The Repository Root Cause

Aloha 243: The Inspection Order Named One Seam of Six

October 14, 2026 · Runtime 11:06 · Watch on YouTube

In November 1987 a federal order told airlines where to inspect the roof of an old Boeing 737 for fatigue cracks. It named one seam; the manufacturer's bulletin it came from named six. Five months later, eighteen feet of an Aloha Airlines cabin roof tore away along a seam that was on the bulletin's list and not on the order's.

Chapters

Transcript

The order named one seam 0:00

In November of 1987, a federal order told airlines where to inspect the roof of an old Boeing 737 for fatigue cracks. It named one seam. The manufacturer's bulletin that the order was taken from named six. Five months later, an airplane that had been through that order's inspection leveled off at twenty-four thousand feet. About eighteen feet of its cabin roof came away. It had opened along a seam that was on the bulletin's list, and not on the order's.

Flight 243, 24,000 feet 0:31

Aloha Airlines Flight 243 left Hilo for Honolulu at 1:25 in the afternoon on April 28th, 1988. Ninety-five people were on board. It was the airplane's ninth flight of the day, and the first eight had been uneventful. At 1:46, as it leveled at twenty-four thousand feet, both pilots heard a loud clap and then wind. The captain looked back and saw, in his words, blue sky where the first-class ceiling had been. One flight attendant was lost through the opening. The pilots descended, steered for the island of Maui and landed the airplane at Kahului about thirteen minutes later. The other ninety-four people landed with it, eight of them seriously hurt.

A crack seen from the jet bridge 1:16

Here is the detail the National Transportation Safety Board put on page five of its report. A passenger said that as she boarded in Hilo, she saw a crack in the fuselage. It ran along the upper row of rivets of a seam on the left side of the roof, halfway between the door and the edge of the jet bridge. She said nothing. Nobody had asked her to look, and by the airline's accepted procedures the crew was not required to walk around the airplane between flights. The report records it without blame. What matters is the row of rivets, and what was supposed to hold it together.

Glue in the seam 1:52

Boeing built the first 291 of these airplanes with a skin that was held together partly by glue. The skin was only 0.036 of an inch thick. Where two panels overlapped by about three inches, an epoxy-soaked cloth was bonded between them, and three rows of rivets crossed the joint. The idea was that the glue would carry the load of the pressurized cabin, so the skin could be thin and light. The glue was so reactive that it had to be stored on dry ice until the moment it was used. And when the bond came out poor, it came out poor at random: some areas of those seams never stuck at all. Boeing stopped building the joint that way in 1972. The airplanes already built stayed in service.

The knife edge in every rivet hole 2:37

When the glue lets go, the rivets have to do its job. And the rivets of that skin sit in countersunk holes. The skin is so thin that the countersink goes all the way through the outer sheet, leaving a sharp edge, which the report calls a knife edge, at the bottom of every hole. Every flight pressurizes the cabin and stretches the skin, and the stress concentrates on that edge. The pressure pulls on the skin around the fuselage twice as hard as it pulls along it. So the cracks start at the rivet holes and grow along the row, one hole after another.

The safety net: tear straps 3:11

Boeing had designed a safety net for exactly this kind of crack. The fuselage had to survive a forty-inch tear without breaking up. Every ten inches, strips of metal called tear straps ran around the skin. If a crack ran into a strap, the strap would turn it sideways. The skin would peel open like a flap, and the cabin would lose its pressure in a controlled way. Boeing proved it in certification with two fifteen-inch blades driven through a pressurized test fuselage. But the straps were glued on, too. A disbonded strap cannot stiffen anything, and the crack runs straight past it.

What the certification test never tested 3:49

There was a second thing the design had not considered. The 737 was designed for an economic life of seventy-five thousand flights. Boeing tested the fuselage design on half a section of a body, pressurized a hundred and fifty thousand times. No fatigue cracks developed, and no joint came unstuck. But that test said nothing about airplanes whose glue had failed, because the glue in the test never had. The fatigue evaluation gave no consideration, the report says, to the possibility of disbonding. The airplane that came to Maui had been pressurized eighty-nine thousand, six hundred and eighty times.

Sixteen years in print 4:28

The problem was in print, and had been for sixteen years. In July of 1972, Boeing issued a service bulletin about corrosion in these glued seams. In February of 1974 it was revised, with disbonding found on thirty airplanes. It said that prolonged operation with large areas of disbonding will eventually result in fatigue cracking. None of it was mandatory. In 1984 Aloha itself reported a seven-and-a-half-inch crack along a seam of another of its 737s. In 1987 an airline reported multiple cracks on three airplanes, and Boeing made the bulletin an Alert.

Twelve joints on the list, two in the law 5:10

The Alert bulletin listed the seams to inspect: stringers four, ten, fourteen, nineteen, twenty and twenty-four, left and right. Twelve lap joints. The cracks that had triggered the Alert had been found on the seams at stringers four, ten and fourteen. Then the Federal Aviation Administration wrote its Airworthiness Directive, effective November 2nd, 1987. An airworthiness directive is the document that makes an inspection mandatory. It required the inspection at stringer four, left and right. Two lap joints.

Why only stringer four 5:44

Why only stringer four? An FAA employee testified that the decision was based on statistical information, and on the scope of the work required. The Safety Board was blunt: the intent of the Alert bulletin was altered significantly by the FAA, to reduce the scope of the inspection. It said that had a full inspection of all the lap joints been required, the likelihood of this accident may have been reduced. And it did not blame the order alone. Aloha had the full bulletin, and its own seven-and-a-half-inch crack on the books, and it inspected what the order said.

The inspection that did happen 6:21

That inspection did happen, over three days in November of 1987. Two inspectors examined the seam at stringer four and found cracks. Two sheet metal repairs were made, and the airplane flew on. If cracks turned up, the directive required a second check, with an electrical probe, along the entire row of rivets. The maintenance log shows the visual inspection, and no record that the probe was used. Afterwards, the Board measured the five largest cracks it sampled in that seam: between 0.11 and 0.15 inches. Working backward, they had been about 0.09 to 0.13 when the inspectors were there. The probe, tested after the accident, reliably found anything above 0.08.

What the work looked like 7:11

The Board said the cracks should have been found, and then it asked what the work looked like. The directive called for a close look at about thirteen hundred rivets in one panel. The inspectors climbed scaffolding, carried a bright light and, at times, roped themselves to the hangar rafters. Sometimes under layers of paint, they scanned one rivet after another, and most were fine. A human factors expert told the hearing that in this kind of search people raise the size of the defect they will report. And the inspector who found the cracks had received two hours of formal training in non-destructive inspection, two months earlier.

A maintenance clock set for the wrong airplane 7:49

The airline's maintenance program also had a clock problem. Boeing had planned for about one and a half flights for every hour an airplane spent in the air. On Aloha's island hops, the airplanes made about three. Fatigue counts pressurizations, not hours, so these airplanes were aging twice as fast as the schedule assumed. The structural check was split into fifty-two small packages, done on overnight stops, and there was no spare airplane. The airline had no engineering department to read bulletins. The highest-cycle 737 in the world was one of theirs, at ninety thousand flights.

How the roof opened 8:28

That is how the roof opened. The Board concluded that along the seam at stringer ten, on the left side, there was enough fatigue cracking, loose straps, or both, to switch the safety net off. Small cracks ran from rivet to rivet along the top row, and, in the Board's finding, they linked up quickly. A skin without sound straps does not flap. It peels. Boeing's own fatigue test fuselage had shown the other ending: its cracks joined into a forty-inch tear, and the skin flapped and let the pressure out. At Maui, the net had stopped working.

The other airplanes 9:03

The airplanes that were inspected the same way tell the same story. Two other high-cycle 737s of the airline had also been inspected under the directive. After the accident, both were found to be beyond economical repair. A third, stripped of its paint six months later, showed fatigue cracks at many rivets along the seam at stringer fourteen. That seam was on Boeing's list too, and not on the directive. In the FAA's own database since 1983, among 1,352 reports on the 737 fuselage, ten were cracks at or near lap joints. The signal was there, and it was ten lines in thirteen hundred.

The verdict, and a dissent 9:46

The Safety Board named the airline's maintenance program as the probable cause. It listed as contributing the failure of the FAA to require the inspection of all the lap joints, and the lack of any permanent fix for the glued joint. One member dissented. He wrote that the best engineers in the industry had reviewed this airline's records, and that nobody predicted the hull was about to rupture. If something failed, he said, it was a system. The report's own findings say that a similar lack of attention to lap joints was an industry-wide deficiency.

A rule has an edge 10:20

Look at the sizes of the boxes. Boeing's bulletin drew a box around twelve joints, and the order drew one around two. The inspection looked at about thirteen hundred rivets in that box, and cracks big enough to be found were still there afterwards. Under it all, the one safety net was glued as well, and its straps were coming loose too. A rule is a kind of knowledge, but it has an edge, and after the edge nobody is looking. Next case is another airliner, the following year. United Flight 232, July 1989: one engine in the tail broke apart, and the three hydraulic systems that were meant to back each other up were all lost. Three independent systems, and one disk that reached them all.

Description and sources

Aloha Airlines Flight 243, Hilo to Honolulu, 28 April 1988, 95 on board. As the Boeing 737-200 leveled at 24,000 feet, about eighteen feet of the upper cabin skin separated from the airplane. One flight attendant was lost; the crew landed at Kahului, Maui, about thirteen minutes later.

The early 737 fuselage was glued as well as riveted. When the bond failed, the load went to countersunk rivets whose knife edges started fatigue cracks, and the design's safety net -- tear straps meant to turn a running crack into a controlled flap -- depended on bonds that had come unstuck too. Boeing's service bulletin had warned of disbonding since 1972, and was upgraded to an Alert in August 1987 after cracks were found on three airplanes at stringers 4, 10 and 14.

The FAA's directive of November 1987 required the inspection at stringer 4 only: two lap joints of the twelve the Alert listed. The NTSB found the intent of the Alert was significantly altered to reduce its scope. It also found that cracks large enough to be detected were still in the stringer 4 seam after the November inspection, and that the airline's program, running about three flights per flight hour against Boeing's assumed one and a half, had not caught up. The fuselage failed along stringer 10 on the left side. One Board member dissented, calling it a system failure.

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

National Transportation Safety Board, Aircraft Accident Report NTSB/AAR-89/03: Aloha Airlines, Flight 243, Boeing 737-200, N73711, near Maui, Hawaii, April 28, 1988 -- read in full (sections 1.1, 1.6.2-1.6.3.4, 1.16.3, 1.17.1-1.17.4, 2.3-2.5.1, Conclusions, dissent): https://www.ntsb.gov/investigations/AccidentReports/Reports/AAR8903.pdf
Aloha Airlines Flight 243: https://en.wikipedia.org/wiki/Aloha_Airlines_Flight_243

ARCHIVE PHOTOGRAPHS — the real structure, with title, author, date, archive and licence. All public-domain.

Fig. 1 — NTSB — 1989 | NTSB/AAR-89/03 | Public domain
Fig. 3 — NTSB — 1989 | NTSB/AAR-89/03 | Public domain
Fig. 6 — NTSB — 1989 | NTSB/AAR-89/03 | Public domain
Fig. 5 — NTSB — 1989 | NTSB/AAR-89/03 | Public domain

The 9 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 — the photographs above are the archive, the drawings are ours, not photographs of the hardware, and no person is depicted in any of them.

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

#aviation #engineering #rootcause