The Space Reviewin association with SpaceNews

Agena spacecraft being prepared for launch in August 1961. This mission carried both a TAKI payload for intercepting Soviet TALL KING air surveillance radar emissions and a SOCTOP payload for determining if the vehicle was being tracked or the Soviets were attempting to take control of it. The payloads may have been mounted under the gray box structure on the aft rack of the Agena. The CORONA photo-reconnaissance payload has not yet been attached to the front of the Agena. (credit: Peter Hunter)

The wizard war in orbit (part 2)

Black black boxes

Bookmark and Share

By fall 1959, a number of CORONA photo-reconnaissance spacecraft had already been launched under cover of the Discoverer program, but none had operated successfully. Program officials became concerned that the Agena spacecraft that carried CORONA might be vulnerable to tracking by Soviet radars, or possibly even deliberate electronic interference. They did not think this explained CORONA’s early string of failures, but it was a possibility they worried about. At the time, Harold Willis was working in the Office of ELINT located at CIA Headquarters when CORONA officials briefed him about their program and told him about their concerns.

Willis also learned about the Samos Subsystem F signals intelligence satellite program, which at the time consisted of the F-1 and F-2 payloads. The former was a relatively small payload that would fly attached to a Samos photo-reconnaissance satellite and the latter a larger and more capable payload that would occupy the front end of an Agena spacecraft. Although the specialized F-2 satellite might be able to detect Soviet transmissions or interference, it was then scheduled to fly years after CORONA became operational. Willis thought that the Soviet threat to CORONA and other military satellites could develop sooner and they should not wait for the Samos signals intelligence satellites to provide data. He was not simply worried about problems over the Soviet Union but even far out over the oceans: the Soviets also had ships and trawlers with radomes, and nobody knew what they were for.

Program officials became concerned that the Agena spacecraft that carried CORONA might be vulnerable to tracking by Soviet radars, or possibly even deliberate electronic interference. They did not think this explained CORONA’s early string of failures, but it was a possibility they worried about.

Willis discussed his concerns with Lockheed’s Bill Harris, who was working on Subsystem F payloads. Willis concluded that the Agena upper stages carrying CORONA cameras should be equipped with a small payload for detecting Soviet radar tracking or interference with the spacecraft’s S-band beacon. The beacon was used to announce the satellite’s presence to American ground tracking sites so that they could communicate with it. CIA officials approved Willis’ proposal in November 1959. The Agena spacecraft had an aft rack to which instrument boxes could be attached, like a luggage rack on a car. “AFTRACK” became the collective name for a large number of payloads carried on numerous spacecraft during the first half of the 1960s.

Samos Subsystem F was classified at the DoD Secret level, but CIA officials determined that AFTRACK should be handled on a strict need-to-know basis. Lockheed Missiles and Space Company arranged for office space on Hanover Street in Palo Alto for the new project. Harris would conduct payload development and integration for the AFTRACK system.

AFTRACK became the satellite concept demonstrator for what was already known within the electronics intelligence world as a “quick reaction capability” (or QRC) payload that would require minimal development time. As early as the Korean War, the Air Force had undertaken QRC projects to deal with new radar threats to their aircraft. Now the CIA was going to apply the concept to spacecraft.

Fortunately for Harris, the first task he oversaw was relatively straightforward. The Agena vehicle developers imposed a mandatory requirement that there had to be a fuse in the power line of the AFTRACK payload so that it could not affect the primary payload power system, but otherwise they placed few requirements on the new system. Earth-pointing antennas could be easily mounted on the Agena’s aft rack so that they were pointing in the proper direction. The early CORONA vehicles had lifetimes of five or six days because they relied entirely on batteries. This limited collection time for AFTRACK.

Willis worked for the CIA’s Office of ELINT. Signals intelligence, or SIGINT, is the overall term encompassing several subfields including electronic intelligence (ELINT)—primarily the interception of radar signals—and communications intelligence (COMINT). The Agena SIGINT spacecraft sometimes conducted both ELINT and COMINT. But although outsiders often referred to America’s smaller intelligence satellites as “elints,” this designation was not common within the intelligence community. One declassified document that regularly referred to AFTRACK payloads as “elints” had the word crossed out in pencil wherever it appeared. AFTRACK payloads eventually evolved to include much more than simply ELINT missions. Finding the signals and deciphering them was an esoteric art. Lockheed reports on new AFTRACK systems soon had a cover image of a witch on a rocket-powered broomstick orbiting the Earth and carrying a banner that said “Applied Witchcraft.”

Applied Witchcraft
Image from cover of a Lockheed report on an AFTRACK payload.


After being given his new task and learning about the Samos SIGINT work, Willis contacted Major John Copley, who was overseeing Subsystem F. It was being built by the Airborne Instruments Laboratory at Mineola, Long Island, New York. According to The SIGINT Satellite Story, an official history written in 1994, Gene Fubini was at AIL and “became an enthusiastic supporter of the AFTRACK concept and suggested a small payload called SOCTOP, which received signals in the 2500–3200 MHz frequency band,” the same band for the S-band beacon on the Agena.

SOCTOP only required on/off commands and a few telemetry channels. Willis arranged to mount SOCTOP on Discoverer 13, which was the fourteenth attempt to launch a spacecraft under the cover of the Discoverer program. Despite the unlucky number, Discoverer 13 was a lucky flight: the first successful flight in that program. The recoverable capsule, containing only an American flag, returned to Earth and was displayed in the White House. It is now on display in the Milestones of Flight Gallery at the Smithsonian’s National Air and Space Museum. Meanwhile, Willis’ SOCTOP payload had produced some alarming data.

SOCTOP was the first of what came to be known as a “vulnerability payload” intended to determine if the spacecraft was vulnerable to tracking or interference. Eventually this type of payload flew on almost every low-altitude reconnaissance system that the United States launched.

Analysis of the SOCTOP data indicated that the Soviets were tracking the CORONA spacecraft. Almost every time an American tracking station obtained readout data from Discoverer 13, SOCTOP had reported that it was being tracked by radar. SOCTOP had no data recorder, so ground controllers could not determine if it was being tracked while directly over the Soviet Union, only when it was in sight of an American ground station. Willis passed the data on to the intelligence community. But upon further analysis Willis determined that the radar SOCTOP was detecting was not Soviet in origin. Instead, SOCTOP was receiving signals from American VELORT (very long range tracking) radars that happened to be collocated with the ground stations receiving the data—but at least it worked.


SOCTOP was the first of what came to be known as a “vulnerability payload” intended to determine if the spacecraft was vulnerable to tracking or interference. According to the recently-declassified history, eventually this type of payload flew on almost every low-altitude reconnaissance system that the United States launched. Very early on, recorders were added to the payload to enable them to gather data while deep over the Soviet Union, not only on the periphery. After the first SOCTOP flew in August, several more were launched until August 1961, although not every CORONA vehicle carried one.

A byproduct of operating the payloads was that they could discover new radar characteristics or variations in their patterns. The payload configurations were modified over time as Soviet radars changed. In early 1963 the Special Projects Office held a competition to design a more sophisticated payload that could detect and return characteristics of signals in a still-classified frequency band. This payload would also include a recorder. The Electronics Defense Laboratory (EDL) at Sylvania in Mountain View, California, won the competition to build the payload.

Gene Fubini, at the Airborne Instrument Laboratory in New York, also developed a simplified version of the Samos F-2 forward rack payload. It would simply scan the 0.4 to 1.5 gigahertz band to detect radar activity in the Soviet Union. This included suspected anti-ballistic missile and tracking radars. The mission was almost the reverse of the original SOCTOP, focusing on detecting ground radars rather than detecting radar tracking of the satellite, so they named it TOPSOC. It used super-heterodyne receivers and omnidirectional antennas. The Samos F-2 satellites had a directional antenna, but SOCTOP had an omnidirectional antenna that “scooped up a large number of interleaved signals, horizon to horizon, including sidelobes and main beams,” according to the history.

As the engineers and emissions experts developed these new detectors, they thought that the radiofrequency band of 400 to 1,600 megahertz was going to be relatively quiet. But when TOPSOC was launched on September 12, 1961, it encountered a very noisy signal environment over the Soviet Union. It was impossible to process the data. “The first lesson in matching the collection system to the processing system had been learned,” the official history declared.

What early American signals intelligence satellites like the Navy’s GRAB, Samos Subsystem F, and the AFTRACK payloads quickly revealed was that there were far more radars in the Soviet Union than intelligence experts thought. Signals analysts also learned that intercepting a signal without pinpointing the location of the transmitter was of very little value unless the intercept was unique and very high priority, so adding location detection capabilities to the American signals intelligence satellite program became higher priority.

From TAC to TAKI

Before the SOCTOP launches started, the Air Force had sponsored an annual review at Stanford Electronics Laboratory in Palo Alto focused on electronic intelligence activities, called the Technical Advisory Committee (TAC) meetings. TAC meetings were widely attended by contractors and agencies involved in electronic warfare and were a great networking opportunity for experts in the exclusive field of detecting electronic signals.

In August 1960, Harris, Major John Copley, and Phil Doersam, the Lockheed Missiles Subsystem F manager, attended the TAC meeting. They were looking for AFTRACK payload concepts, and that’s when they met with James de Broekert of Stanford.

Although hunting air surveillance radars in the Soviet Union was a target rich activity, the big game that the intelligence analysts were after was a large and powerful Soviet radar that had been designated HEN HOUSE.

De Broekert was an electrical engineer and had gone to the Stanford Applied Electronics Laboratory in the late 1950s in order to continue his education. He had previously worked on designing radar displays for the F-100 fighter aircraft. At Stanford, one of his first jobs was converting electronic receivers that used vacuum tubes to transistorized solid-state technology. His work had broad applications, for many of the Air Force’s satellites would benefit by using more robust and less power-hungry solid-state components.

De Broekert and his colleagues soon developed a small handheld battery-powered electronic intelligence, or “elint,” receiver. He regularly attended TAC meetings where government agencies and contractors involved in electronics intelligence systems met to discuss their work. “Harris visited SEL during this TAC meeting, and saw the small battery-powered receiver we engineered,” de Broekert explained in an interview published in a 2002 book about the recollections of pioneers of national reconnaissance. “We took a subsequent pocket-sized version of that receiver on airline flights and held it up to the window, logging radar signals as we flew. Harris took an interest in our small elint receiver,” de Broekert remembered. Then Harris asked him a question: “Do you think you can put that on a satellite?”

De Broekert said that it could be adapted to a satellite, and because it could detect Soviet TALL KING air traffic control radars, they named it TAKI. It included a tape recorder for recording signals over the Soviet Union that would be commanded to read back its data when it was in view of an American ground station. In October 1962, TAKI was modified after the first mission to limit its coverage only to the TALL KING radar frequency. Several more TAKI payloads flew before their mission was taken over by other systems.

Don Grace became the Stanford Electronics Laboratory manager who handled AFTRACK payloads. Grace established a lab in the basement of their building on the Stanford campus. Don Eslinger built ten payloads there single-handedly, like a watchmaker assembling a very exclusive product, all while students went to classes on the floors above. The Stanford Electronics Laboratory would design and build the first of a new series and then industry would take on production.

“This was an exciting opportunity for us,” de Broekert remembered. “Instead of flying at 10,000 or 30,000 feet, we could be up at 100 to 300 miles and have a larger field of view and cover much greater geographical area more rapidly. The challenges were establishing geolocation and intercepting the desired signals from such a great distance,” de Broekert said in the 2002 book.

“Another challenge was ensuring that the design was adapted to handle the large number of signals that would be intercepted by the satellite. We created a simple model that was used to determine the probability of intercept on the desired radars. The model also determined the interference environment from the other radar signals that might be in the field of view. That was very challenging,” de Broekert explained.

“My function was primarily to develop the system concept and to establish the system parameters. I was the team leader, but the payloads were usually built as a one-man project with one technician and perhaps a second support engineer. Everything we built at Stanford was essentially built with stockroom parts. We had a qualified parts list and we knew which parts to avoid, but there were no high reliability parts in any of these systems. We built the flight-ready items in the laboratory, and then put them through the shake and shock fall test and temperature cycling. We did not do thermal vacuum tests in those days because we did not think it was necessary. These were relatively simple systems compared to today’s sophisticated payloads, but they did a good job.”

After they had produced the hardware, they would take it over to Lockheed Missiles and Space Division in Sunnyvale, only a few kilometers away. There the instruments were integrated with the Agena spacecraft.

Reconnaissance image of two HEN HOUSE radars at the sprawling Soviet Sary Shagan anti-ballistic missile research and development facility. HEN HOUSE was a major target for American signals intelligence satellites in the 1960s. (credit: NRO)

The Hunt for HEN HOUSE

Although hunting air surveillance radars in the Soviet Union was a target rich activity, the big game that the intelligence analysts were after was a large and powerful Soviet radar that had been designated HEN HOUSE when it was first spotted in April 1960 by the second-last U-2 spy plane mission, and then again photographed by CORONA reconnaissance satellites. Located near Sary Shagan on the edge of Lake Balkash in Kazakhstan, the massive radar building was long and thin, like the buildings seen on chicken farms. Photo-reconnaissance satellites would continue to take increasingly higher-resolution photographs of HEN HOUSE radars throughout the 1960s, sometimes gathering incredibly detailed images of the radars’ construction—soon one HEN HOUSE radar building was joined by two, then by even more. Eventually the Soviets had eight HEN HOUSE radar pairs located at five different sites around their territory to detect American ballistic missiles. The photographs were valuable, but what the intelligence analysts really wanted to capture were the radar’s signals.

After TAKI, Stanford’s engineers invented WILD BILL in spring 1961. It might have been named after Bill Harris or Bill Rambo. WILD BILL was to search for HEN HOUSE signals from the radar under construction at Sary Shagan. Nobody was sure what frequency it used other than 50 to 400 megahertz, based upon the size of the antenna and the requirement for tracking missiles and satellites.

Stanford Electronics Laboratory built WILD BILL and WILD BILL 1, covering the frequency range of 50 to 150 megahertz that signals analysts and radar experts thought would be the most probable band. WILD BILL was launched on July 7, 1961, on the back of a CORONA spacecraft and operated for two days, finding nothing. WILD BILL 1 was launched on February 27, 1962, on the aft rack of an Agena hosting another CORONA spacecraft, and operated for only two orbits, also not finding anything. ATI built later versions of WILD BILL. The company had been formed in Palo Alto by former Stanford Electronics Laboratory engineer John Grigsby. Lockheed contracted with ATI to build the follow-on versions of Stanford payloads. Multiple WILD BILL missions also flew, trying to gather HEN HOUSE signals.

The signals generated by HEN HOUSE were directed into space, so normally only a receiver in space could collect them. But after the Soviets detonated an atomic weapon in the atmosphere in October 1962, creating a high-altitude atmospheric phenomenon that reflected the radio waves back to the ground, a signal was received at the CIA’s ground station in Iran on the southern shore of the Caspian Sea. That signal was at approximately 600 megahertz and was in a format that could be HEN HOUSE. The signal was very distorted. Grigsby’s first payload, WILD BILL 2, was launched on December 12, 1962, covering the frequency range of 550 to 620 megahertz. But like its predecessor, it also did not detect anything.

Whereas chasing HEN HOUSE signals was perhaps the sexiest job in the signals intelligence field for a while, there were still many demands for precisely locating radars that posed a threat to American bombers that would have to penetrate Soviet air defenses.

Another signal that was also detected in October was at approximately 160 megahertz. So Grigsby designed the WILD BILL 3 payload for 150 to 230 megahertz. The payload was launched on June 12, 1963, and two weeks later it made the first confirmed satellite intercept of the HEN HOUSE target-tracking radar. Sary Shagan’s radar was emitting while the Soviets were conducting a missile tracking test; a declassified report indicates that they were tracking a missile launched from Makat and tracked at Sary Shagan. WILD BILL intercepted the radar signal when the satellite was 1,800 kilometers from the radar site. But the WILD BILL payload worked poorly and the tapes were noisy. The Naval Research Laboratory’s POPPY SIGINT satellite also detected the HEN HOUSE radar around the same time. POPPY was classified as a “general search” satellite designed to pick up new and unusual signals by examining broader frequency ranges, but that satellite was also carrying WILD BILL 4.

The suspected HEN HOUSE signal that had been detected on the ground was given a still-classified designation, but this designation was changed after WILD BILL 3, 4, and 5 collected much better data on it. The data was processed both at a still-classified location—probably Lockheed in Sunnyvale, California—and at Strategic Air Command in Omaha, Nebraska.

John Grigsby proposed a new payload to define the center frequency for HEN HOUSE, which scanned space by changing the frequency of its transmissions. The payload was known as LONG JOHN—Grigsby was rather tall. The first LONG JOHN mission was November 27, 1963, another was launched on February 15, 1964, and another launched on June 13, 1964, all on the backs of CORONA photo-reconnaissance satellites. The February mission suffered a recorder failure, but the missions managed to gather more HEN HOUSE signals.

Whereas chasing HEN HOUSE signals was perhaps the sexiest job in the signals intelligence field for a while, there were still many demands for precisely locating radars that posed a threat to American bombers that would have to penetrate Soviet air defenses. The last Stanford payload was PLYMOUTH ROCK, which covered the frequency range of 2.0 to 4.0 gigahertz. Strategic Air Command had requested the payload in order to identify and locate as many Soviet S-band tracking radars as possible. PLYMOUTH ROCK 1 was launched on November 24, 1962. It used a sweeping yttrium-iron-garnet filter for frequency discrimination. Two more PLYMOUTH ROCK payloads were built, and the last one was unique for AFTRACK payloads, although what made it unique remains classified.

SOCTOP Discoverer 13
The Agena spacecraft for Discoverer 13. Not only did this mission successfully return the first object from space in August 1960, but the Agena aft rack carried a highly classified payload named SOCTOP intended to determine if the Soviets were attempting to track or interfere with the spacecraft. (credit: Peter Hunter)

SIGINT enters mass production

In addition to the various quick reaction capability signals intelligence payloads produced by the handfuls, there were also some that were essentially mass-produced. According to declassified documents, one payload, named STOPPER, was a substantially improved version of SOCTOP “designed to detect skin-tracking and attempts at command probing” by the Soviets—i.e. trying to take command of the satellite. STOPPER also had possible use as a SIGINT collector. STOPPER was initially developed to be fitted to the back end of the Agena SIGINT satellites (see part 1) but soon added to the prolific CORONA reconnaissance satellites as well. Nearly two dozen STOPPER units were manufactured, but only a handful actually flew. Like many of the AFTRACK payloads, STOPPER was vulnerable to electronic noise from the spacecraft and sometimes from the STOPPER unit itself. STOPPER was described by one official as a big system, and it seems probable that electronics advances made it possible to build smaller units before many of the STOPPER systems could be flown.

STOPPER led to a far more prolific system known as BIT. (The existence of BIT was first revealed by this author in February 2012. See “BIT by precious BIT,” The Space Review, February 27, 2012.) BIT was initially intended to determine if the spacecraft was being tracked by a specific radar at Sary Shagan. Major Murray J. Sherline came up with the idea of tailoring frequency coverage of “BIT boxes” to known radar threats instead of duplicating the signals intelligence satellites and looking for entirely new threats. Starting with BIT 1 and continuing on through many versions, the detectors were built to collect data on Soviet HEN HOUSE, DOG HOUSE, and the TRY ADD radars.

The BIT box output was distributed to the National Security Agency and “other interested agencies” and used to program and even assist the design of other signals intelligence payloads. Over time, the program office changed some of their focus. Responsibility for the “BIT boxes” was transferred from the Los Angeles Air Force Base’s ELINT office to the newly-created satellite vulnerability office.

At NSA, the analog tapes were converted to “visa-corder” paper-roll photographic records. The records, which were often kilometers long, were manually scanned and sections of interest were analyzed in a laborious process.

The primary goal of the BIT boxes was to indicate if the satellite itself was being tracked. But after a number of BIT boxes had flown and not detected anything there was some rumbling within the intelligence community that they were not providing useful intelligence. This prompted someone involved in the program to write a defense of it, noting that the lack of evidence of Soviet tracking of the satellites carrying the boxes was actually a good thing because if the Soviets had been actively tracking most of the missions it would indicate that they could soon have an effective anti-satellite capability.

BIT boxes were carried on most low Earth orbit reconnaissance satellites at least until the later 1960s. The BIT 2 system used an antenna that stuck off the side of the Agena. It featured an inflatable boom with a circular antenna at the end, described as a “lollipop” antenna in declassified documents, although no photographs of this unusual antenna system have been declassified. BIT operated into the 1970s and was almost certainly replaced by more capable systems over time, and it seems likely that today every American reconnaissance satellite that flies carries a sophisticated descendent of this system. (The BIT program will be addressed in a later article that is not part of this series.)

By September 1963, over forty American SIGINT payloads had been placed into orbit. It was not only the CORONA spacecraft that carried AFTRACK payloads during this period. Two Samos E-6 photo-reconnaissance Agena spacecraft were also equipped with payloads designated ST. LOW, although the mission of ST. LOW is not identified in any of the declassified documents or official history.

Texas Pint
Frames from a film showing technicians preparing an Agena spacecraft for launch in August 1961. The TEXAS PINT communications intelligence payload may be visible in these images. It was designed to intercept Soviet communications between aircraft and the ground. (credit: Peter Hunter)

Processing pains

Although many of the details of the number, type, and quality of the collected signals during this early period of signals intelligence collection in space are still classified, what the declassified history indicates is that the ground processing of the data was often more challenging than developing the satellites and payloads to collect the data in the first place. The project’s engineers usually spent more time developing the processing procedures than actually processing the collected data. Data formats changed frequently because the AFTRACK payloads had to compete for positions on the Agena’s telemetry system and these changed from spacecraft to spacecraft, further complicating data processing.

The data from the spacecraft was beamed to ground stations and recorded on magnetic tape, which was then physically shipped to the National Security Agency for analysis. At NSA, the analog tapes were converted to “visa-corder” paper-roll photographic records. The records, which were often kilometers long, were manually scanned and sections of interest were analyzed in a laborious process.

National Security Agency analysis of the TAKI flights revealed a high density of TALL KING early warning radars. It also provided signal parameters for the radars. Some analysts thought that TALL KING might be part of the Soviet anti-ballistic missile system.

PLYMOUTH ROCK data was fed to a processing system on the ground called FINDER, which had originally been developed for the U-2 and other airborne collection systems. Strategic Air Command, using FINDER, produced radar locations with “circular error probable” estimates of 740 kilometers or greater for V-beam and height-finder radars—what this meant was that the radars had a 50 percent chance of being inside a circle that was 740 kilometers in diameter. This was clearly a very low accuracy, and how Strategic Air Command was able to use this data is unknown.

Alcohol seemed to be a popular subject for the signals intelligence community when it came to naming systems. After the Agena SIGINT satellite program got the designation Program 102, analysts named computer software programs for processing the digital data BREW1 and BREW2 after a popular beer known as Brew 102. Eventually names were changed to ROOK and another still-classified designation. A high-speed analog-to-digital converter named BEERMAN was developed in-house by NSA in 1963. Alcohol names were even applied to satellite payloads.

In fall 1961 the National Security Agency and Strategic Air Command both processed the same data. But in September 1962 this arrangement was changed. Thereafter, SAC would process the signals data in response to operational intelligence requirements. They would take tasking instructions from NSA, which would provide planning and technical support.

Responding to a problem of too many signals layering on top of each other, in late 1962 the engineers started working on a device known as a “deinterleaver.” It would separate overlapping signals with the same pulse repetition interval, deinterleaving them. The analog results would then be filmed for analysis. One still-murky aspect of the processing effort is the degree to which Lockheed was directly involved in processing signals. The declassified history hints that Lockheed may have had a processing facility co-located with its satellite production facilities.

Hearing voices

Radar signals were only one kind of electronic emission that interested the intelligence community. They also wanted to go after communications signals, so-called COMINT. In August 1959, Roger Thayer of the NSA wrote “Study Report on Collection of COMINT from Satellite Vehicles,” where he suggested that the Samos Subsystem F payloads could be adapted to COMINT.

Alcohol seemed to be a popular subject for the signals intelligence community when it came to naming systems. A high-speed analog-to-digital converter named BEERMAN was developed in-house by NSA in 1963.

Captain Don Wipperman at the Air Force Security Service in San Antonio came up with the first COMINT satellite concept. Working with the Airborne Instrument Laboratory team they presented a proposal for an AFTRACK payload that could intercept a signal that was then thought to be the air-ground communications network in the USSR. The payload became known as TEXAS PINT and launched on August 30, 1961, on the Agena aft rack of a CORONA spacecraft. But when analysts processed the signals, they determined that the Soviets had apparently started phasing out that system in favor of a more advanced communications network. TEXAS PINT still managed to collect VHF data over the Soviet Union that was later used to design other signals intelligence payloads.

In summer 1961, Sanders Associated developed two NEW JERSEY payloads to intercept the new signals. They were named NEW JERSEY because the original idea had come from ITT in Nutley, New Jersey. The first was launched on July 27, 1962, and the second on January 7, 1963. They intercepted and located several signals using Doppler techniques. Sanders began working on a follow-on payload named NEW HAMPSHIRE, but it ran into difficulties and never flew.

Wayne Burnett of HRB-Singer at State College, Pennsylvania, proposed a concept for intercepting, encoding and recording a radio teletype channel of the Soviet point-to-point VHF multichannel communications signal. The downlink signal from the satellite had to be encrypted to prevent the Soviets from figuring it out. The National Security Agency supplied an encryption device that was used during readout. HRB engineer Conrad Welch invented the intercept electronics, and the payloads were named GRAPE JUICE. They were launched on December 12, 1961, April 17, 1962, and September 17, 1962. Unfortunately, European TV and FM radio stations created so much interference that the GRAPE JUICE payloads were only able to lock onto the desired signal for short periods.

Following the frustrations with GRAPE JUICE, the engineers developed more filtering and better logic systems and produced the more refined VINO payload (sometimes referred to as “RED VINO”) launched on December 4, 1962. By March 1963, the Committee on Overhead Reconnaissance (COMOR) Target Working Group indicated that the highest priority COMINT signal was a major command link among Soviet Rocket Forces. But the working group determined that it would be difficult intercepting these signals. The second highest COMINT priority was going after the RED VINO signals.

A final version named OPPORKNOCKITY was launched on August 21, 1964. The name was an inside joke, a play on the saying that opportunity only knocks once—OPPORKNOCKITY, in contrast, only tuned once. It was designed to hold lock on the signal through extensive interference. OPPORKNOCKITY made 12,000 intercepts. It brought back larger amounts of data, but the designers still did not consider it to be an ideal system for the type of intelligence they required. After the first one flew officials evaluated whether to fly a second one. An extra OPPORKNOCKITY payload, developed as the service test model for the flown item, was in storage. But officials determined that it was obsolete and it apparently never flew.

SQUARE TWENTY was designed to locate Soviet communications links of a type that is still classified, but may have been the Soviet Rocket Forces communications link that COMOR had identified as the highest priority COMINT target. Although no photographs or illustrations of the SQUARE TWENTY payload have been released, it was a substantial system mounted to the back end of CORONA satellites, including two large deployed antennas. SQUARE TWENTY had a pencil beam antenna consisting of a downward-pointing parabolic dish approximately two meters in diameter. It was also equipped with a fan beam antenna that was a two-meter-long curved rectangular section with a parabolic taper that looked forward in the direction of vehicle motion. As the vehicle orbited the Earth, the fan beam antenna would be the first to pick up signals, with the pencil beam antenna capturing more accurate information as the satellite flew overhead, assuming that the emitter was within the antenna’s field of view.

The payload was launched on October 28, 1965 and lasted for 11 days, producing many locations of signal emitters. SQUARE TWENTY made 1,290 intercepts, resulting in 209 communications transmitter locations, producing still-classified accuracy results that were determined after multiple intercepts, and approximately 18-92 kilometer accuracy based upon single intercepts. Although SQUARE TWENTY had the ability to copy the collected data, it was not able to lock on for long enough periods and the data was not useful—apparently the analysts were able to determine where the transmissions were occurring, but not what was actually being communicated. By early 1967, another SQUARE TWENTY payload, this time with an antenna more than three meters in diameter, was scheduled for flight in the fall. But it is unclear if this mission flew and what spacecraft carried it. If it did, it may have flown on one of the Agena SIGINT satellites.

By the time SQUARE TWENTY went into orbit, the AFTRACK payloads were no longer as useful as they had once been. The reasons remain classified, but some of the limitations of this approach to gathering signals intelligence are obvious, including the fact that the payloads had to fly attached to a battery-powered spacecraft that was intended to perform a photo-reconnaissance mission and was not optimized for signals intelligence. In addition, power and lifetime were determined by the primary payload—if signals intelligence analysts wanted more power or to stay in orbit longer collecting more signals, they were out of luck, the satellites were commanded to reenter and burned up in the atmosphere. By 1965, over forty AFTRACK payloads—not counting the prodigious STOPPER and BIT vulnerability payloads—had flown. By 1963 a new series of satellites with longer lifetimes began operating, taking over the quick reaction capability that had driven the development of the AFTRACK payloads. Those satellites will be addressed in part 4 of this series.

Next: Part 3: SIGINT satellites go to war

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 Story declassified in 2015 and 2016, the AFTRACK document declassification, and documents released on the Samos satellite program and obtained by the author over several decades.