sasaki, 39, who was occupying the captain’s role on a final promotion check. Takahama worked the radios. Flight Engineer Hiroshi Fukuda, 46, was the third man.

What they were fighting was a phugoid — a slow ninety-second cycle in which the nose rises, the speed decays, the nose drops, the speed builds and the nose rises again — overlaid with a Dutch roll swinging the aircraft through arcs of around fifty degrees every twelve seconds. Their only input was differential thrust between the left and right engines.

They lowered the gear to damp the phugoid, which worked and cost them what little directional authority they had. They extended flaps electrically. At no point did any of them put on an oxygen mask, and hypoxia is documented in the record.

The report is saying, in the flattened language of accident investigation, that what the crew were attempting could not be done. They kept the aircraft airborne for thirty-two minutes anyway. All three were posthumously awarded the Polaris Award in 1987.

The exchange with air traffic control has a terrible economy to it. Tokyo Control offered a heading and a radar vector to Oshima. The reply was two words: “But now uncontrol.”

At 18:46, ten minutes before impact, Takahama said what everyone in the cockpit already knew.

At 18:56:30 the aircraft struck a ridge with its right wing, then went inverted into a second ridge.

A reconstruction of the accident and the investigation that followed.

The night on the mountain

This is the part of the story that Japan has never entirely settled.

A US Air Force C-130 crew spotted the site about twenty minutes after impact, in daylight, and passed the position to Japanese authorities and to Yokota Air Base, where a helicopter was readied. According to an account published in 1995 by a USAF navigator who was there, an American search-and-rescue effort was prepared and then stood down at the request of Japanese authorities. A contemporary 1985 report in Pacific Stars and Stripes said instead that Yokota personnel were on standby and were simply never called.

A Japanese military helicopter reached the area after dark, could not land, and reported no sign of survivors. On that basis ground teams did not set out overnight. They camped 63 kilometres away and climbed in the morning.

Yumi Ochiai, an off-duty JAL flight attendant and one of the four survivors, said from her hospital bed that she had seen lights and heard rotor blades, and that she had heard other people screaming and moaning through the night, and had heard those voices fade.

Medical examiners later found bodies with injuries indicating that people had survived the crash and died of shock, exposure or injuries that were treatable. A doctor at the site told Tiempo that ten hours earlier they would have found more survivors alive.

The failure sequence and the crew’s attempt to fly on thrust alone.

What it cost, and what it changed

Japan Air Lines lost up to a third of its domestic traffic. Its president, Yasumoto Takagi, resigned. A maintenance manager, Hiroo Tominaga, died by suicide in atonement; so did Susumu Tajima, the engineer who had certified the aircraft airworthy after the 1978 tailstrike.

Technically, the accident forced a reckoning with hidden repairs and with the assumption that four hydraulic systems constituted redundancy when all four ran through the same small volume of tail structure.

Four years later, when a DC-10 lost all its hydraulics over Iowa, one of the men who helped fly it to Sioux City was a check airman named Dennis Fitch, who had studied JAL 123 and had gone into a simulator to practise flying an airliner on engine thrust alone. A hundred and eighty-four of the two hundred and ninety-six people aboard United 232 survived.

Japan Air Lines opened a Safety Promotion Center at Haneda in 2006, where part of the wreckage is displayed. Every new employee visits it, and climbs the ridge.

A concise account of the accident and the four survivors.

Sources: Aircraft Accident Investigation Commission final report on JA8119 (19 June 1987), Japan Transport Safety Board, Aviation Safety Network, Tiempo, Pacific Stars and Stripes