This piece follows “The Crown’s case” and takes the Comet inquiry from its fourth week to its close: the evidence of the Air Registration Board’s chief technical officer, the makers’ engineers and designer, the Board’s new requirements, and the closing speeches. As before, the words of the witnesses and counsel are taken from Flight’s weekly reports of the hearings.
The fourth week: 8 to 12 November
Tye held the floor for three days. In October 1949 the Establishment had suggested raising the ultimate factor to 2.5; the Board “took this to mean that a designer should aim at 2.5 and test at 2”, and de Havilland had done exactly that. In December 1948 the company’s tests had been “fully up to current practice, perhaps a little in advance of current practice”, and the Board had believed that any cracks at working pressure “would be non-catastrophic, and would be inspectable and repairable before reaching dangerously large dimensions”. Between 1950 and 1952 “we had begun to become much more deeply immersed in fatigue problems”, mainly in wings. For the Comet’s cabin it had not seemed urgent, because the stresses were believed to be about 40 per cent in the shell and designed not to exceed 50, and “an average life of at least 10,000 applications of pressure (equivalent to 30,000 hr), and a corresponding minimum safe life of 10,000 hr, had been estimated.” After Elba, cabin failure “was at least a potential explanation, but it still seemed inconceivable that failure at one-third of the estimated safe life could occur. De Havilland tests on the Comet nose-section at this time had reached 18,000 applications of 8¼ to 9 lb/sq in, which might have indicated a life of 54,000 hr. Pre-loading might have influenced this figure, but could not completely rule it out as an indication.” On Tuesday he gave his doubt: “I have doubts whether, looking at the extremely small fracture, one can deduce this new term ‘high-level fatigue’ from that extremely small piece of metal.” “What other theories are you suggesting?—I have no specific theories to suggest as an alternative. . . . So that my answer is a somewhat unhappy one, in the sense that I am saying that I am not very enthusiastic about the conclusion, from looking at a small specimen, that fatigue was present. On the other hand, I cannot produce any shattering alternative.” Of the tank he had no doubt: “I am clear that it must have been fatigue which caused the disruption in the tank.” Mr Scarman, for the Board, read its letter after Naples: “The Board cannot disregard the real possibility that there is some factor influencing the airworthiness of Comet aircraft which the most diligent investigations have failed to discover.” Should every new pressurised aircraft go into a tank? “For the immediate future, yes . . . I hope, thinking rather further ahead, that it will not be a continuing necessity in this respect”, because each small change to a type would otherwise call for an expensive test. “Ideally, one would like to get to a state of affairs where an aeroplane can crack safely.” On Wednesday, under Lord Cohen’s questions, he said that strain gauges should be used to check early calculations, and that if the tank’s 9,000 hours could be shown to be above average his misgivings would disappear.
Mr Bruno Jablonsky, a maker of laminated propellers who had written to the Court, put forward the one rival theory: that rapid changes of temperature on the climb had loosened the Redux glue lines between skin and stringers, and that fatigue came after. He agreed that testing was better than reasoning. Professor Murphy of Birmingham, for de Havilland, was more positive than Hall that the fracture at the starboard rear corner was fatigue, “the shading, smoothness and fine grain”, and that it had begun at the countersunk bolt hole; but a stress-raiser was not enough on its own, “their stress-concentrating effect must have been imposed on a high general level of stress.” Harper, de Havilland’s chief structural engineer, then took the Court through the company’s tests: 500 investigations and some thousands of actual tests; the two fuselage sections; the window panel to 19¼ pounds; the nose section to 13.5, then 2,000 reversals of 9, then 15, then in 1948 and 1949 to 16.5; and by 1953 “a total 18,000 reversals of at least working pressure”. The 2,000 cycles had indicated a life of 6,000 hours and the 18,000 about 54,000; “the pre-loading effect was not ignored, but the safety margin was felt to be quite satisfactory for the foreseeable future.” He had been satisfied that the stresses did not exceed 50 per cent of the ultimate, did not accept the Establishment’s 70, and thought that in future “they would have to work to a lower general level of stress than 50 per cent”. He explained why calculation could not be trusted on its own: a riveted joint should, on paper, see three times its average stress at the rivet holes and break, and did not, “which shows that these peak stresses either disappear or become altered under load, and so one must resort to fatigue tests of the particular piece of structure under consideration”. On the difference between the company’s 26,000 and Hall’s 36,000 to 40,000, worked “on the back of an envelope”: “The results did not depend on mathematics, they depend on opinion.”
Ronald Bishop, the designer, said on Thursday 11 November that the cabin “was one of the biggest problems that we had to face, since we were very alive to the consequences of a failure. It was quite obvious right from the start that if the structure of the pressure cabin failed or if a window blew out at 40,000ft, it was at least a very serious situation; and a great deal of thought and testing was done with the object of making sure that that pressure cabin was safe. We obviously missed the main point.” Fatigue “did not begin to be seriously considered as a practical problem until about the beginning of 1952. We thought we were covered by these tests that we had done, and it was a very great surprise to me personally when the fuselage blew up in the tank at 9,000 hr. We thought we had a very much longer life than that.” For the future he differed from Hall on order, not substance: the company could not afford to find the weak points by tank first and specimens after; it would test riveted joints and small sections first, “and then, finally, I believe we shall have to finish up, for a few years anyway, with a major tank test with the wings on the fuselage”. On Friday Heald announced the Board’s new requirements: tank tests of complete cabins “at any rate until knowledge of the fatigue problem has become much more exact”, two airframes of each prototype, one for static and one for fatigue testing, and scientific members added to the Board’s council. Povey, the production director, explained the concession procedure and the skin crack at the aerial window that had been drilled and never reported: “I think that he probably omitted to call the attention of the supervisor or of the inspector, and unfortunately they closed the assembly”.
The fifth and sixth weeks: 17 to 24 November
At the Institution of Civil Engineers the last witnesses were heard, and the company and the Establishment put in an agreed statement on their calculations: “in ordinary and proper design practice reliance would not be placed upon calculations except as a preliminary approach before testing”; both sets were free of error; de Havilland’s had been made “merely in order to forecast the ultimate static stress at which failure would occur on test”; and “there must have been present in the region adjoining the cut-out stresses sufficient in conjunction with local stress-raisers to cause the fatigue”. Hall, recalled, explained the two figures: his was the maximum at the edge, the company’s a mean over three inches. Why had stresses not doubled when the pressure doubled? “Within the yield point of the material twice two is four; above the yield point it is not.” Yoke Uncle “did not have to be better than average to explain the Elba and Naples disasters, but it might easily have been stronger than average.”
Counsel then spoke. Mr Grant, for the Ministry, said that if the Minister had been told in March what the Court now knew, services would not have resumed, and that “nobody could have suspected after the first disaster that the second would be likely to occur.” Mr Shaw, for the relatives, said that Lord Brabazon’s “hostages to fate” had not been of great comfort to the public, and asked whether the Board and the makers had misunderstood each other about what the company’s calculations were for: “if there is scope in the system for a misunderstanding between the A.R.B. and de Havillands or any other manufacturer as to the basis on which they are designing an aircraft, it certainly looks as if some alteration should be made in the procedure so that that is not possible.” Mr MacCrindle, for the pilots, said that nobody could blame the company for flying the aircraft in the state of belief of the time, but “if they had chosen to do the calculations which we have heard of from Sir Arnold Hall, or if they had chosen to use the strain-gauges at the appropriate parts of the aircraft . . . it may well be that they would have found a stress level which would have caused them sufficient concern to make them—to use the phrase of Mr. Bishop—think again.” Mr Tilling, for the engineers, proposed that the Board be empowered to send a novel aircraft to Farnborough or Boscombe Down for independent test at public expense before certifying it.
Shawcross closed for de Havilland over three days. “If Messrs. de Havillands had not thought that the Comet aircraft were as safe as human skill and diligent testing could make any aircraft, they would never have allowed the aircraft to fly. They thought the aircraft was safe; so did other great authorities. It has turned out now that they were wrong, and the question is ‘why?’” The aircraft had “one fatal fault, namely, this proneness to fatigue in the pressure fuselage. Beyond that it has been established to be an aircraft of quite remarkable excellence in design and in workmanship.” “There must be scores of stress-raisers in an aircraft like this where fatigue could commence. If you have a high stress, which is the pre-condition for fatigue and which we now know existed in these particular areas of the pressure cabin, it is really fortuitous where and at what particular stress-raiser the fatigue will first manifest itself.” On the belief that had failed: “We believed the static-loading test would cover fatigue. We know now that that was wrong, but not only did we believe that, the Royal Aircraft Establishment believed it, the Air Registration Board believed it, the International Civil Aviation Organisation believed it and the Civil Aeronautics Authority of America believed it.” On the further 16,000 loadings of 1953: “although possible static preloading effects were not ignored, there was considered to be a sufficient margin between the hours then flown by any Comet (less than 4,000) and the 54,000 hours implied by the total of 18,000 repeated loadings.” He put in the company’s statement of future action, and said that “all new types of aircraft would have to be tested as a whole in tanks of the kind established at Farnborough”. Mr Rimmer, for the airline, said that “out of these disasters and the subsequent investigations, air travel for future generations would be materially safer.” Mr Scarman, for the Board, defended Tye’s doubt as that of a man “who in the future will not be hurried into advising a certificate of airworthiness if he feels there is still anything to be examined or investigated.”
Heald closed for the Crown on Wednesday 24 November. The two cabin sections, Yoke Uncle’s and Yoke Peter’s, exhibited side by side at Church House, were “a classic example in the field of circumstantial evidence”, and the small piece of Yoke Peter’s roof bore, in the opinion of the two experts, “the fingerprints of fatigue”. Sabotage, freak weather, temperature, explosives in baggage, hydraulic mist, pressurisation defects and collision had each been ruled out. “We can all feel humbly proud to have taken part in such an enterprise.” Lord Cohen thanked counsel for “a single and whole-hearted desire to get at the truth of the matter” and said he would prepare the report as soon as he reasonably could. He signed it on 1 February 1955.