The DC-10 Memo That Predicted the Crash and Vanished

Chapters

1. 0:00 June 27th, 1972
2. 0:31 The story everyone tells
3. 0:55 A contract with a gag clause
4. 1:24 Hydraulic, swapped for electric
5. 1:56 How Boeing and Lockheed built it
6. 2:19 Nine ways to crash, found in 1969
7. 2:54 Long Beach, May 1970
8. 3:32 Certified by its own employees
9. 4:02 Windsor, June 12th, 1972
10. 4:39 The memo
11. 5:12 Not only about safety
12. 5:30 "Nothing new to Douglas"
13. 5:59 A deal struck the night before
14. 6:32 Cut out of its own decision
15. 6:50 Courting Turkish Airlines
16. 7:10 The fix that existed
17. 7:33 Three signatures, no memory of the work
18. 7:46 March 3rd, 1974
19. 8:27 "Human failure"
20. 9:05 What was actually missing
21. 9:39 Twelve cases, a new shape
22. 9:59 Next: same airplane, different mechanism

Transcript

June 27th, 1972 0:00

June 27th, 1972. An engineer named Dan Applegate sits down and writes a memo to his own bosses. He doesn't send it to the government. He doesn't send it to the press. He predicts, in writing, that a passenger jet's cargo door will blow open in flight — and that it will usually cost the plane. Six hundred fourteen days later, it happens exactly as he described. Three hundred forty-six people die.

The story everyone tells 0:31

Here's the story everyone tells about that memo: a lone whistleblower, ignored by a cold company. That memo is real, and that sentence in it is real. But Applegate wasn't the first person at his company to know this. The reason his warning went nowhere wasn't disbelief. On paper, telling anyone else simply wasn't his company's job.

A contract with a gag clause 0:55

Convair didn't design the DC-10. It built the fuselage — cargo doors included — as a subcontractor to McDonnell Douglas. The contract specified that door down to the latch and the actuator, plus a standing order to save weight wherever possible. Convair's engineers had no vote in any of it. And the contract went further: it barred Convair from contacting the FAA about the airplane at all.

Hydraulic, swapped for electric 1:24

The original latch design used a hydraulic actuator. If the latches weren't fully locked, cabin pressure would just push the mechanism back open — a soft failure, on the ground. At American Airlines' request, McDonnell Douglas switched to a lighter electric one, to cut weight and maintenance cost. An electric actuator holds its position with a ratchet. It doesn't fail soft. It fails once, under pressure high enough to break that ratchet in the air.

How Boeing and Lockheed built it 1:56

Boeing and Lockheed built their cargo doors differently: the small vent door could only close after the lock pins were fully engaged. On the DC-10, the vent door and the lock pins were driven off the same handle — closing one didn't guarantee the other had happened. Forcing that handle could make the door look sealed without ever locking it.

Nine ways to crash, found in 1969 2:19

In 1969, Douglas asked Convair for a formal analysis: every way that door could fail. Convair's engineers found nine sequences that could bring down the airplane. One was almost exactly what would happen in 1974. Douglas had instructed them, in writing, not to trust cockpit warning lights — and even less, the alertness of ground crews. That report never reached the FAA. Convair's contract didn't allow it, and Douglas never mentioned it in its own filings.

Long Beach, May 1970 2:54

May 1970, Long Beach. The first DC-10 ever built sits on a ground stand, being pressurized to test its air conditioning. The rear cargo door blows open. The cabin floor collapses into the hold, severing the same lines that steer the tail. Nobody dies — the fuselage never left the ground. Investigators blame a mechanic who didn't latch the door correctly. The fix that follows is small: a peephole window, so a ground crew can see the lock pins. Nothing about the floor changes.

Certified by its own employees 3:32

A year later, the DC-10 goes through FAA certification. Three out of every four inspections are carried out by Douglas's own employees, deputized for the day as FAA representatives. It's a common arrangement — the agency doesn't have the staff to inspect a new airliner alone, so it borrows the manufacturer's own engineers. The DC-10 is certified for passenger service on July 29th, 1971.

Windsor, June 12th, 1972 4:02

June 12th, 1972. American Airlines Flight 96 climbs out of Detroit. A ground agent had struggled to latch the rear cargo door, and forced the handle shut with his knee. At eleven thousand seven hundred fifty feet, that door blows open. The cabin floor partly collapses, snapping control cables to the tail. The captain had trained for hours, on his own initiative, for exactly this failure. He flies the crippled jet back to Detroit on the wing engines alone. Everyone survives.

The memo 4:39

Two weeks after Windsor, Dan Applegate sits down to write. He was Convair's chief product engineer — in charge of that door. His memo says the Windsor failure was inevitable, and spells out exactly what comes next if nothing changes. He writes that in the next twenty years, DC-10 cargo doors will keep coming open — and that it will usually cost the airplane. That sentence would prove exactly right, six hundred fourteen days later, over a forest outside Paris.

Not only about safety 5:12

But read the memo's own text, and safety isn't the whole story. Applegate writes that Douglas was trying to make the door's failure into Convair's mistake to pay for. Not a design choice Convair ever had a vote in. The warning and the invoice were the same document.

"Nothing new to Douglas" 5:30

Convair's management reviewed the memo. Their conclusion, recorded later in testimony, was almost reasonable. Everything in it was already well known to Douglas — nothing new. Douglas's own engineers had lived with that door for years. Raising it again risked nothing but Convair's standing in an ongoing contract dispute. The memo stayed inside the company. It was never sent.

A deal struck the night before 5:59

The NTSB investigated Windsor and reached its own conclusion: the doors needed a mandatory fix, backed by a public Airworthiness Directive. The FAA's regional office had already drafted one. The night before it was due to be issued, it was shelved. Douglas's president struck an informal deal with the FAA's administrator instead. Douglas would fix the doors itself, through routine service bulletins. No public directive. No grounded fleet.

Cut out of its own decision 6:32

The FAA's own regional office — the one that actually oversaw Douglas's factory — had helped draft that Airworthiness Directive. It learned about the private deal only after it was struck, and asked for a seat at the table. Douglas held that meeting without them.

Courting Turkish Airlines 6:50

That summer, Douglas was mid-tour on a sales campaign, courting airlines worldwide to buy the DC-10 — Turkish Airlines among them. A federal directive naming a design defect would have made international news at the worst possible moment. A quiet service bulletin would not.

The fix that existed 7:10

The promised fix was two small parts: a peephole to check the lock pins, and a metal support plate to stop the linkage from bending under force. Cheap. Simple. Correct — if it actually reached every airplane. The support plate never reached the DC-10 that would become Turkish Airlines Flight 981.

Three signatures, no memory of the work 7:33

Three separate quality inspectors had stamped that aircraft's paperwork, certifying the support plate installed. When investigators looked afterward, none of the three could recall actually doing the work.

March 3rd, 1974 7:46

March 3rd, 1974. Turkish Airlines Flight 981 climbs out of Paris, bound for London. At about eleven thousand feet, the same door fails the same way it failed at Windsor — six hundred fourteen days later. The floor collapses onto the control cables to the tail. All three independent sets sever together — the FAA minimum, one fewer set than competing airliners carried. The airplane pitches down and cannot recover. Seventy-seven seconds later, it hits a forest outside Paris. Three hundred forty-six people die.

"Human failure" 8:27

Within days, Douglas's president told the press the company had always gone beyond federal safety rules — and pointed at the ground crew. The airline's own president called it, in his words, human failure, citing a baggage handler who couldn't read an English warning label. In the hearings that followed, Douglas also tried blaming Convair, then the FAA, then Turkish Airlines itself. Every lawsuit eventually settled through the same insurance syndicate, Lloyd's of London, whoever a court might have blamed. No court ever ruled on which company was legally responsible.

What was actually missing 9:05

The common version of this story is that one warning got buried. The documents tell a stranger one. By 1974, at least three separate efforts had already mapped this exact failure. Convair's 1969 study, the 1970 ground test, and Applegate's 1972 memo. Nobody in that chain lacked information. At every link, telling someone else cost more than staying quiet. And the people best placed to speak had signed a contract that told them not to.

Twelve cases, a new shape 9:39

Twelve cases into this channel, and this is the first one driven by a contract clause instead of a control room. The earlier warnings on this channel died from distance, or from too many hands passing them along. This one died because speaking up had a price tag attached to it, and staying quiet didn't.

Next: same airplane, different mechanism 9:59

Five years after this crash, the same airplane would fail again — a different mechanism entirely. May 25th, 1979. American Airlines Flight 191, taking off from Chicago. A maintenance shortcut on the engine mounts — one the manufacturer itself had recommended — cracks a fitting nobody is watching.

Description and sources

On June 27th, 1972, Convair's chief product engineer, Dan Applegate, wrote a memo to his own management predicting that DC-10 cargo doors would keep coming open in flight for the next twenty years, and that it would usually cost the airplane. Six hundred fourteen days later, on March 3rd, 1974, Turkish Airlines Flight 981 crashed outside Paris the exact way he described. All 346 people aboard died.

Convair didn't design that door — it built the fuselage under contract to McDonnell Douglas, which specified the latch and actuator down to the last part, and barred Convair from contacting the FAA at all. In 1969, Douglas had asked Convair for a formal failure analysis; Convair's engineers found nine sequences that could bring the plane down, one nearly identical to 1974. That report never reached a regulator. In May 1970, the first DC-10 ever built proved it on a ground stand: the door blew open during a pressure test, collapsing the cabin floor onto the same control lines that would fail twice more.

Applegate's memo came two weeks after a near-disaster — American Airlines Flight 96, whose door failed at 11,750 feet and whose captain flew it back to Detroit on the wing engines alone, saving everyone aboard. Read in full, the memo isn't only a safety warning: Applegate also wanted it on record that Convair had "no voice" in decisions Douglas was now trying to bill back to it. His own managers reviewed it and concluded, reasonably enough, that everything in it was already well known to Douglas. The memo stayed inside Convair. It was never sent.

The NTSB wanted a mandatory Airworthiness Directive. The night before one was due, Douglas's president struck a private deal with the FAA instead: quiet service bulletins, no public directive, no grounded fleet — while Douglas courted airlines worldwide, Turkish Airlines among them. The promised fix, a support plate for the door's linkage, never reached the aircraft behind Flight 981. Three inspectors had certified it installed; none could recall the work.

This wasn't a warning lost for lack of information. Three separate efforts, in three organizations, had already mapped this exact failure. At every link, speaking up cost more than staying quiet — and the party best placed to speak had signed a contract against it.

PRINT-READY, FROM THIS CHANNEL

The Failure Atlas, Vol. 01 — Tacoma Narrows · Citicorp Center · Millennium Bridge · Apollo 13 · the 2003 blackout · Hyatt Regency

https://therepository.gumroad.com/l/failure-atlas

PRIMARY SOURCES

• NTSB-AAR-73-2, American Airlines DC-10-10 N103AA near Windsor, Ontario, June 12, 1972 — read directly, page by page, from the scanned original: https://www.ntsb.gov/investigations/AccidentReports/Reports/AAR7302.pdf
• The Applegate memo, the 1970 ground test, the FAA "Gentleman's Agreement," and Flight 981 are drawn from Fielder & Birsch, The DC-10 Case (SUNY Press, 1992) and Eddy, Potter & Page, Destination Disaster (1976) — cited via a secondary source quoting both with page numbers, cross-checked independently. Not read directly from those two books or the French BEA report in this production.
• Cross-checked: en.wikipedia.org/wiki/American_Airlines_Flight_96 and en.wikipedia.org/wiki/Turkish_Airlines_Flight_981

The 26 technical plates in this video are illustrations generated for the channel by a diffusion image model, styled to match its cyanotype identity. They are diagrams of the system, not photographs of the hardware, and no person is depicted in any of them.

Root Cause investigates why engineered systems fail, using official investigation reports and primary technical literature. Sources for this episode are linked above.

The technical drawings in this video are cyanotype-style illustrations produced for the channel — diagrams, not photographs of the real hardware, documents, or wreckage. Charts and dimensioned comparisons are drawn from the sources listed above.

#engineering #failureanalysis #rootcause #aviation #dc10