On July 23, 1958, Johnson attended a meeting at which he presented Archangel I and Gusto 2A. Bissell was present and Johnson noted in his log that both designs were “well received.” A Navy commander also present alluded to a Navy idea for an inflatable aircraft, and Bissell requested Kelly’s comments on the concept. On July 31 the advisory panel met to formally discuss both Lockheed concepts, during which Bissell noted that reasonable progress was being made, but that he thought the way forward would become clearer after their September meeting.
Having been provided with details about the intriguing Navy concept, Johnson conducted a thorough evaluation for Bissell. Under the Navy project name Champion, Goodyear was proposing a reconnaissance vehicle with inflatable wings that could be rolled up whilst in transit aboard an aircraft carrier, then inflated for launch. It was envisaged that the ramjet-powered vehicle would be lifted to altitude by balloon and would cruise at 125,000–150,000ft. A quick calculation by Johnson determined that the balloon would need to be over a mile in diameter! A more detailed study followed during which a tug aircraft was also evaluated, but the unreliability of the concept for reconnaissance over highly sensitive areas was enough to ensure that the design didn’t progress.
On August 18, Johnson made a note in his log to conduct studies into a modified version of Archangel I utilizing ramjets on the wingtips to gain a further 10,000ft in altitude. However, Archangel II made no concessions to reducing RCS and during meetings held in Washington, DC and Boston from September 17–24, Johnson was once again reminded that the President required a replacement reconnaissance aircraft that was invisible to radar.
The Land Panel convened in Boston on September 22–23, and both Johnson and Widmer presented their designs independently. The panel decided to terminate further studies into Champion, voiced interest in the Convair design, but rejected Archangel II primarily on the grounds of its poor RCS. Gusto 2A, the panel contended, required further development.
Five days after the Boston meeting, Johnson began making notes on what would become his A-3 design. Consisting of two wingtip-mounted ramjets and two JT-12A turbojets, Johnson assigned the concept to Dan Zuck, Ed Baldwin, and Henry Combs. Continuing the angel theme, the team referred to them as Cherubs, since they were smaller than Archangel, and they worked on several design permutations. On October 30 Ed Baldwin completed the final A-3 design, which had a length of 62.3ft, a span of 33.8ft, and weighed in empty at 12,000lb. Despite the relatively low levels of thrust generated by the two turbojets, the A-3 could reach Mach 3.2 at a cruise altitude of 95,000ft and had a range of 4,000nm. But weight issues dogged the design and when “LD” MacDonald and his team conducted RCS measurements on two scale models, the returns averaged out as the same for those of the U-2, which were much greater than those for the all-metal Gusto 2S, which were themselves larger than for Gusto 2A, which incorporated RAM.
On November 12, 1958, Johnson and Widmer presented their designs to the Land Panel — Johnson the A-3 and Widmer FISH. Three days later the panel communicated its findings to James Killian — the President’s science advisor. Their recommendation was that FISH be selected, but should this design later prove unable to meet “the desired technical features” the panel would wish to review other alternatives before recommending firmly a second choice (the A-3). Bissell called Johnson with the bad news on November 26, pointing out that actually the two designs compared quite favorably, but that the crunch issue had again been RCS, an area in which Convair was clearly ahead.
On December 22, Convair was given the go-ahead to proceed with further detailed design development of FISH. Twelve aircraft were to be bought and it was thought that a minimum of one B-58 launch-aircraft would be required for every four FISH. Since Strategic Air Command (SAC) was operating a wing of B-58 Hustler bombers from Carswell AFB in Texas, the base provided ready cover and security for Project FISH — the close proximity of the base to Convair’s Fort Worth plant was also an advantage.
A major operational disadvantage of FISH was its reliance upon a carrier aircraft for launch. The effect upon Johnson of the Land Panel’s decision to reject the A-3 caused him to focus his next set of designs on reducing RCS values to a minimum, whilst retaining the vehicle’s ability to launch itself independently. Therefore, beginning on November 26 and over the next two months, Johnson and his team began working on designs designated A-4, A-5, and A-6. The designs utilized an idea of Frank Rodgers to incorporate chines beginning at or near the nose that blend into the fuselage and merge into the leading edge of the wings. In addition, the designs were of small physical size and vertical surfaces were hidden above the wings. However, at the beginning of January 1959, before design studies had been completed on the A-6, Johnson instructed his team to work on a series of small, non-stealthy designs, designated A-7 through A-9. Powered by a single J58 turbojet and two ramjets, none of the designs were subsequently judged to be viable — their range was inadequate, typically 1,640nm — and despite their size, the RCS was too great.
Following the November 12 meeting, Convair began developing production plans, awarding subcontracts for the design of various sub-systems, in addition to conducting work that would refine and test various elements of their FISH design. But it was during the nearly 300 hours of wind tunnel testing of 1/17th scale models that problems were identified. Firstly the drag coefficient acting on the B-58 with FISH in place was nearly double that of a “clean” aircraft. This meant that instead of the usual three minutes to accelerate from subsonic speed to Mach 2, it would take nearly nine. In addition, FISH needed to be lengthened in order to improve stability and carry additional fuel; this meant that the B-58 needed to be 5ft longer. Neither of these were issues for the B-58B, as it was both longer and had uprated J79-9 engines; the problem was that the new bomber was unfunded.
