From Scramble - The Aviation Magazine
History of the XST
In 1974, the US Defense Advanced Research Projects Agency (DARPA) initiated a program known as PROJECT HARVEY, after a well-known comedy about an invisible giant rabbit, that requested designs of an "eXperimental Survivable Testbed" (XST) aircraft with a low RCS. It is a misunderstanding that XST stands for "eXperimental Stealth Tactical".
Lockheed was not among the companies contacted by DARPA with this request, but in 1975 the company got wind of the project and lobbied the government successfully to have them included.
Two Lockheed "Skunk Works" employees, mathematician Bill Schroeder and computer scientist Denys Overholser, worked on the XST program. Schroeder realized that it would be much easier to compute RCS if the shape of an aircraft could be reduced to a set of flat surfaces, or "facets". Schroeder approached Overholser with the idea, and within five weeks Overholser had written a computer program named "Echo I" that could determine the RCS of a "faceted" aircraft. Armed with Echo I, Schroeder came up with an initial XST design that he called the "Hopeless Diamond", and handed programme director Rich a sketch of it in May 1975.
Not everyone was impressed by the Hopeless Diamond. Lockheed's prestigious Kelly Johnson disliked it, likely on the traditional aircraft designer's belief that "no aircraft that looked that ugly could possibly be any good." Johnson favored a flying-saucer-like design, but nobody could figure out how to make a saucer shape fly very well. To strengthen his case, Rich had a three meter (ten foot) sized mockup of the Hopeless Diamond built and tested against radar in September 1975. The RCS proved to be as small as predicted.
Work on the Northrop demonstrator was conducted by a team under John Cashen, who had come to Northrop after working at Hughes on how targets appeared to radar and infrared sensors, and Irv Waaland, a designer who had come over to Northrop from Grumman. Although the Lockheed team used a computer program to come up with their design for the XST, Cashen said later that Northrop didn't have such a luxury, and worked up their design using a combination of theoretical analysis and cut-and-try experiments.
By October 1975, the DARPA XST competition had been reduced to two finalists: one from Northrop, and a refined version of the Lockheed Hopeless Diamond. The Northrop entry was a delta with a faceted fuselage, with the jet engine mounted on the back and the intake above the cockpit. A fine mesh screen sealed the air intake from radar, while a pair of tilted fins hid the exhaust. The Lockheed design had two wings with a sweep of 72.5 degrees and a pointed rear fuselage, giving the rear of the aircraft a "W" outline. Twin intakes were placed along the cockpit and covered with grills, while the exhausts were arranged as long slots along the rear fuselage. Such "platypus nozzles" reduced the "infrared signature" of the exhaust, which was also concealed by twin inward-canted fins.
Since RCS is a function of viewing angle, both designs were optimized to have the lowest RCS from the bottom and front, the most probable viewing angle of adversary ground defences. Northrop and Lockheed built one-third scale models in December 1975. The models were put through preliminary RCS tests, and full-scale models were tested the next year.
The Lockheed design had a tenth of the RCS of the Northrop design. It was so invisible to radar that a radar operator performing tests on the model at White Sands, New Mexico, thought it had fallen off the pole. Bird droppings increased the RCS by 50 percent, and so the model had to be regularly cleaned. Lockheed won the competition in April 1976.
The Northrop team was heartbroken, even though their engineers admitted Lockheed had the better design. They returned to their calculations and would eventually catch up with Lockheed at the stealth game, with the Tacit Blue and eventually the B-2 Spirit bomber.
The Lockheed contract specified construction of a pair of XST aircraft for flight and RCS tests, with $32.6 million USD provided by DARPA and the USAF. Lockheed had to furnish $10.4 million USD more out of its own pocket, which was a big risk at the time, since Lockheed was in difficult financial circumstances.