The Airframe Test
Written June 1992 by Jack E. Barth, a Douglas Aircraft Structural Engineer during the 1940s, 1950s, and 1960s.
To better explain the testing of aircraft structures, I need to define two terms. The conditions and resulting loads expected to be encountered by the aircraft are
defined as "limit loads." The structure is designed to take 50% higher loads than the "limit loads." These are termed "ultimate loads" and are applied to the
structure during lab tests. Assuming the structure has passed the ultimate test, one of the more significant conditions are carried to failure.
The incident I'm writing about was a case where the part I had strength-checked and approved, failed at loads less than ultimate.
I started at Douglas, El Segundo, in the stress group, October 1941. I had minor assignments on the XA26, XSB2D 1, and the XTB2D 1 before a major responsibility
on the XTB2D 1 (later the AD, then A1 Skyraider).
The forward bulkhead was an important part of the fuselage, receiving loads from the engine mount. The two upper legs of the four legs of the engine mount attached
at the upper corners of this bulkhead. The two lower legs reached farther aft to another bulkhead. The analysis of the forward bulkhead was relatively simple. The
critical vertical loads were carried down the outer cap (angle) of the bulkhead and fed into the fuselage outer skin.
The analysis of the engine mount (not done by me) was less simple. The engine mount structure was not a welded steel truss, but a built-up aluminum structure.
The legs could carry bending and shear loads as well as axial loads. The upper and lower legs were of different lengths, making the determination of the distribution
of the vertical loads to the fuselage a difficult problem. I used the loads given to me and showed no extra margin in the strength check. The critical item was the
0.064 24 ST outer cap, in compression.
Now, to the lab test. The flight condition represented was the plus low angle of attack (+L.A.A.), 9G limit. A complete aircraft structure, dedicated for testing, was
subjected to the calculated loads. Hydraulic jacks applied the up load (aerodynamic lift) of approximately 170,000 pounds, less wing weight inertia. Similar jacks
pulled down on the fuselage, engine mount, and tail, representing the balancing inertia and tail loads. Since this was the most critical static test, there was
considerable tension as the loads increased.
The fuselage outer skins began to show deep wrinkles and the wing showed considerable deflection at the tips. This tension was broken when at about 135% of limit,
the outer cap of my bulkhead buckled in compression. All the jack loads were relaxed, and the inquiry began. What failed? Where is the test analysis? And who
did it?
Of course, my boss looked into it. He soon decided that my work was O.K. So, the next thing was to look at the loads. The analysis of the engine mount was
reviewed and corrected, giving my bulkhead higher loads. The critical outer angle was changed from 0.064 24 ST to 0.080 75 ST. 75 ST aluminum alloy was just being
introduced in the U. S., but I believe the Japanese used a similar alloy as early as the Zero.
As for me, the notoriety was good. I was first suspected of a boo boo. Then, I was recognized as doing a good job. Over the years, the structures that I
approved were many times tested to ultimate and failure loads. Working with the design engineer, I never padded the loads, worked to zero margins, and never
had an embarrassing failure.
From 1941 to 1962 I worked for an engineer named Pete Shaw. We became very close, and I greatly admired him. At times he negotiated with me to approve
designs that I felt were inadequate. Since these structures would be tested to ultimate loads, the affect of an early failure would be that the test part would have to
be repaired and early production parts retrofitted. We were always striving for minimum structural weight. The management sometimes requested that marginal
structural parts be approved to minimize weight. It was understood that the stress engineers would not be held negligent.
My next effort in this series was about my analysis and approval of the design of the D558 1 engine access door that I believed to be marginal.