Convair's proposal depended on two uncertain factors. First and foremost was the unproven technology of the ramjet engines. At the time, no aircraft in existence could carry a large, ramjet-powered aircraft into the sky and then accelerate to sufficient speed for the ramjet engines to be ignited. Since ramjet engines had only been tested in wind tunnels, there was no available data to prove that these engines would work in the application proposed by Convair. The second uncertain factor was the B-58B bomber that was supposed to achieve Mach 2.2 before launching the FISH above 35,000 feet. The version of the B-58 was still in the design stage.
Convair's proposal suffered a major setback in June 1959, when the Air Force canceled the B-58B project. Conversion of the older, slower B-58A into the supersonic launching platform for the FISH was ruled out by the high cost and technical difficulties involved. Moreover, the Air Force was unwilling to part with two aircraft from the small inventory of its most advanced bomber. Even had the B-58B program not been canceled, the FISH proposal would probably not have been feasible. Convair engineers had calculated that the added weight of the FISH would prevent the B-58B from achieving the speed required to ignite the parasite aircraft's ramjet engines.
The Convair proposal was therefore unusable, but the Lockheed design with its high radar cross section was also unacceptable to the Land committee. On 14 July 1959, the committee rejected both designs and continued the competition. Lockheed continued to work on developing a design that would be less vulnerable to detection, and Convair received a new CIA contract to design an air-breathing twin-engine aircraft that would meet the general specifications being followed by Lockheed.
Following recommendations by the Land committee, both Lockheed and Convair incorporated the Pratt Whitney J58 power plant into their designs. This engine had originally been developed for the Navy's large, jet-powered flying boat, the Glenn L. Martin Company's P6M Seamaster, and was the most powerful engine available. In 1958 the Navy had canceled the Seamaster program, which left Pratt Whitney without a buyer for the powerful J58 engine.
Although the Land committee had not yet found an acceptable design, it informed President Eisenhower on 20 July 1959 that the search was making good process. Concerned about the U-2's vulnerability to detection and possible interception and aware that the photo satellite project was encountering significant problems, the President gave his final approval to the high-speed reconnaissance aircraft project.
By the late summer of 1959, both Convair and Lockheed had completed new designs for a follow-on to the U-2. Convair's entry, known as the KINGFISH, used much of the technology developed for the F-102, F-106, and B-58, including stainless steel honeycomb skin, planiform wing design, and crew capsule system, which eliminated the need for the pilot to wear a pressurized suit. The KINGFISH had two side-by-side J58 engines inside the fuselage, which significantly reduced the radar cross section. Two additional important design features that contributed to a small radar return were fiberglass engine inlets and wings whose leading edges were made of Pyroceram.
Lockheed's new entry was much like its first, but with several modifications and a new designator, A-12. It too, would employ two of the powerful J58 engines. Lockheed's major innovation in reducing radar return was a cesium additive in the fuel, which decreased the radar cross section of the afterburner plume. This improvement had been proposed by Edward Purcell of the Land committee. Desiring to save weight, Kelly Johnson had decided not to construct the A-12 out of steel. Traditional lightweight metals such as aluminum were out of the question because they could not stand the heat that would be generated as the A-12 flew at Mach 3.2, so Johnson chose a titanium alloy.
On 20 August 1959, Lockheed and Convair submitted their proposals to a joint Department of Defense, Air Force, and CIA selection panel. The two aircraft were similar in performance characteristics, although the Lockheed design's specifications were slightly better in each category. The Lockheed design was also preferable in terms of overall cost. In the vital area of vulnerability to radar detection, however, the Convair design was superior. Its smaller size and internally mounted engines gave it a smaller radar cross section than the Lockheed A-12.
Some of the CIA representatives initially favored the Convair KINGFISH design because of its smaller radar cross section, but they were eventually convinced to support the Lockheed design by the Air Force members of the panel, who believed that Convair's cost over-runs and production delays on the B-58 project might be repeated in this new project. In contrast, Lockheed had produced the U-2 under budget and on time. Another factor favoring the A-12 was security. Lockheed had experience in running a highly secure facility (the Skunk Works) in which all of the key employees were already cleared by the Central Intelligence Agency.
The CIA selected Lockheed's A-12 over a Convair proposal called KINGFISH.
Despite its vote in favor of the Lockheed proposal, the selection panel remained concerned about the A-12's vulnerability to radar detection and therefore required Lockheed to prove its concept for reducing the A-12's radar cross section by 1 January 1960. On 14 September 1959, the CIA awarded a four-month contract to Lockheed to proceed with anti-radar studies, aerodynamic structural tests, and engineering designs. This research and all later work on the A-12 took place under a new codename, Project OXCART, established at the end of August 1959 to replace its more widely known predecessor, Project GUSTO.
By mid-January 1960, Lockheed had demonstrated that its concept of shape, fuel additive, and non-metallic parts would reduce the OXCART's radar cross section substantially. Richard Bissell, however, was very upset to learn that the changes had led to a reduction in the aircraft's performance, which meant it would not be able to attain the penetration altitude he had promised to President Eisenhower. Kelly Johnson then proposed to reduce the aircraft's weight by 1,000 pounds and increase the fuel load by 2,000 pounds, making it possible to achieve the target altitude of 90,000 feet. Afterward, he noted in the project log: "We have no performance margins left; so this project instead of being 10 times as hard as anything we have done, is 12 times as hard. This matches the design number and is obviously right". Johnson, "Archangel log", 21 January 1960.
These changes satisfied Bissell, who notified Johnson on 26 January that the CIA was authorizing the construction of 12 of the new aircraft. The actual contract was signed on 11 February 1960. Lockheed's original quotation for the project was $96.6 million for 12 aircraft, but technological difficulties eventually made this price impossible to meet. Recognizing that fabricating an aircraft from titanium might involve unforeseen difficulties, the CIA included a clause in the contract that allowed costs to be reevaluated. During the next five years, this clause had to be invoked on a number of occasions as the A-12's costs soared to more than double the original estimate."
![](data:image/svg+xml;charset=utf-8,<svg height="540px" width="720px" xmlns="http://www.w3.org/2000/svg" version="1.1"/>)
![](data:image/svg+xml;charset=utf-8,<svg height="784px" width="617px" xmlns="http://www.w3.org/2000/svg" version="1.1"/>)
![](data:image/svg+xml;charset=utf-8,<svg height="1px" width="1px" xmlns="http://www.w3.org/2000/svg" version="1.1"/>)
![](data:image/svg+xml;charset=utf-8,<svg height="16px" width="16px" xmlns="http://www.w3.org/2000/svg" version="1.1"/>)