The Space Reviewin association with SpaceNews

Titan 3D image
A Titan IIID rocket being prepared to launch the first KH-9 spysat in 1971. (credit: USAF)

The flight of the Big Bird (part 1)

The origins, development, and operations of the KH-9 HEXAGON reconnaissance satellite

Bookmark and Share

In summer 1971 a Titan IIID rocket stood poised on its launch pad on the picturesque central California coast. It was the largest rocket that Vandenberg Air Force Base had ever seen, and in its payload shroud was a brand new, highly-classified satellite. Those on the base had started to refer to the rocket and its payload as “the big bird.” For decades later, whenever articles appeared in newspapers and magazines about this particular type of satellite, reporters referred to it as Big Bird. It had followed a tortuous path to the launch pad.

The story of the KH-9 HEXAGON is a complex one, more complex than the programs that preceded it. In the early days of satellite reconnaissance the technical challenges were great, but the intelligence goals were in many ways simpler.

This was actually the first KH-9 HEXAGON reconnaissance satellite, and over the next decade and a half it would be followed by nineteen more. It was indeed a massive payload with a unique mission. HEXAGON was intended to replace the venerable CORONA series of spysats in the vital role of conducting searches of large areas of the Soviet Union. The HEXAGON had the ability to see objects on the ground—known as resolution—significantly better than the CORONA. It achieved this with what people familiar with it say remains the most sophisticated mechanical device ever placed in orbit. Even though the last HEXAGON was launched in the mid-1980s, and its technology is long obsolete, it remains classified.

But that has begun to change. Some information on HEXAGON, including the KH-9 designation (but not the code-name) was officially released in 2002. In recent years, the National Reconnaissance Office (NRO), which operated the HEXAGON (the CIA paid for the camera), has declassified even more information on the program, such as the dates of major milestones and the involvement of key personnel. In many cases the information the NRO released has merely confirmed what was leaked about the program over the years, or has been speculated by independent observers. But some of this declassified information is indeed new, and may signify that the US government may be close to declassifying the HEXAGON. If this happens, the public will finally learn about this venerable spysat, which those involved with it believe was an amazing example of American technological innovation and ingenuity.

What follows is an admittedly sketchy and incomplete history of the development of the KH-9 HEXAGON, based heavily upon a declassified official history of the program as well as other declassified documents and some published works. This remains a draft, and will serve as the basis for a more detailed, and footnoted, print article to be published in the next year. Even that published version will remain a draft; until the program is formally declassified, it will not be possible to fill in many of the still-gaping holes in the HEXAGON story. But we are closer today than ever before.


The story of the KH-9 HEXAGON is a complex one, more complex than the programs that preceded it. In the early days of satellite reconnaissance the technical challenges were great, but the intelligence goals were in many ways simpler. For the United States’ first successful reconnaissance satellite, CORONA, which had its first success in summer 1960, the most basic goal was to launch a camera into space, take pictures at reasonable resolution—then defined as essentially anything better than about 9 meters (30 feet)—and return them to Earth. Simply getting something into orbit and back was tough, and accomplishing that was the objective; the satellite and its camera system could be refined later. For the next recoverable satellite, the GAMBIT, first launched in summer 1963, the goal was to produce significantly higher resolution, roughly good enough to provide technical intelligence about the objects that the camera saw. This called for resolution around 0.9 meters (three feet) or better. One of CORONA’s highest priorities was finding Intercontinental Ballistic Missile (ICBM) sites inside the Soviet Union. One of GAMBIT’s highest priorities was determining if there was an actual ICBM on the launch pad at those sites. Both systems returned film to the ground where it had to be developed before it could be viewed, and this meant that on average the photographs were at least several days old by the time they were viewed by a pair of human eyes.

HEXAGON was started when these early goals had been met and intelligence officials started asking about what they could accomplish with a reconnaissance satellite and what value would it provide. There was not clear agreement on these questions. Furthermore, HEXAGON emerged at a time when various intelligence officials were feuding over control of this powerful new intelligence tool.

In the late 1960s the late Robert Perry began writing a series of histories of American satellite reconnaissance programs for the NRO. Perry wrote a history of the KH-9 that was finished in the early 1970s, soon after the program entered service. According to Perry, HEXAGON resulted from “two years of variously intense controversies about requirements, schedules, technology, and organizational imperatives.” That was a mild way of putting it. The origin of HEXAGON, to the extent that it can be understood now, before the program is fully declassified, lies in a murky soup of high-level studies of reconnaissance technology by the nation’s premier optics and engineering experts as well as various proposals for new reconnaissance satellites pushed by aerospace and camera contractors. At the moment, it is extremely hard to separate out the relevant details from the irrelevant ones. But any discussion of HEXAGON has to begin with its immediate predecessor.


The first successful reconnaissance satellite of any kind was known as CORONA. CORONA was a CIA project that was substantially supported by the US Air Force. The Air Force attempted to launch the first preliminary payload in the program in early 1959. But the program suffered a long string of failures—13, in fact—before achieving its first success in summer 1960. CORONA’s record of successes improved gradually—25% successful in 1960, 50% successful in 1961, and 75% successful in 1962.

CORONA was a brilliant bit of engineering, but also a compromise.

The US Air Force had started a reconnaissance satellite program in the mid-1950s. When the CORONA came along, the CIA became involved. The CIA procured the camera payload and the Air Force procured everything else—the rocket, spacecraft, recovery aircraft, and ground control systems. This ad hoc relationship worked well at first. But by 1961, senior Department of Defense and CIA officials sought to formalize the arrangement with the creation of a new, highly secret, DoD organization named the National Reconnaissance Office, or NRO. But even after the NRO was created, CIA and Pentagon officials did not agree on its authority.

CORONA was a brilliant bit of engineering, but also a compromise. CORONA’s designers at the upstart camera company Itek, located outside of Boston, Massachusetts, had combined a reasonably powerful set of lenses—60-centimeter (24-inch) focal length—with a new way of exposing film by sweeping the image across a long film strip. This created an image that had decent resolution—the ability to resolve objects on the ground that were reasonably close together—with tremendous area coverage. Each sweep of a CORONA camera photographed a massive amount of ground underneath, about 16 kilometers (10 miles) north to south, but 190 kilometers (120 miles) wide. The camera would sweep back and forth, “mowing the lawn” over the Soviet Union, capturing significant amounts of territory and revealing what was being built there—as long as the objects on the ground were relatively big, like missile sites. Trucks and tanks would barely show up on CORONA photos.

But the CIA had sort of backed into this design. The CIA wanted decent pictures of the Soviet Union, and were happy with anything that they got. The first photographs were really poor, with resolution around 30 feet. Itek’s engineers kept tweaking each camera before it was launched and the imagery kept improving, until it was around 4 meters (13 feet) by 1963. But the project had not started with the CIA clearly defining requirements of what they needed to see and what CORONA needed to produce. The agency’s leadership only started to ask those questions after CORONA started flying.

The initial CORONA (KH models 1, 2 and 3) had only a single camera. Itek soon added a second camera in a system known as the KH-4 MURAL, but later primarily referred to as the CORONA MURAL. Both cameras tilted inwards slightly, enabling one camera to take pictures forward of the spacecraft at one angle, and a few moments later the other camera would photograph the same area, now behind the spacecraft, at a slightly different angle. This enabled photo interpreters to take measurements of objects on the ground and determine their height.

In March 1962 the CIA endorsed a radical upgrade to CORONA known as the M-2 search system (for CORONA-MURAL-2). This involved the use of a single 100-centimeter (40-inch) f3.5 lens and a dual-platen film system (the platen is the surface that the film lays on during exposure). The system offered to return broad-area photography with a resolution of 1.2–1.5 meters (4–5 feet). The M-2 was formally presented for NRO review on July 24, 1962.

CORONA already had some competition in the form of the Air Force’s Samos E-6 system. Had it been successful, the E-6 might have developed an honorable legacy for itself. It was a big system with a long, pointy reentry vehicle. The Samos E-6, with two 36-inch focal length cameras, could provide nominally better coverage than CORONA. But by late 1962, the difference between improved CORONAs and the E-6 had narrowed. The primary difference was that CORONA had been sponsored by the CIA, and the Samos E-6 was sponsored by the Air Force. Also, CORONA worked.

CORONA was working, although not consistently, which put it far ahead of any of the theoretical replacements. But there were a lot of people involved in the reconnaissance program who viewed CORONA as an inherently limited design.

Samos E-6 had its own problems. The first two Samos E-6 missions had serious camera system malfunctions and had suffered recovery failures. In July and August 1962 the third and fourth missions also failed. By October the NRO’s Director of Special Projects—the Air Force component of the NRO—Air Force Major General R.E. Greer, and NRO Director Joseph Charyk suspended purchases of further systems. The fifth E-6 was damaged during reentry and sank in the Pacific, although its camera may have operated successfully in orbit. There was a proposal for an “improved” Samos E-6 known as SPARTAN. SPARTAN would have eliminated one of the E-6’s two cameras and used the proven CORONA reentry vehicle instead of the troublesome E-6 version. But although details on the E-6 and SPARTAN remain sketchy, it appears as if the only advantage it really had over CORONA was that the CIA was not involved—this at a time when many in the Air Force reconnaissance program were hoping to push the agency out of satellite reconnaissance completely.

Yet another updated CORONA, known as the CORONA-J, entered engineering design in July 1962, with first launch scheduled for May 1963, although first launch did not occur until August 1963. The CORONA-J introduced a number of improvements to the camera design. The primary upgrade was the addition of a second reentry vehicle, enabling the satellite to stay in orbit longer and take more pictures. In short, CORONA was working, although not consistently, which put it far ahead of any of the theoretical replacements. But there were a lot of people involved in the reconnaissance program who viewed CORONA as an inherently limited design.

Not much is known about Itek’s M-2 configuration. Although it offered to substantially improve resolution, it still retained the large moving parts that made a number of reconnaissance experts uneasy. One reason that M-2 was not approved at this time was that there was no stated requirement for a relatively high resolution search system.

In January 1963, the National Reconnaissance Office’s Directorate of Special Projects awarded a study contract to Eastman Kodak. Although the details of this study are unknown, it involved some kind of high-resolution goal. Soon this resulted in Kodak proposing a high-resolution satellite camera that would be launched by a Titan IIIC rocket, then the most powerful rocket in the Air Force’s inventory. The proposal was designated “S-2,” for “Search 2.”

page 2: Purcell and Drell >>