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

<< page 1:CORONA

Purcell and Drell

In spring 1963 the Director of Central Intelligence created the Panel for Future Satellite Reconnaissance Operations, headed by Edward Purcell. The Purcell Panel, as it became known, was briefed on the various technologies that people in the optics field believed could have a major impact on the development of reconnaissance satellites.

Land believed that what was really needed was a system that had CORONA’s territorial coverage and GAMBIT’s high resolution, which initially started at 1.2 meters (four feet), but eventually got as good as 0.6 meters (two feet).

The CIA director’s decision to create the panel may have been prompted by the proliferation of new proposals for reconnaissance systems. In spring and early summer 1963 CIA reconnaissance specialists proposed two alternatives to the M-2. One was a covert vehicle “that could be represented to be something other than a reconnaissance vehicle,” in the words of historian Robert Perry. The idea was a rather bizarre one: disguise a reconnaissance satellite as a prototype orbiting nuclear weapon. To date this proposal is only sketchily described in declassified histories, but it is beyond head-scratching. After all, many of the people who were involved in developing reconnaissance satellites were very concerned that the Soviet Union might start to shoot them down and wanted the Soviet Union to ignore them. Painting a big fat radiation sign on the side of one was not exactly a way of de-escalating the situation.

The second concept “was even more controversial,” according to Perry. Unfortunately, what exactly this was has not been declassified. The idea was passed on to the Purcell Panel, which determined that it “would not be a wise investment of resources.”

In its July 1963 report, the Purcell Panel recommended that the CIA (and NRO) should focus most of its attention on getting consistent quality from the existing CORONA satellites rather than pursuing new capabilities. The panel identified several factors that were leading to inconsistent quality on the CORONA missions and said that they could all be fixed or improved. The panel also identified several other potential technology improvements such as beryllium mirrors with diameters up to 150 centimeters (60 inches). But they concluded that the primary goal should be to improve what was already working before starting new projects.

Arthur Lundahl, the head of the National Photographic Intelligence Center where experts looked at photographs and counted missiles and bombers and submarines, had earlier told the Purcell Panel that there was really no difference between the intelligence that an interpreter could derive from 3-meter (10-foot) resolution (what CORONA was essentially seeing) and that he could derive from 1.5-meter (5-foot) resolution imagery (which the proposed M-2 could achieve). Thus, substantially improving the CORONA to provide approximately 5-foot resolution did not make much sense.

But one member of the committee, Edwin “Din” Land, who ran Polaroid and had served as a senior intelligence advisor on reconnaissance systems for many years, reached a different conclusion. Land believed that what was really needed was a system that had CORONA’s territorial coverage and GAMBIT’s high resolution, which initially started at 1.2 meters (four feet), but eventually got as good as 0.6 meters (two feet).

As Robert Perry wrote in a history of this period, the Purcell Panel was criticized by some because it was not considered bold enough. However, Perry noted that this came at a time when the Samos E-1, E-2, E-5, and E-6 had all failed, and the KH-6 LANYARD, which used the modified Samos E-5 camera in an attempt to achieve better resolution than CORONA, was also suffering problems. In the context of all these failures, the panel’s recommendation to perfect what was already working seemed reasonable.

But CIA and NRO officials interpreted the Purcell Panel differently.

As a result of the panel’s recommendations, the new Director of the NRO, Brockway McMillan, ordered cancellation of M-2 work at Itek in July 1963. The M-2 concept reappeared infrequently over the next several years, under several different names, including CORONA J-4, but it never became anything more than a contractor proposal.

What the Drell Committee’s tests indicated was that CORONA’s resolution—then around 3.6 meters (12 feet), was inadequate for the kinds of information that they needed.

The CIA, however, took a different approach, largely because of the views of a CIA newcomer, Albert “Bud” Wheelon. Wheelon took over as head of the agency’s stagnating science and technology office around the same time that McMillan became director of the NRO and he quickly started building it into a major directorate. He did this in numerous ways, including obtaining special hiring authority that allowed him to pay highly skilled engineers good salaries to work for him—something that the DoD-based NRO simply could not match. Wheelon started to think big. He wasn’t interested in simply fixing the problems in existing satellites, he wanted to build new satellites. But before he started doing that, he realized he needed to establish both requirements and boundaries. In October 1963 Wheelon created the Drell Committee, named after its chair, Stanford University professor Sydney Drell, “to explore the whole range of engineering and physical limitations on satellite photography…”

One of the things that the Drell Committee did was to test photo-interpreters by giving them aerial images of different quality and asking them to determine what they saw. By this method they were able to determine how much intelligence data could be gleaned from a satellite image with say, 0.6-meter (2-foot) resolution versus one with 1.8-meter (6-foot) resolution.

NRO Director McMillan did not learn about the Drell Committee until after it was established and thought that its charter was too broad. He also argued that Wheelon had no official role in the satellite program. Wheelon and McMillan were not on good terms to start, and Wheelon simply ignored him.

What the Drell Committee’s tests indicated was that CORONA’s resolution—then around 3.6 meters (12 feet), was inadequate for the kinds of information that they needed. However, camera specialists believed that in order to get resolution much better than 2.1 or 2.4 meters (7 or 8 feet) for about 50% of the returned film, the original CORONA would not be sufficient. In other words, the requirement for better intelligence about targets detected in satellite images demanded a more powerful satellite, rather than upgrades to existing ones. (Somewhat ironically, these unnamed experts were actually wrong about CORONA, because the CORONA system eventually achieved this goal.)

Drell’s group apparently approached the resolution question differently than the Purcell committee. The earlier group had essentially looked at the existing system—CORONA with its (at best) 3-meter (10-foot) resolution—and its then-proposed upgrade, CORONA M-2 with its 1.2- to 1.5-meter (4- to 6-foot) resolution—and concluded that the upgrade would not provide better intelligence. In contrast, the Drell group, at Wheelon’s urging, looked at what kind of intelligence could be asked at five different resolution levels. Although the resolution levels that they evaluated remain classified, it seems likely that they ranged between a high of (0.3 meter) 1-foot ground resolution and 3 meters (10 feet). The magic number seems to have been around 0.6-meter (2-foot) ground resolution. At that point it became possible to determine substantially more information about the objects seen in satellite photographs. Why that number was special remains unclear, but it was probably tied to the dimensions of Soviet missiles, which were then the highest priority targets for American intelligence collection. If you can measure the length and width of a missile to relatively high precision you can determine what its fuel load is, and if you know the fuel load, you can begin to calculate how far the missile can fly, and if it can hit the United States.

Drell’s group also looked at another question: how much could CORONA be improved? They concluded that it was essentially reaching its limit. As Jeffrey Richelson wrote in his 2001 book The Wizards of Langley, Bud Wheelon said that the two answers led to the conclusion that “something a lot better was needed.”

This was, however, Wheelon’s conclusion. Neither the Purcell or Drell Committees had recommended the development of a new search satellite. Wheelon’s people soon approached experts at both Itek and optics company Perkin-Elmer about their ideas for developing a new system and chose Itek over Perkin-Elmer, which at that time had no experience with reconnaissance satellite cameras.

On November 18, 1963, the NRO’s West Coast directorate contracted with a still-classified contractor to conduct feasibility studies, presumably about producing a higher resolution search satellite. Two months later the CIA separately authorized a separate study with Itek, but with more ambitious design goals. Itek was funded to conduct preliminary development work under a program code-named FULCRUM.

Bureaucratic wars

In December 1963, a DoD official, Eugene Fubini, proposed to Director of Central Intelligence John McCone that the CIA assign total CORONA responsibility to the NRO in return “for a free hand in development of a new search system.” But Brockway McMillan was apparently unaware of the offer and rejected it when he learned about it. McMillan was in a bind, because he lacked much authority in his position as Director of the NRO as long as the CIA handled all the contracts for the successful CORONA system, and the CIA leadership chose to essentially ignore him. In a May 1964 meeting the President’s Foreign Intelligence Advisory Board (PFIAB)—the nation’s highest level group of intelligence advisors—recommended giving McMillan more power, but a presidential directive to this effect was never signed. McMillan still lacked the power he needed to take more forceful control over development of new reconnaissance systems.

At a meeting of an intelligence advisory committee, a senior CIA official stated that the 60-degree off-axis requirement—which Itek officials thought was unnecessary—had been an Itek recommendation. Levison viewed this as an outright lie and a breach of trust.

The bureaucratic skirmishing continued, with Wheelon and McCone proposing that CIA responsibility for development and operation of the new search system should be formally confirmed. According to Perry, the CIA also forbade “contractors to release information about their progress to any agency other than the CIA” thus keeping the NRO in the dark. They also proposed establishing an internal project office, “providing technical support and serving as system integrating contractor.”

Repeating the earlier experience with the Drell Committee, it was not until late June 1964 that Brockway McMillan even learned about FULCRUM. After he found out about FULCRUM, he sought to keep it from progressing to full-scale development. He did not want the CIA developing reconnaissance hardware considering that, in his view, the NRO had been created in 1961 to lead that effort. He sought to get an agreement between senior DoD and CIA officials that the agency would only study FULCRUM’s feasibility while the NRO conducted comparative studies. However, he was apparently outmaneuvered by Wheelon, who sent him a letter indicating “the various tasks for which we require immediate NRO funding.” That list included funding not only for the spacecraft, but also the rocket to launch it into space and “assembly, integration, and checkout” contracts. While McMillan wanted to confine the CIA to studying future systems, Wheelon was already sending him the bill for hardware.

While McMillan sought to convince the President’s Foreign Intelligence Advisory Board that the summer 1963 Purcell panel had only recommended perfecting CORONA, he also started his own alternative to the CIA’s FULCRUM. The NRO signed a contract with Eastman-Kodak to begin studying a “fairly conventional” system designated the S-2. Apparently the study also included a smaller system designated MATCHBOX. By that time Kodak was an experienced manufacturer of reconnaissance cameras. It had built the failed Samos E-1, E-2 and E-6 systems. But more importantly, it was regularly manufacturing the KH-7 GAMBIT cameras.

Throughout 1964, Itek continued designing FULCRUM. But the company’s relationship with the CIA was deteriorating. The CIA’s reconnaissance experts had imposed a requirement on the system that it be able to image from horizon to horizon—essentially sixty degrees to the left or the right of the satellite’s path above the Earth. But the farther a camera looks to either side, the greater distance it is looking toward the ground, thus degrading performance, meaning that the imagery very far “off axis” would be essentially useless.

The head of Itek’s camera development, Walter Levison, who had previously developed the CORONA camera, thought that the CIA’s requirement was pointless, and not worth the effort of developing a system capable of scanning the full 120 degrees. This technical argument apparently raged for some time. But in February 1965 the relationship completely fell apart. According to Richelson, at a meeting of an intelligence advisory committee attended by Wheelon, McMillan, Levison, Itek’s president Frank Lindsay, and others, a senior CIA official stated that the 60-degree off-axis requirement—which Itek officials thought was unnecessary—had been an Itek recommendation. Levison viewed this as an outright lie and a breach of trust. He recommended to Lindsay that they needed to cancel their contract with their biggest customer. Lindsay agreed.

The Itek decision sent shock waves through the CIA and the NRO, although NRO officials who had been annoyed at the CIA’s independence and arrogance were also slightly amused that it had finally bit them in the ass. An emergency meeting involving Levison, Itek’s FULCRUM program manager John Wolfe, McMillan, and Din Land was called. According to Richelson, Levison told them that Itek officials had concluded that they could not operate under “the domination of the CIA” and that the agency had created an “immoral environment.”

CIA officials suspected that they had been undercut by the NRO, which had promised Itek work on the newly-approved massive reconnaissance camera being developed for the Manned Orbiting Laboratory and eventually designated the KH-10 DORIAN. They quickly sent personnel over to Itek to remove all FULCRUM project records and drawings, equipment, as well as a scale working model of the camera.

Now they had to figure out what to do for their next step.

Next: Perkin-Elmer and AQUILINE