The Space Reviewin association with SpaceNews

Recovery illustration
The recovery sequence for American film-return satellites during the 1960s and into the 1980s, developed before it was possible to beam useful images down to Earth. Although unusual, this method was highly reliable. (credit: NRO)

Ike’s gambit: The KH-8 reconnaissance satellite

An Air Force officer walks into a bar…

No, this is not the beginning of a joke.

An Air Force officer walks into a bar in the early 1970s. The bar was in the officer’s club on an airbase in England, probably RAF Mildenhall. He was an American officer, apparently in England to attend the annual Farnborough air show, one of the largest air shows in the world. While tipping back a few beers, the officer hears a couple of other officers—it is unknown if they were American or British—discussing an object that a farmer had found on land near where one of the officers lived off-base. The local authorities had been called in to look at the object and had no idea what it was. But the person telling the story noted that whatever it was, “it had a lot of glass in it.”

The details of the story are sketchy, unconfirmed, and it is unclear if the visiting officer instantly knew what they were talking about. But what the farmer had found in his field were apparently pieces of a KH-8 GAMBIT high-resolution spy satellite that had fallen to Earth only a short time before. According to a cryptic entry in a declassified reconnaissance satellite history, intelligence officials knew that one of their spysats had unexpectedly fallen to Earth, but they did not suspect that it had actually come down on land, or that it had been found by a civilian. It was not the first time that a top secret spy satellite had been discovered by a farmer (see: “Spysat down!” The Space Review, February 18, 2008), but fortunately the officer’s intervention prevented the story from leaking.

For a top secret satellite, the KH-8 has a fascinating history.

There was good reason for the intelligence community to want the find to remain as secret as possible. The KH-8 was the highest resolution reconnaissance satellite ever built. Even today, it apparently holds the record for the best reconnaissance photographs returned from orbit by any spacecraft, a combination of both a powerful camera and the ability to dramatically lower its orbit, to “swoop” in over a target at altitudes of apparently only 70 nautical miles (130 kilometers). The KH-8 could apparently see objects on the ground as small as a baseball and had the ability to photograph people with enough resolution to see their arms and legs. Later satellites had bigger mirrors, but flew at higher altitudes and could not return pictures as good.

The fallen spysat story is only one of several fascinating anecdotes about the KH-8. There was the time that some Air Force officers used one to take a self-portrait. And the time that the National Reconnaissance Office used one to photograph the crippled Skylab space station. For a top secret satellite, it has a fascinating history.

Upgrading the KH-7

The KH-7 GAMBIT was initially limited by the amount of attitude control gas that the Orbital Control Vehicle (OCV) carried, meaning that a number of missions had to return before they had used up all their film because they were unable to continue pointing the camera. But as the program progressed the amount of film returned increased dramatically. Whereas mission 4005, launched in February 1964, made only 13 photo passes, mission 4025, launched in February 1966, made 74 photo passes.

An inherent limitation of film-return reconnaissance systems was that it took a long time for their pictures to reach the interpreters in Washington. This was a result of many factors, including the transportation time for getting the returned film from the Pacific Ocean far west of Hawaii a third of the way around the world to Kodak’s processing center in Rochester, New York, and also the time it took to process the film. The NRO had several efforts underway to reduce this time, including evaluating installing a film processing machine on board a C-135 transport based in Hawaii so that it could process the film in flight to Washington. But for unknown reasons such a capability was never developed.

Another limitation was that it took a long time to prepare a mission for flight. If a crisis occurred, by the time a satellite could be prepared for launch, put in orbit, return its film to Earth, and the film processed, the crisis might be over before anyone saw the photos. The NRO developed a goal to keep reconnaissance systems ready for quick launch. At least one KH-4 CORONA was kept in “R-7” status at all times, meaning that it was ready for launch in seven days. By summer 1965 the NRO had also demonstrated an ability to keep a CORONA vehicle in R-1 status—only a day away from launch—for extended periods of time. Although GAMBIT’s figures remain classified, it apparently had similar capabilities.

The early GAMBIT missions had ground resolution of approximately four feet (1.2 meters). This was on average about three times better than the CORONA search systems then in operation. By the later missions the resolution had been improved to approximately two feet (0.6 meters). This meant that the camera could produce an image that depicted objects on the ground that were at least two feet on a side.

Additional equipment was also incorporated into later missions. At some point a Stellar/Index camera manufactured by Itek Corporation and nearly identical to one carried aboard CORONA missions was also added to some GAMBIT missions. The Stellar/Index camera was used for taking broad images allowing photo-interpreters to identify key landmarks in higher resolution images. It could also be used for mapping.

After a GAMBIT SRV reentered and was caught in mid-air by a C-130 aircraft, the film would be transported to a processing facility, usually at Eastman Kodak in Rochester, although the Air Force also operated a facility in Maine. Duplicates were made and shipped to Washington, DC, where they were evaluated by photo-interpreters at the National Photographic Interpretation Center (NPIC). The photo-interpreters specialized on specific targets and during their initial evaluations they looked for any changes since the last reconnaissance mission had photographed the area. A GAMBIT photo could reveal a tremendous amount of information about a target. New and noteworthy information was quickly sent to the Pentagon and CIA headquarters via secure cable. The rest was summarized in reports prepared over several days or weeks.

For example, in late May 1967 the last KH-7 mission flew for a week over various targets. When it returned its film NPIC produced a mission report that mentioned what was seen at each target. The satellite made three passes over the Tyura-Tam missile test centre in Kazakhstan, photographing targets that the CIA had designated Complexes F and G and Launch Groups K, L, M, and N. Complex G was a large multi-use facility that included launch pads for the Soviet Proton and Tsyklon launch vehicles. NPIC had designated the Tsyklon pads as G5 and G6. The mission report stated:

The 4-legged apparatus is positioned on pad G5 and a probable transporter and a long vehicle/piece of equipment are nearby. At least 8 vehicles are parked between the pads and one long vehicle/piece of equipment is on pad G6. Approximately 45 cylinders are positioned along the east fence line and the previously reported light-toned possible missile components are in the same position along the south fence line. The gantry is back near the end of its track.

Each mission report listed similar information for hundreds of separate targets, a deluge of data. On a regular basis, or by request, NPIC would produce detailed site reports for individual targets. For instance, starting in 1965 every October NPIC produced a report on the massive Complex J that served as the assembly and launch site for the N-1 lunar rocket, which the CIA had designated the “J vehicle” and which was designed to carry cosmonauts to the Moon.

All of these reports were incorporated into the highest level intelligence documents, the National Intelligence Estimates, or NIEs.

The last KH-7 GAMBIT was launched on June 4, 1967. Mission 4038 flew an eight-day mission. But long before this happened, the KH-7’s successor was already under development.

Titan launch
An early KH-8 GAMBIT satellite launch. The KH-8 initially had only one satellite recovery vehicle. A second was added later. (courtesy Jonathan McDowell)

First light for the new bird

On July 29, 1966 a Titan 3 launch vehicle lifted off of its pad at Space Launch Complex 4 at Vandenberg Air Force Base. The tall and powerful Titan 3 carried a thinner Agena D and its cylindrical payload into the sky, looking much like a ballpoint pen aiming for the stars. This highly classified launch pushed the first KH-8 GAMBIT into orbit. Known as mission 1701, the spacecraft entered a 98 by 155 nautical mile (181 by 287 kilometer) orbit at 94.12 degrees inclination. Seven days later its single Satellite Recovery Vehicle (SRV) returned to Earth. This new spacecraft was also named Air Force Program 110 and although many of its components were new, it was essentially a direct upgrade of its predecessor, the KH-7. Today intelligence officials refer to the “GAMBIT program” as a whole, but the upgrades to the camera were clearly sufficient to justify a change in the KEYHOLE designation.

Although the KH-8 camera modifications remain unknown, what is known is that major changes were made to the spacecraft.

In 1964 the Air Force had initiated a GAMBIT upgrade program. One declassified document appears to indicate that at this time the NRO planned on improving the absolute resolution of the camera by 20%. Because little is known about the resolution of the KH-7 other than its initial and final capabilities, as opposed to its theoretical limits, it is impossible to determine what the resolution goal was for this upgrade. One CIA critic of the plan noted that although such an upgrade was possible, so many things affected satellite resolution that improvements in the theoretical limits of a camera system would not automatically correspond to an equivalent improvement in the resolution returned by the photographs.

According to one person familiar with the KH-8, the primary mirror was not increased in size for the KH-8. Back in summer 1963 the CIA had initiated a special study of the requirements for future search satellite systems. That study, known as the Purcell Panel after its chairman, Dr. Edward M. Purcell, had evaluated the future of very large optics for reconnaissance purposes and determined that mirrors as large as 60 inches (152 centimeters) in diameter could soon be developed. By the late 1960s optics firm Perkin-Elmer had developed 72-inch (183-centimeter) mirrors for the KH-10 DORIAN camera system scheduled for the Manned Orbiting Laboratory (MOL). But at least for the switch from the KH-7 to the KH-8 the camera apparently still maintained is 44-inch (112-centimeter) diameter primary mirror and 77-inch (196-centimeter) focal length.

Although the KH-8 camera modifications remain unknown, what is known is that major changes were made to the spacecraft. Probably around the same time that the NRO sought an improvement in camera resolution, the agency initiated a contract competition to select an upgrade to the Program 206 (KH-7) spacecraft.

General Electric apparently proposed a modified version of its existing OCV spacecraft. Lockheed Missiles and Space Division proposed using a modified version of its versatile Agena upper stage. Both would support the upgraded camera system. The requirements of this upgraded spacecraft are unknown, but probably predictable—longer operational lifetime, higher stability and pointing accuracy, greater film load, and higher reliability.

Lockheed’s program manager, Robert M. Powell, proposed using the Agena upper stage equipped with a roll joint at its forward end and an electric motor. This enabled it to rotate the payload to point the camera to either side of the ground track without expending control gas to do so. The idea was not new. In fact, the company had developed a roll joint to accomplish the same task on the panoramic KH-6 LANYARD, which had flown three times with one success. With the General Electric OCV no longer needed, much of its mass and volume could be traded for other things, such as greater film capacity and more control gas, both of which had been limited with the KH-7.

Determining the physical characteristics of the KH-7 and the KH-8 is difficult. One independent analyst has calculated that the KH-7, minus its recovery vehicle, probably had a mass of around 4,400 pounds (2,000 kilograms). How much of this was devoted to the camera system and its film is unknown, but eliminating the OCV would not eliminate the requirement for many of its systems such as batteries, control gas, and some structure. Such a change might have only freed up a few hundred extra pounds of mass.

Surprisingly, the upgrade to a Titan 3 Agena D launch vehicle—which lacked the massive solid rocket boosters common to other variants of the Titan 3—apparently added only a few hundred extra pounds of payload capability. A very rough guess is that the changes may have only freed up only about 800 pounds (363 kilograms) to use for other purposes. Years after the upgrade, the satellites could carry far more film as well as a second reentry vehicle, but for now the initial switch from the KH-7 to the KH-8 remains somewhat mysterious.

Lockheed won the contract for what was soon designated the KH-8 GAMBIT. General Electric had lost the Samos E-6 contract the previous year and now it had lost the GAMBIT. Lockheed’s win had the effect of further securing the company’s position as the primary manufacturer of reconnaissance satellites for the United States.

With the Agena capable of providing all of the functions of the General Electric Orbital Control Vehicle, General Electric was soon demoted from building the entire GAMBIT spacecraft and integrating all of its parts to manufacturing the GAMBIT reentry vehicle. Instead of prime contractor on a major reconnaissance project, GE was now a supporting contractor.

Ellis Lapin felt that the early GAMBIT missions that had kept the vehicle attached to the Agena had established that the Agena could perform the basic mission. The roll joint had made the choice obvious. He remembered that soon after the decision he was at a meeting of GAMBIT officials. “Logan Cowles, the GE program manager, and Jack Katzen, his assistant and I discussed the news. Jack commented ‘How can one compete a complete satellite against a couple of pulleys and some rubber bands?’” Of course the spacecraft did not really use pulleys and rubber bands, but the simplicity of the roll joint solution was obvious. The NRO was already paying for the Agena, so why pay for an additional spacecraft as well?

According to Lapin, the additional payload mass that was made available by the rocket upgrade and elimination of the OCV was used for a modified film transport, more expendables such as control gas and batteries, and Agena propellant. It was not used for a bigger primary mirror to increase resolution.

Some people in the Air Force were apparently concerned about the ability of the Agena to conduct this task, and in 1966 they asked a civilian official, the Air Force Chief Scientist, Bob Cannon, to review their plans. Cannon determined that there were no problems with the proposal.

Early (1959) photo of the General Electric built Satellite Recovery Vehicle that was initially designed for the CORONA satellites and later adapted for the KH-7 and KH-8 GAMBIT satellites. The recovery vehicle had to hold a roll of nine-inch-wide film for the GAMBIT satellites. Note the engineer holding the small "life cell" containing four mice. The mice later died on the launch pad. (credit: NRO)

Program 110

Although the KH-8 apparently did not have a larger primary mirror larger than the KH-7, various technological developments probably enhanced the GAMBIT’s resolving power. For instance, during the mid-1960s the Itek Corporation, which by 1966 had been shut out of developing new reconnaissance cameras for the Central Intelligence Agency, had an active optical research and development program. One of the projects that Itek worked on was a means to accurately determine the amount of image motion compensation (IMC) needed for a camera. According to the head of Itek’s R&D department, early reconnaissance spacecraft had been sent into orbit with several possible settings for IMC. Once ground controllers knew the precise altitude that a satellite had reached, they would then select the proper IMC setting for that orbit. But if the amount of image movement could be measured inside the camera while it was in orbit, then more precise IMC could be applied to the camera, reducing image smear even more. Whether or not Itek’s solution to this problem was incorporated into the KH-8 version is unknown, but it seems reasonable that precise control of the film through the camera was a good way to reduce image smear and increase sharpness and resolution.

Undoubtedly other developments, like improved optical coatings, also contributed to the KH-8’s better resolution. Whereas the KH-7 GAMBIT had started at 4-foot (1.2-meter) ground resolution and by 1966 had improved to 2-foot (0.6-meter) resolution, the KH-8 was reported to be far better. Several sources have stated that the KH-8 achieved an average ground resolution of around six inches (15 centimeters), but this may not have occurred until many years later. It seems more likely that at least initially the KH-8’s resolution was equal to or slightly better than the 2-foot resolution of late model KH-7 GAMBITs.

Although the satellite’s average resolution was apparently six inches, several sources have stated that the KH-8 eventually achieved accuracy as great as 2.5 inches (6.3 centimeters), and one source has indicated that the KH-8 was able to resolve golf-ball size objects.

In 2000 the NRO gave awards to forty-six pioneers in satellite reconnaissance. Two of these were apparently associated with the KH-8: Colonel Lee Roberts and Colonel James T. Mannen, both of whom started their careers in national reconnaissance in 1971, five years after the KH-8 entered service. Mannen’s citation read: “As director of a vital imagery satellite program, Colonel James Mannen introduced procedures that improved target tasking and significantly increased ground resolution and on-orbit system reliability.” Roberts’ citation read: “Colonel Lee Roberts directed improvements in an important Program A satellite reconnaissance effort that produced high-resolution imagery of the Earth’s surface.”

Although the satellite’s average resolution was apparently six inches, several sources have stated that the KH-8 eventually achieved accuracy as great as 2.5 inches (6.3 centimeters), and one source has indicated that the KH-8 was able to resolve golf-ball size objects. This is still insufficient to read license plates or newspapers, or to recognize faces, but it is far better than commercially available imagery today.

Current American reconnaissance satellites reportedly have mirrors far larger than the 44-inch (112-centimeter) mirror carried by the KH-7, but do not have the same resolving power as the KH-8. The reason is that current American spy satellites stay in orbit for years and in order to do this they must operate in higher orbits so as not to be dragged down by the tenuous upper atmosphere. The KH-8 was a disposable satellite and it achieved its high resolution by descending to a low perigee over its targets, but then had to be reboosted so that it did not reenter the atmosphere.

The NRO sought to develop a quick reaction capability for the KH-8, as it had for its predecessor. In September 1969 John McMahon, the CIA’s Deputy Director for Science and Technology, wrote the CIA’s Deputy Director for Plans concerning satellite scheduling for crises situations. The intelligence community had already learned a harsh lesson about the limitations of satellite scheduling during August 1968, when Soviet troops invaded Czechoslovakia. A CORONA satellite returned the second of its SRVs to Earth and its photos clearly showed Soviet troops massing for an invasion. But the film was not recovered and processed until after the invasion had already occurred, an event that had shocked the intelligence community.

McMahon noted that neither the CORONA nor GAMBIT “provides more than one payload at a time which could be used in advance of a scheduled launch date.” He also pointed out that processing times—including the time to transport the exposed film to the processing center and then the photo-interpreters—also imposed delays. “As an example, the minimum time from the taking of a photograph in China to viewing of the film is about two to two and one-half days.” Another problem was that the KH-8 had to have a pre-selected orbit before launch, with the targets upon which the orbit was selected provided 17 days prior to launch. CORONA was much more flexible and could be targeted on-orbit from a general-purpose orbit.

If the intelligence community was going to get more timely intelligence about what an adversary was planning, what they classified as “indications and warning intelligence,” they would need to change from the recoverable film technology to something better. This became one of the major quests of the satellite reconnaissance community during the 1960s.

page 2: FROG >>