Dark Eagles: A History of the Top Secret U.S. Aircraft

by Curtis Peebles

  Because the situation in Bosnia was highly fluid, the GNAT-750s needed to be operational by October 1993 at the latest. A development effort with so short a time frame could only be accomplished, however, if freed of the funding and acquisition constraints of a military program. There had been a similar situation some forty years before. The solution was the same; the CIA was given control of the program.[620]

  The GNAT-750 had now become a Black airplane.

  A PALER SHADE OF BLACK

  By the late summer of 1993, the CIA GNAT-750 effort was under way.

  The CIA acquired two GNAT-750s from General Atomics, which were then modified with the relay antennae. A large teardrop-shaped dome was added to the top of the GNAT-750 to house the data link. The flight tests were done at the El Mirage site. The CIA GNAT-750s were to be ready for operations by October 1, 1993; once operational, they were to be based in Italy.

  In spite of its new Black status, the project did not go smoothly. The CIA GNAT-750 program had to face the political realities of spying in the 1990s.

  Not even congressmen had been told of the U-2 overflights. In 1993, however, the autonomy of CIA Director R. James Woolsey Jr. was far less than that enjoyed by Allen Dulles in 1956, when the U-2 program was about to begin. Now it was necessary for the CIA to consult with the Defense Department and Congress on the GNAT-750 effort.

  The overall UAV program itself was caught in a tug-of-war between the various factions who were soon faulting the CIA's handling of the GNAT-750 program. They objected to the CIA's unwillingness to tell others what was going on with the project. One congressional staffer complained (rather incongruously), "They tried to turn it into a secret program."[621]

  Not surprisingly, word of the CIA GNAT-750 program leaked. The early details were fragmentary, with one report referring to the GNAT-750 as "an unmanned SR-71 follow-on." The image that phrase brings to mind is far from a plane with a wooden propeller.[622] But more detailed accounts were soon appearing.

  Clearly, the GNAT-750 was a very different shade of Black than the U-2.

  The concept of Black itself had also changed. HALSOL was classic Black — secret from start to finish. Since then, the concept had become blurred. Teal Rain was Black, yet the end result, the GNAT-750, was White.

  It was even sold to an Allied country. Then, suddenly, the White GNAT-750 became Black with the start of the CIA program.

  The net result was that the CIA GNAT-750 effort was a Black project that was almost open. It was a very strange picture; yet it would become stranger still.

  The political infighting over the CIA GNAT-750 was aggravated by technical problems. It was reported that the GNAT-750 was suffering from computer and software problems, which resulted in cost overruns, schedule delays, and "annoyed congressional staffers." A "Pentagon official" complained that the CIA lacked experience with systems integration or developing tactical communications links. Defense officials spoke of a "hobby shop approach" to development and said that the CIA was hurrying the project too fast.[623]

  By late October, the GNAT-750 was judged ready for its final systems integration test flight. On the night before the flight, the CIA contract personnel modified the software program. The new program shut down the GNAT-750's motor and data link when its speed fell below 40 mph. It was assumed that if the UAV was at this low a speed, it would be on the ground.

  During the flight the next day, the GNAT-750 was hit by a gust of wind.

  This produced a low-speed indication, and the software shut down the motor and data link. The GNAT-750 rolled over on its back, went into a flat spin, and crashed. The surviving GNAT-750 was grounded pending an investigation. In all, the technical difficulties and the crash caused a three-month delay, as well as a $1 million cost overrun.[624]

  The program regrouped from the mishap. On December 2, 1993, a GNAT-750 flying over the El Mirage test site successfully transmitted im-agery of moving targets, such as a tank, to a relay aircraft, which then passed it on to the Pentagon. Tests also indicated the GNAT-750 was quite stealthy. The radar at Edwards Air Force Base did not pick it up. It was believed that this was due to several factors — the shaping of the fuselage, its heavy use of carbon epoxy materials, and, ironically, its slow speed of 90 knots. The radar's own software filters out such weak and slow-moving targets as returns from birds or weather.[625]

  BOSNIAN OVERFLIGHTS

  Under the revised schedule, the remaining GNAT-750 was to be operational by February 1, 1994. CIA Director Woolsey reportedly demanded that it be operational before his next appearance at congressional budget hearings. There was, however, a new political problem. The Italian government was having second thoughts about playing host to the CIA operation.[626] By late January 1994, they formally turned down the U.S. request, citing the possible danger to civil aircraft from the GNAT-750.[627]

  The UAVs had lost their base of operation. But the world had changed since the U-2 had taken off on that first overflight. There were other possibilities, ones that would have been unthinkable in the mid-1950s. The CIA GNAT-750 would be based in Albania.

  When the Italians refused permission, the U.S. government made a direct appeal to senior Albanian officials. They approved the operation. The CIA unit would be flown in from Germany in a C-130. The single GNAT-750 and a satellite transmission station would be set up at an Albanian military base on the Adriatic coast. The three prime candidates were Scutari, Durres, and Tirane, the capital.

  Pentagon acquisition chief John Deutch endorsed the recommendation Albania's location had advantages for the GNAT-750 overflights. They would have direct access to Montenegro, Kosovo, and Macedonia without having to overfly a third country. It was in these regions that U.S. ground troops were assigned as part of the UN peacekeeping force. Bosnia Herzegovina was also well within range of the GNAT-750, as was the Serbian capital, Belgrade.[628]

  It was a stunning example of how the world had changed in just a few years. Between 1949 and 1953, the CIA and the British Secret Intelligence Service had tried to establish networks of agents within Albania in an attempt to overthrow its Communist government. The effort ended in disaster — the agents were caught and killed. Throughout the 1970s and 1980s, Albania remained a closed, Stalinist country. It was the last Communist country in Eastern Europe to undergo the Revolution of 1989.[629]

  The CIA unit was in place by early February. Its deployment in Albania was marked by a major article in Aviation Week and Space Technology magazine. The results of the GNAT-750's Bosnian overflights were mixed.

  It was reported that of the thirty attempted overflights, only twelve were successful, due to continuing bad weather, maintenance problems, and difficulties with the GNAT-750's data link.[630] One of the UAV's missions was to follow UN convoys. Other flights were targeted against entrenchments; from 6,000 feet they were able to identify decoy artillery and SAM sites.

  Airfields, troop and artillery movements, supply dumps, and tank locations were also monitored. Best resolution was eighteen inches. The GNAT-750s were quite stealthy — at no time did people on the ground realize the UAVs were overhead.

  The overflights did reveal several problems. The relay aircraft was a two-man Schweitzer RG-8 powered glider. Because of Albania's location, it could only spend about two hours on station. Another six hours was spent flying to and from the area. Having to coordinate the GNAT-750s with the manned RG-8 meant the UAVs could only be flown in specific areas, and at certain times and altitudes. Real-time changes in the route, to more closely examine side roads and buildings, also disrupted the GNAT-750 activities.

  By the summer of 1994, the U.S. European Command requested the renewal of GNAT-750 overflights. The CIA preferred to fly from an Italian base, but this was again refused. The unit was based in Croatia. This allowed better use of the RG-8 relay plane. The GNAT-750s were also equipped with an improved IR scanner.[631]

  The CIA also wanted to expand its GNAT-750 unit with three more UAVs and an additional ground station. Two of
the UAVs would be used for reconnaissance, while two others would be used as relay aircraft. The reconnaissance GNAT-750s were also expected to be fitted with ELINT receivers that could pick up both radar signals and transmissions from walkie-talkies.

  This would be done by buying GNAT-750s originally ordered by Turkey.

  The Turkish government was in a budget crunch and lacked the money to pay for the three GNAT-750s already delivered, much less the four still at the General Atomics factory. Each GNAT-750 cost $800,000, while the ground station cost $1.2 million. The CIA asked Congress for a u t h o r i t y to reprogram funds to buy the UAVs.[632]

  THE CIRCLE CLOSES

  The HALSOL and GNAT-750 represent the future of airborne reconnaissance. The smaller post-Cold War air force is unlikely to have the billions of dollars needed to develop and fly a new manned reconnaissance aircraft.

  In contrast, a unit of UAVs can be built and operated for less than one-tenth the cost of a single F-15E or F-117A. The GNAT-750's similarity to a big model airplane is an advantage, rather than a shortcoming. This allows it to combine long-duration flight times with the simplicity of the small tactical […] value […] Black projects.

  Despite all the problems, it was still done faster and at a lower cost than conventional methods could have accomplished. The criticism was based as much on jealousy as on the problems. At the same time, the GNAT-750 was a very different Black than the U-2, A-12, or even the Have Blue (which also started out White and then went Black). The involvement of congressional staffers meant that the lines of authority were less clear and direct.

  The GNAT-750 effort also marked the return to Black aviation by the CIA after two decades. The Nationalist Chinese U-2 overflight program had ended in 1968. The CIAA-12s were retired the same year. Between 1969 and 1971, the CIA-sponsored D-21 Tagboard made their few disappointing flights. By the mid-1970s the CIA U-2 operation had been closed down, and the surviving planes had been transferred to the air force and NASA.

  With this, the GNAT-750 project brings the story of the U.S. Black airplanes full circle.

  The A-12, D-21, and HALSOL had all been kept secret for a decade or more before their existence was revealed. Up to that time, there had been no hint of these Dark Eagles. Such secrecy, along with the large number of stories about the stealth fighter, inevitably gave rise to speculation about other Black airplanes, ones that were still secret. Reports and sightings of these secret airplanes were soon being whispered about. From time to time, the stories would be published.

  They were tales of darkness and shadows.

  CHAPTER 10

  The MiGs of Red Square

  Have Doughnut and Have Drill

  Therefore I say, "know the enemy and know yourself; in a hundred battles, you will never be in peril."

  Sun Tzu ca. 400 B.C.

  The longest continuing U.S. Black airplane program is the secret test flying of MiGs and other Soviet aircraft. This effort's tentative beginnings were in the mid-1950s, before the U-2. It began in earnest a decade later, contemporarily with the A-12 and D-21, and has continued to the present day.

  Unlike the other Black airplane programs, such as the Have Blue, F-117A, or HALSOL, MiG operations still remain Black. The program can not even be acknowledged.

  It is not known exactly the actual number or types of aircraft involved, where they came from, or the complete history of the program. There are only a few, limited accounts, and it is probable that many of these are, at best, incomplete, and at worst, wrong. In one case, a published MiG tale proved spectacularly wrong.

  It is known that these Dark Eagles brought about a fundamental change in air-combat tactics. They revitalized the art of dogfighting at a time when, seemingly, it had nearly been forgotten. The knowledge gained from these planes was reflected in the success of U.S. Navy air operations over North Vietnam in the final year of the war, as well as in the founding of the Navy's Top Gun school.

  THE BLACK YAK

  The program started with a C-124 Globemaster landing at Wright-Patterson Air Force Base in the mid-1950s. Its payload was a Soviet-built Yak 23 Flora. This was a small single-seat, single-engine jet fighter. The 217 aircraft had first been exported to the Polish, Czechoslovakian, Bulgarian, and Romanian air forces in 1951. Although quickly superseded by the more advanced MiG 15, it was the first jet operated by these Eastern European air forces. For the U.S. Air Force, it introduced the secret testing of Soviet aircraft.

  According to one account, an Eastern European intelligence officer contacted an American intelligence officer and offered the "loan" of the plane.

  The deal was made and the Yak 23 was packed in shipping crates and left aboard a railroad car. After its arrival, the parts were photographed, to insure they could be repacked correctly.

  The project pilot for the Yak 23 flights was Capt. Tom Collins. In September 1953, he had become the first U.S. pilot to fly a MiG 15. It had been flown to South Korea by a defecting pilot. Unlike that effort, the tests Collins conducted with the Yak 23 were top secret. The plane carried U.S. Air Force insignias, while its buzz number of "FU-599," and the "0599" serial number actually belonged to an F-86E. If the curious asked what kind of plane it was, they were told it was an "X-5."

  The flight tests of the Yak 23 lasted about a month. The plane was then disassembled, loaded aboard a C-124, and flown away. It would be another forty years before the existence of a U.S. Air Force Yak 23 would be revealed. In December 1994, four photos of the plane were published. Where the plane had come from, and where it went after leaving Wright-Patterson, are secret still.[633]

  In the years following the Yak 23's brief stay at Wright-Patterson, a new generation of fighter aircraft, such as the F-4 Phantom II, was developed.

  The F-4 was the first fighter designed from the start with only air-to-air missiles — the radar-guided Sparrow and the shorter-range Sidewinder infrared-guided missile. With the new missiles came the new attitude that dogfighting was obsolete. The air-to-air training given to new navy F-4 crews was extremely limited. It involved about ten flights and provided little useful information. By 1964, few in the navy were left to carry on the tradition of classic dogfighting.[634] Then came Vietnam.

  THE AIR WAR OVER NORTH VIETNAM

  The early years of the air war over North Vietnam showed the faith placed in missiles was terribly in error. Between 1965 and the bombing halt in 1968, the U.S. Air Force had a 2.15 to 1 kill ratio. The navy was doing slightly better with a 2.75 to 1 rate. For roughly every two North Vietnamese MiG 17 Frescos or MiG 21 Fishbeds shot down, an American F-4, F-105, or F-8 would be lost. This was far worse than the 10-plus to 1 kill rate in Korea.[635]

  More serious, the percentage of U.S. fighters being lost in air-to-air combat was growing. During 1966, only 3 percent of U.S. aircraft losses were due to MiGs. This rose to 8 percent in 1967, then climbed to 22 percent for the first three months of 1968.[636]

  In 1968, navy Capt. Frank Ault was assigned to learn the reasons for this poor showing against the MiGs. The "Ault Report" was issued on January 1, 1969. It found 242 problems that ranged from hardware to crew training. The Sidewinder and Sparrow missiles showed very poor reliability. A full 25 percent of Sparrows failed because their rocket engines never fired. The Sidewinder and Sparrow were both limited to 2- to 2.5-g maneuvers, as they had been designed for use against nonmaneuvering bombers rather than fighters.[637] It took a full 5.2 seconds to fire a Sparrow; yet, the average time the F-4 crews had to fire was 2.2 seconds. To hit the target, the F-4's radar beam had to be kept on the MiG. This was extremely difficult in a turning dogfight.

  Far more important was the training of the crews. Few F-4 crewmen knew the firing parameters for the missiles. These changed with altitude, and whether it was a tail, head-on, or side attack. The crews lacked the knowledge to judge the ever-changing parameters in the midst of the fight.

  The result: of some six hundred missiles fired between 1965 and 1968, only one out of ten or eleven had any chance t
o hit its target.[638]

  Finally, the emphasis on interception meant the F-4 crews had only the sketchiest knowledge of dogfighting. The design of the F-4 made it ill-suited for a tight-turning dogfight. In contrast to the MiG 17, the F-4 was large and heavy. When a tight turn was made, the F-4 would lose speed. The MiG 17's superior turning capability then allowed it to close to gun range.

  All too often, hits from the MiG 17's "outmoded" cannons would then destroy the F-4.

  The key to survival in the skies of North Vietnam, as it had been in every air war, was to make the enemy pilot fight on your terms. This meant knowing his weaknesses, while using your plane's strengths to maneuver into position to […] the enemy […] was acquiring the knowledge.

  Out in the Nevada desert, a MiG 21 awaited.

  HAVE DOUGHNUT

  In 1967, the Defense Intelligence Agency secretly acquired a single MiG 21. The country the MiG 21 came from, and the means by which it came to the United States, remain secret to this day. Because U.S. possession of the MiG 21 was, itself, secret, it was tested at Groom Lake. A joint air force-navy team was assembled for a series of dogfight tests. The project was code-named "Have Doughnut."[639]

  Comparisons between the F-4 and the MiG 21 indicated that, on the surface, they were evenly matched. At a speed of Mach 0.9 at 15,000 feet the instantaneous turn rates of the two planes were nearly identical, at 13.5 degrees per second. At Mach 0.5, the MiG 21 held the edge at 11.1 degrees versus 7.8 degrees for the F-4.[640] But air combat was not just about degrees per second of turn rate. In the final analysis, it was the skill of the man in the cockpit. The Have Doughnut tests showed this most strongly.