Early American signals intelligence satellites

Dwayne Day

June 21, 2016

The wizard war in orbit (part 1)

Tales of espionage are filled with lanky men in trenchcoats walking through cold Berlin streets at the height of the Cold War. But the most important intelligence—in terms of volume and reliability—was gathered by reconnaissance satellites far overhead. These satellites were precise, they collected vast amounts of information, and unlike spies, they did not forget, embellish, lie, or go rogue. Photographic reconnaissance satellites like CORONA, GAMBIT, HEXAGON, and KENNEN were in many ways the most prolific spooks. But they were also accompanied by other satellites, signals intelligence, or SIGINT, satellites that listened for the electronic whispers of radars and radios, engaged in a high-tech war of electrons against an enemy that could vanish and emerge at will.

During the Cold War the United States intelligence community gathered signals intelligence from the Soviet Union via a variety of means. These included ground stations, cable-tapping and bugging operations, airborne platforms such as the RC-135 Rivet Joint and RB-47 Stratojet, and signals intelligence satellites. Any history of SIGINT satellite operations during the Cold War is going to be limited in scope because much of the story remains classified, and unlike the reconnaissance photographs, signals intelligence is an arcane and esoteric subject.

What the history reveals is that SIGINT satellite operations by the United States during the 1960s were far more complex than independent observers ever imagined.

In 1998, the National Reconnaissance Office (NRO), which develops and operates intelligence satellites, declassified the first signals intelligence satellite named GRAB, which was launched in April 1960. GRAB was developed by the Naval Research Laboratory (NSL) in Washington, DC, and part of the NRL was later subsumed as a component of the super-secret NRO. In 2007 the NRO declassified the follow-on to GRAB, named POPPY. But both of these declassification actions were limited, leaving out many major details such as the appearance of some of the satellites, the variety and types of signals they collected, and even how long they operated. The NRO released further details in dribs and drabs over the next several years, but GRAB and POPPY operations remained shrouded in a certain amount of mystery and confusion. But while the GRAB and POPPY revelations were significant, the reality was that they represented only a small part of the story. Throughout the 1960s the NRO operated many other SIGINT satellites and platforms, most developed by the US Air Force, and these remained shrouded in secrecy. Until now.

Starting last year, the NRO declassified portions of its official history of signals intelligence satellites. Titled The SIGINT Satellite Story and written in 1994, it was produced by four authors: Major General David D. Bradburn, US Air Force (retired); Colonel John O. Copley, US Air Force (retired); Raymond B. Potts, National Security Agency (retired); and one other retired National Security Agency employee whose identity remains classified. One chapter of the history tells the story of the GRAB and POPPY satellites. Two chapters in The SIGINT Satellite Story explain much of the history of the other low Earth orbit satellites operated by the NRO during the Cold War. What the history reveals is that SIGINT satellite operations by the United States during the 1960s were far more complex than independent observers ever imagined, and the NRO operated satellites with names like MULTIGROUP and STRAWMAN, and numerous SIGINT payloads with names like NEW JERSEY, DONKEY, and OPPORKNOCKITY. Although other parts of this effort remain classified, over the years details have also emerged about a class of small SIGINT satellites with names like FARRAH, RAQUEL, and URSULA.

Perhaps the most startling revelation was just how diverse and prolific these projects were. The spooks were lofting signals intelligence collectors into orbit at every opportunity in a major effort to find Soviet radars, figure out how they operated, and even intercept communications between air defense operators and Soviet fighter pilots. Much of their focus was on Soviet anti-ballistic missile radars as well as Soviet air defense radars so that the United States could develop countermeasures.

Samos Subsystem F

The first Air Force reconnaissance satellite program was started in 1956 and initially designated Weapons System 117L. After Sputnik it was split into several programs, the largest of which was named “Samos.” Samos quickly expanded to encompass several projects including an increasing number of different photographic payloads designated E-1, E-2, and so on. It also included a number of signals intelligence payloads designated F-1, F-2, F-3, and eventually F-4. The letters had been assigned sequentially based upon the major subsystems in the program— Subsystem F, also abbreviated as “S/S F”, followed Subsystem E in the sequence. “Ferret” was a term applied to aircraft used to collect electronic emissions because like the furry animal, ferret aircraft sometimes flushed out hidden prey—so the F designation was appropriate.

The main parts of the F-1 satellites were built by Airborne Instrument Labs at Mineola and Deer Park, Long Island, New York. They used components of an electronic intelligence receiver known as the AN/APR-9 and originally developed in 1948 for Air Force aircraft. Airborne Instrument Labs modified the equipment for space use. The F-1 was unique as the only vacuum-tube-type electronic intelligence payload ever flown by the United States, although the Soviet Union continued using vacuum tubes in many of their spacecraft for decades.

The more unusual aspect of the payload was that the F-1 shared the front of an Agena spacecraft with the first Samos camera system.

The F-1 payload covered a frequency range of 2.5 to 3.2 and 9.0 to 10.0 GHz. These frequencies included World War II era S-band and X-band area search radars. At the time, 100 to 400 MHz was used by higher power ground-based early warning radars, but the F-1 payload could not carry the larger antenna required. The F-1 was intended to detect the sidelobes of radar antennas rather than their main beams. Whereas the main beam of any electromagnetic emitter is the most powerful part of the emission, it is also focused and therefore the smallest part, even after spreading out over long distances. The sidelobes, on the other hand, spread out from the sides of the emitter like the petals of a flower, and although they are much lower power they are far bigger, making them easier to intercept.

The Samos F-1 was a rather clunky design. For instance, it had motor-driven mechanical cams for frequency scanning. But the more unusual aspect of the payload was that the F-1 shared the front of an Agena spacecraft with the first Samos camera system, which was designed to take photographs on a special film that was developed in orbit and then electronically scanned for transmission to the ground. Because real estate on the front of the Agena was limited, the F-1 S-band horn antenna was mounted in front of the Samos E-1 camera lens, which looked Earthward through a hole cut in the antenna. This design limited the field of view of the E-1 camera. The solution was to detach the F-1 by firing a squib after the twenty-first orbit, ejecting the F-1 antenna and providing the E-1 camera with a full field of view so that it could continue the photographic mission.

The first launch took place on October 11, 1960, when an Atlas Agena rocket lifted off from Vandenberg Air Force Base in California with the F-1/E-1 payload aboard. But the Atlas did not operate properly and both payloads were lost, falling into the Pacific Ocean.

On January 31, 1961, the Air Force tried again, launching Samos 2 from the same pad at Vandenberg. This time the Atlas operated like it was supposed to and placed the F-1/E-1 combination into the proper orbit. The spacecraft worked, with the F-1 collecting signals data and the E-1 taking low-resolution photographs of the ground. According to the official history, the F-1 operated long enough to produce 69 data interceptions. The spacecraft fired its squib on orbit 21 to remove the F-1 antenna. But something went wrong. The shock from the squib may have damaged part of the spacecraft, but the exact cause of failure was unknown. Ground controllers received no further communications from the spacecraft. The parasitic F-1 had killed its host.

Program officials canceled the third E-1/F-1 launch to save money and because the E-1 was a low-performance camera with no intelligence value. In April 1961 an Atlas carrying the more capable Samos E-2 photographic camera—but no signals intelligence payload—blew up on the pad and the Samos readout system was canceled. The third E-1/F-1 payload was placed in storage.

On December 23, 1963, project officials decided to use Thor rockets for launch of future F-2 and F-3 spacecraft. The signals intelligence payloads alone did not require the more powerful and expensive Atlas rocket to reach orbit, and improvements in electronics meant that the F-2 was less than two-thirds the weight of the F-1. The F-2 also covered more radio frequency spectrum using three frequency bands.

The original plan was to launch four F-2 and four more advanced F-3 satellites. Colonel Bill King, who had overall control of the Samos program, was concerned about the poor success rate of American launch vehicles at the time and thought that four launches of each payload would provide assurance that at least one of each would be successful. As it turned out there were no launch failures for any of the signals intelligence satellites.

Under the new plan using the Thors, the first launch was scheduled for February 1962. In March 1961 Joseph Charyk, the Director of the newly-created NRO, allocated $35 million for four F-2 launches in 1962 and four F-3s in 1963. In December 1961 the program was cut back to seven flights, four in 1962 and the remainder in 1963. An analog satellite payload was eliminated and an on-orbit programmer for the analog missions was also canceled, but the program remained largely intact.

The digital payloads were general search collectors and also provided electronic order of battle information—essentially classifying the type and location of Soviet radars so that this information could be plotted out for Strategic Air Command bomber crews who had to fly into Soviet territory.

The Subsystem F signals intelligence payloads were classified as Department of Defense “Secret” level with strict “need-to-know.” This was a lower level of classification than the CIA-led photographic reconnaissance program known as CORONA. CORONA was so highly classified—blacker than black—that even its existence was classified. In contrast, people working on other Samos satellite programs were allowed to know that Subsystem F existed, although they were not allowed to know any details unless they specifically worked on the satellite program. The spacecraft were given mission numbers that were part of the TALENT-KEYHOLE security compartmentation system.

Lieutenant Colonel Edwin J. Istvan of the Air Force Office of Missiles and Space staff at Los Angeles Air Force Base developed a system that identified the payloads according to their radio frequency bands and the output data. The frequency bands were numbered 1, 2, or 3. And they were designated digital or analog. For example, Group 2D was a digital system that provided radio frequency coverage from 0.059 to 0.130, 2.5 to 3.2, and 8.2 to 12.4 GHz.

TypeFrequencies

Group 1 Unknown

Group 2 0.059 to 0.130, 2.5 to 3.2, and 8.2 to 12.4 GHz.

Group 3 640 MHz to 8.28 GHz.

The digital payloads were general search collectors and also provided electronic order of battle information—essentially classifying the type and location of Soviet radars so that this information could be plotted out for Strategic Air Command bomber crews who had to fly into Soviet territory. The output was in the form of a 10-kilobit digital data stream. The digital payloads used frequency-sweeping superheterodyne receivers with lightweight solid-state components providing improved versatility and reliability. Electronic frequency scan and switching were also improved over the F-1’s electromechanical methods.

The first Samos F-1 signals intelligence payload mounted atop the first Samos E-1 photo-reconnaissance camera in 1960.

First launch of the Agena SIGINTs

The leftover F-1 equipment that had been placed in storage in spring 1961 had been incorporated into its own spacecraft, now part of the F-2 series and designated a Group 0 digital mission. It was given the TALENT-KEYHOLE designation Mission 7151, and was launched on February 21, 1962 from Vandenberg atop a Thor Agena vehicle. Once the Agena separated from its Thor first stage it fired its engine to place the entire spacecraft into orbit.

CORONA reconnaissance satellites, which had first started launching three years earlier, operated by flying horizontally over the Earth with the camera pointed out of the side of the spacecraft, down toward the ground. But the signals intelligence satellites had their antennas located under a fairing at the front of the Agena. Once they reached orbit, small gas jets tipped the Agena so its nose pointed toward the ground and its engine bell pointed out toward space. This orientation lined up with the Earth’s pull on the spacecraft, which was slightly stronger at the end nearer the Earth. The end result was that once the spacecraft was pointed down it stayed pointed down and no further attitude control was required—or at least the satellite designers thought at the time; later experience demonstrated that this was not always true.

During the second evening of operation, around 4 AM, ground controllers made a critical error—instead of sending a command to the onboard tape recorder to read out its data, they unknowingly sent a command for it to read-in, or record, data

The spacecraft was battery operated and Mission 7151 operated successfully for six days in orbit until its batteries were depleted. The spacecraft successfully collected data over the Soviet Union and transmitted it to American ground stations. But when signals analysts began processing data from Mission 7151 they noticed that the location accuracy for the detected signals were poor, meaning that they could not pinpoint the locations of the Soviet radars that the satellite detected. To verify the data, Ed Stillman of Lockheed used American radars like the MSQ-11 in Alaska. The MSQ-11 emissions had also been detected by the satellite. Because the locations of those radars were known, Stillman could compare their actual locations with what the satellite was indicating and correct the biases. He discovered that the arithmetic signs of spacecraft pitch and roll had been entered into the computer reversed. Once they corrected for this error, the location accuracy improved.

Four months after the first spacecraft had launched, the Air Force launched Mission 7152. On June 18, 1962, another Thor Agena roared into the sky at Vandenberg Air Force Base, this time carrying a Group 2D digital payload. The spacecraft was equipped with an AMIE tape recorder developed by RCA in Camden, NJ. The AMIE recorded only when the receiver recognized a specific signal, resulting in short recording times. In order to preserve the magnetic tape from wearing out from constant stopping and starting, the payloads were allowed to collect signals for many orbits before the data was read out from the tape recorder when the satellite was in view of a ground station.

During the second evening of operation, around 4 AM, ground controllers made a critical error—instead of sending a command to the onboard tape recorder to read out its data, they unknowingly sent a command for it to read-in, or record, data, which caused the tape to run in one direction. When that did not result in the data being played back, they sent a read-out command. The two reels around which the magnetic tape was wrapped pulled in opposite directions and broke the tape, ending the mission. At a post-mortem the engineers decided that they needed to redesign the recorder command system to prevent it from destroying the tape, and they changed operating procedures to avoid another fatal error.

After the February and June 1962 launches, the next mission involved a Group 1D payload and was initially scheduled for November using a Thor Agena equipped with solid rocket motors around its base to provide additional thrust. But apparently delays in the payload construction caused both a launch delay and a switch in launch vehicle. Although the official history is unclear, it appears that the Mission 7153 satellite was switched to a different rocket and successfully launched on January 16, 1963. Apparently, the original Thor Agena for the payload was reassigned for a CORONA photo-reconnaissance satellite that launched a few weeks later, but one of the Thor’s solid motors did not ignite and the CORONA did not reach orbit.

Although Mission 7153 did reach orbit, the spacecraft suffered a battery failure after only two days. Thus, whereas the first spacecraft had operated for six days, the second two spacecraft had only operated for a total of less than four days. But with every failure the spacecraft designers and ground controllers learned something and improved their equipment and their procedures, and in summer 1963 they prepared to try again.

They proved that the Agena SIGINT satellites could work and could provide useful signals intelligence data on Soviet air defense systems. But while they were launching every six months, it was possible for the Soviets to improve their air defenses at an even faster rate.in, or record, data

On June 29, 1963 another Thor Agena lifted off from Vandenberg carrying Mission 7154, equipped with another Group 1D payload. The satellite achieved its proper orbit, pointed its antennas toward the ground, and turned itself on. This time there were no failures and no mistakes and the spacecraft established a new record of ten days of orbital operations and approximately 140,000 good intercept data points.

Those first four launches of the Samos F-2 program were all digital—one Group 0, one Group 1D, and two Group 2D payloads. They proved that the Agena SIGINT satellites could work and could provide useful signals intelligence data on Soviet air defense systems. But while they were launching every six months, it was possible for the Soviets to improve their air defenses at an even faster rate. The Soviets were quickly introducing new radars and radar operating modes as well as mobile radar systems like the FAN SONG radar used with the deadly SA-2 Guideline surface-to-air missile that had shot down Gary Powers in his U-2 spyplane in 1960. Because of this, the people designing these signals intelligence satellites came up with another method of gathering intelligence, one that could operate on a far quicker turnaround time. It was called AFTRACK, and it was so classified that unlike the larger satellites, its existence was essentially unknown for nearly five decades.

Next: Part 2, AFTRACK

About this series: The information in this series has been derived from numerous sources, but the primary sources are the different versions of The SIGINT Satellite Storydeclassified in 2015 and 2016, the AFTRACK document declassification, and documents released on the Samos satellite program. Part 2 deals with AFTRACK, part 3 with the other Agena SIGINT satellites, and part 4 with the P-11 and other small SIGINT satellites.