The Twin That Was Too Strong
The Comet jet broke up in 1954, and the usual story blames its square windows. The cracks actually began at small holes beside openings no passenger looked through, and the extra-careful test that should have caught them hid them instead.
FactUnion VerifiedEvery fact on the episode's fact sheet was checked against its sources. Wrong facts were corrected. What this meansFacts checked 18 Oct 2026
Chapters (7)
About this episode
In 1954 two de Havilland Comets, the world's first jet airliners, broke up in the air. The popular lesson is square windows. The investigators traced something smaller: fatigue cracks starting at fastener holes, one beside a little radio aerial window in the cabin roof and one beside an escape hatch. Owen and producer Felix follow a piece of roof brought up from the sea off Elba, go through the Court of Inquiry's paragraphs, and look at a test that went further than the rules asked.
That extra care is the twist. The fuselage section de Havilland fatigue-tested had first been squeezed harder than any airliner in service ever was. That very likely made the stand-in tougher than the aircraft carrying passengers. This episode is about who designs a test and what the test is able to show.
Hosted by Owen and Felix, performed by synthetic voices.
Facts and sources
Every fact used in this episode, with the sources it was checked against. Download as data
133 of 133 recorded lines matched the checked script.
Paragraph 77 of the Cohen report says it is probable the cabin disruption on G-ALYP began near the starboard aft corner of the rear ADF window, at the edge of a countersunk bolt hole where fatigue was identified.
The Court report records that three cracks were found in the reinforcing plates of the ADF windows and were approved by the Chief Inspector and Design Department, with the crack ends drilled with a 1/16 in drill (para 137).
In the Farnborough water tank, G-ALYU failed after 1,221 real flights plus 1,836 simulated cycles (3,057 in all), with the fatigue crack starting at a rivet hole near the forward port escape hatch.
BCAR of the era required a design pressure of 2P and a proof test of 1.33P, while de Havilland designed the Comet 1 cabin to 2.5P and proof-tested to 2P (16.5 psi).
Before its fatigue test, de Havilland’s forward-fuselage test section had taken 30 applications of pressure between 1P and 2P and about 2,000 applications up to ‘rather over P’, and the same section was then fatigue-cycled up to P.
Show all 20 factsShow fewer facts
De Havilland’s fatigue-tested prototype fuselage failed at 16,000 pressurization cycles.
- De Havilland DH-106 Comet 1faa.gov
- Common Comet Misconceptions and Collaborative Contribution to Safetyaerossurance.com
- The de Havilland Comet, the world's first jet airliner, had a test fuselage that survived 16,000 pressurization cycles before it cracked. In 1954 two Comets in airline service broke apart in the air,linkedin.com
Production Comet 1s, including G-ALYU, did not receive the 2P proof test, so they lacked the cold-working that inadvertently improved the test specimen’s fatigue life.
- The de Havilland Comet, the world's first jet airliner, had a test fuselage that survived 16,000 pressurization cycles before it cracked. In 1954 two Comets in airline service broke apart in the air,linkedin.com
- DeHavilland Cometglobalsecurity.org
- On the morning of January 10, 1954, BOAC Comet G-ALYP climbed out of Rome's Ciampino Airport on the final leg of a long-haul route from Singapore to London. Thirty-five people were on board.linkedin.com
De Havilland estimated the average stress at window corners at about 28,000 psi.
The Comet 1 cabin skin was 22-gauge (0.71 mm), thickened to 20-gauge (0.91 mm) around the windows.
The RAE report R&M 3248 describes the Comet I window apertures as reinforced by peripheral members bonded with Redux adhesive and also secured with countersunk-head rivets at the corners.
The Comet 1’s passenger windows were near-square, with corners that were rounded, but not as much as on modern square windows.
- Why weren’t the Comet’s overhead ADF windows roundified along with its passenger windows?aviation.stackexchange.com
- On the morning of January 10, 1954, BOAC Comet G-ALYP climbed out of Rome's Ciampino Airport on the final leg of a long-haul route from Singapore to London. Thirty-five people were on board.linkedin.com
Operation Elba Isle (20 January to 11 April 1954) used the RFA Sea Salvor and several Royal Navy ships to recover wreckage of the BOAC Comet G-ALYP, which crashed into the sea off Elba.
An Admiralty Research Laboratory team led by George MacNeice developed underwater TV cameras that were used to help locate and salvage wreckage in deep water.
At Farnborough, RAE investigators mounted the recovered fragments on a frame to rebuild the fuselage like a jigsaw.
South African Airways Flight 201 (G-ALYY), a charter operated for BOAC, broke up near Naples on 8 April 1954 and killed all 21 aboard, 14 passengers and 7 crew.
After the Naples accident, BOAC suspended all Comet flights and the aircraft’s certificate of airworthiness was withdrawn.
Attorney General Sir Lionel Heald told the inquiry that the fatigue had developed near the radio direction-finding aerial window in the cabin roof, and that up to 70 percent of ultimate stress was concentrated at the window corners.
The Science Museum Group holds a section of the upper fuselage of G-ALYP showing the fatigue cracks around its ADF openings (object co202453).
The fuselage of the last surviving short-fuselage Comet 1XB, G-APAS, left the RAF Museum Midlands at Cosford on 9 September for the South Wales Aviation Museum at St Athan.
The Comet 1A fuselage F-BGNX at the de Havilland Aircraft Museum is described by the museum as the only surviving Comet fuselage with the original square-shaped windows.
Removed during checking (9)
Paragraph 78 says a manufacturing crack of about 0.2 in at the forward port corner of the same ADF window was dealt with under the de Havilland Concession procedure by drilling small holes at its ends, and that no witness could identify fatigue at that spot.
An ARC paper on pre-loading reports that tensile pre-loading creates compressive residual stress at stress concentrations, raising fatigue life by anything from moderate amounts to about a hundredfold, and warns that proof-testing a specimen before a fatigue test can give an inflated estimate of service life.
The same ARC pre-loading paper cites earlier studies, including work on Meteor tailplanes and by Johnstone and Payne, showing the effect was known in the fatigue literature.
RAE strain gauges on G-ALYU found peak local stresses of about 45,700 psi (315 MPa) at window corners, roughly 70% of the ultimate tensile strength of the DTD 546 alloy (about 450 MPa).
Scholz's 2019 review, discussing Paul Withey's work, says the failed bolt hole on G-ALYP had a defect in its chamfer.
A former RAE engineer writing in 2012 accepted punch-riveting as a factor in starting cracks, said it was standard practice for the era, and called the 'carpenters versus metalworkers' story 'scuttlebutt'.
BOAC Flight 781 (G-ALYP) killed all 35 aboard, 29 passengers and 6 crew, on 10 January 1954; among them were the journalist Chester Wilmot and Captain Alan Gibson.
The Court of Inquiry under Lord Cohen sat from 19 October to 24 November 1954 and cleared de Havilland, finding it had applied the scientific precautions known at the time.
The fatigue engineer Jaap Schijve points out that the beneficial effect of the earlier proof test on the Comet fatigue specimen was overlooked at the time.
