Deep Space 1
Introduction:
Deep space 1 was an extraordinary historic launch with fascinating new discoveries for the innovative world. The successful launch took place on October 24, 1998 and lasted for roughly three years. The objective of the mission was originally an engineering test flight for a dozen new high-risk advanced technologies. Such technologies included SCARLET II, xenon ion engine, solar concentrator array, autonomous navigation plus two other autonomy experiments, small transponder, Ka-band solid state power amplifier, and experiments in low power electronics, power switching, and multifunctional structures. Deep Space 1 also encountered Comet Borrelly and returned the best images and other science data ever from a comet. The Deep Space 1 was the first of a series of technology demonstration probes developed by NASA’s New Millennium Program. Its primary mission was to prove innovative new technologies for future spacecraft. The project manager for Deep Space 1 was Dr. Marc Rayman of NASA’s Jet Propulsion Laboratory. While the spacecraft accomplished a planetary-class mission, the asteroid encounter was not a required aspect of the mission. The encounter was only a bonus in the primary mission. NASA approved the extended mission to Comet Wilson Harrington in August 1999. Then the spacecraft’s sole star tracker failed in November 1999. A seven-month rescue effort succeeded in recovering the capability to operate the spacecraft without the star tracker, but the rescue excluded encountering both comets. On December 18, 2001, with no further technology objectives and no further science objectives, the spacecraft was commanded into a storage configuration with the transmitter off but the receiver left on. The goal here is to present the Deep Space 1, its origins, and its contributions. Specs of Deep Space 1:
Deep Space 1 spacecraft was built on an octagonal aluminum frame, 1.5 meters high, 1.1 meters deep, and 1.1 meters wide. Total dimensions with antennae deployed were 2.5 meters high, 2.1 meters deep, and 1.7 meters wide. Batteries and two solar panel “wings” attached to the sides of the frame, which spanned 11.75 meters when deployed, powered the probe. The spacecraft weighed 486.3 kilograms with fuel and cost the U.S. $149.7 million to build. The solar panels, designated SCARLET II (Solar Concentrator Arrays with Refractive Linear Element Technology) constituted one of the technologies tested on the spacecraft. “A cylindrical lens concentrated sunlight on a strip of GainP2/GaAs/Ge photovoltaic cells and acted to protect the cells. Each solar array consisted of four 160 centimeters by 113 centimeters panels” (Encyclopedia Astronautica). The array furnished 2,500 watts at 100 volts at the beginning of the mission, and less as the spacecraft moved further from the sun and the solar cells aged. Communications were via a high-gain antenna, two low-gain antennae, and a Ka-band antenna, all mounted on top of the spacecraft, and a third low-gain antenna mounted on the service boom. Propulsion was provided by a xenon ion engine mounted in the propulsion unit on the bottom of the frame. Total propellant aboard included 81.5 kilograms of xenon and 31.1 kilograms of hydrazine for the reaction control system. Of the 81.5 kilograms of xenon, “73.4 kilograms was expended by the end of the hyperextended mission in December 2001” (Encyclopedia Astronautica). The thirty centimeters diameter engine consisted of an ionization chamber into which xenon gas was injected. Electrons were emitted by a cathode traverse discharge tube and collided with the xenon gas, stripping off electrons and creating positive ions. The ions were accelerated through a 1,280 volt grid at 31.5 kilometers per second and ejected from the spacecraft as an ion beam, producing 0.09 Newtons of thrust at maximum power. Deep Space 1 was the only mission ever to rely on ion propulsion as the primary propulsion. It operated the ion engine for 16,265 hours, far longer than any other mission (using any kind of propulsion) had operated its propulsion system.
Launch Details:
Deep Space 1 was successfully launched from pad 17-A at the Cape Canaveral Air Station at 12:08 UT, the first launch under NASA’s Med-Lite booster program, on a Delta 7326-9.5 with three strap-on solid propellant rockets. Telemetry was received by NASA Deep Space Network one hour and thirty-seven minutes after launch, a thirteen minute delay from the expected time. The reason for the delay was radiation from the Van Allen belts causing false locks in the star tracker, thus delaying the spacecraft in acquiring its initial altitude after separation. Deep Space 1 was originally scheduled to “fly by the asteroid 3352 McAuliffe in 199 and comet P/West-Kohoutek-Ikemura and the planet Mars in the year 2000 but because of a launch delay these targets were no longer possible” (DS1 Technology Validation Reports). At launch, the primary mission was planned to last until September 18, 1999, with the possibility of an extended mission to fly by the comet Borrelly in September 2001. Deep Space 1 was to fly by the near-Earth Asteroid 1992 KD on July 28, 1999 at a distance of five to ten kilometers. The comet encounters were never a part of the primary mission at all. At launch, the concept was that if the primary mission were successful, and if the technologies worked, and extended mission would be proposed to encounter comet Borrelly. It was not until several months after launch that the extended mission concept was modified to include comet Wilson-Harrington. In the event, the primary mission exceeded its success criteria, and NASA approved the extended mission in August 1999. Then, the spacecraft’s sole star tracker failed, and a seven-month rescue effort succeeded in recovering the capability to operate the spacecraft without the star tracker.

Technologies:
Two of the technology demonstrations included the Miniature Integrated Camera-Spectrometer (MICAS) and the Plasma Experiment for Planetary Exploration (PEPE). Deep Space 1’s payload consisted of twelve advanced technologies, two of which happened to be these instruments. Their purpose was not to make scientific measurements at the encounters, as the encounters were not part of the primary mission requirements. Their purpose was to demonstrate the in-space flight operations and quantify the performance. Both MICAS and PEPE were ambitious new designs for science instruments, with each integrating what normally would be three or four separate units into one, thus consuming much less power and mass. Deep Space 1 was designed to determine whether and how well these new instruments could work. Each also performed another function for different advanced technology. MICAS provided the “visible images used by the autonomous optical navigation system while PEPE aided in characterizing the effect of the ion propulsion system on the space environment” (DS1 Technology Validation Reports). The science data at the asteroid were a bonus from their inclusion in the flight. Now in the extended mission, which was quite distinct from the primary and hyperextended mission, the focus was indeed to use these instruments as well as the others to collect scientific data.
The rescue following the loss of the star tracker was one of the most remarkable rescues in deep space exploration. The star tracker had caused occasional brief problems during the primary mission. The unit was not one of the new technologies; it was a commercial, off-the-shelf device. On November 1999, during the extended mission, it stopped operating, depriving the spacecraft of its three-axis attitude knowledge. Deep Space 1 was the “lowest cost interplanetary mission NASA had ever conducted as measured in same-year dollars, including launch vehicle and mission operations” (Solar System Exploration). There was only limited redundancy and the star tracker did not have a back-up. This was such a significant failure that termination of the extended mission was given serious consideration, particularly since the primary mission had already exceeded its success criteria. Indeed, before launch, the loss of the star tracker was considered catastrophic. Without the star tracker, the spacecraft was only capable of pointing one axis toward the Sun and slowly spinning around that axis. Nevertheless, the operations team devised a method to point the high-gain antenna (HGA) to Earth using the downlink radio signal measured at the Deep Space Network as an indicator of spacecraft attitude. This was significantly complicated by the spacecraft being 1.7 AU away at the time. Once the HGA was pointed to Earth, a thorough diagnosis of the star tracker could be conducted. It was not recoverable.
Star Tracker:
Attention then turned to new ways to fly the mission. Over the next four months, new operational procedures were devised and new software was developed to use the camera as an attitude sensor in place of the star tracker. This was complicated by the camera having less than “one percent of the star tracker’s field of view, a brighter limiting magnitude, an output rate 100 times slower, and an output format of image files as opposed to the star tracker’s quaternions” (Remote Agent). In four months, the new software was designed, developed, tested, and integrated. In the beginning of June 2000, the entire flight software was replaced. This was a delicate and complex operation as the spacecraft did not have a redundant computer. The new system worked right away, restoring the spacecraft’s three-axis attitude knowledge and control. The ion propulsion system was reactivated, allowing its attempt to reach comet Borrelly to resume on June 28, one week ahead of the ambitious schedule that had been set in January. During the subsequent eighteen months of operation, the system worked extremely well, allowing the spacecraft to reach comet Borrelly with few problems.
Data Returned:
The spacecraft was not designed for a comet encounter. For instance, it did not have shielding from the cometary environment and the attitude control system was not built to track a moving target at finite distance. As a result of the rescue from the star tracker loss, the camera had to be used for both attitude reference and science data acquisitions at the comet. The rescue consumed more “hydrazine than normal operations would have, and conserving the hydrazine was one of the highest priorities” (Remote Agent). Expending the hydrazine before the encounter was one of the highest risks. Since the spacecraft did not have reaction wheels, the mission would end within three hours of exhausting the hydrazine. Still, the encounter proceeded flawlessly, exceeding the science measurement goals. Visible images, the first ever to allow geology to be studied on a “comet’s nucleus, and infrared spectra were collected with MICAS, ion and electron energy, and angle spectra and ion composition data were collected with PEPE, and plasma wave and magnetic field measurements were made with the IPS (Ion Propulsion System) diagnostic sensors” (Remote Agent). The IDs was included on the flight as part of the testing of the IPS. The suite of sensors, designed to assess the effect of the IPS on the spacecraft and space environment, was reprogrammed in flight to collect data at the comet. The encounter was September 22, 2001. Following its return of data from the comet, the Deep Space 1 extended mission was complete. Then the Deep Space 1 hyperextended mission began. It turned attention back to technology testing, taking advantage of the mission’s unexpectedly long life. All nine hardware technologies were exercised, some for the first time in over two years, to investigate the effects of aging and radiation. The focus of the hyperextended mission was on the IPS, exploring modes that were too risky or otherwise inappropriate for earlier in the flight. All tests were completed successfully.

AutoNav:
Prior to the star tracker failure, the autonomous onboard optical navigation system (AutoNav) was operated a great deal, thereby reducing work for the operations team in addition to testing this technology. The rapid pace of the recovery from the loss of the star tracker did not permit modification of AutoNav so it could function with the new design of ACS. As a result, the optical navigation images on approach to Borrelly were analyzed by the navigation team rather than onboard the spacecraft. The initial detection of the comet required “co-addition of the images but, as the range between the comet and spacecraft diminished, the comet became detectable in individual frames” (Solar System Exploration). The optical navigation data proved to be very powerful. The cometary ephemeris, as determined from the optical navigation images, differed by 1,500 kilometers from the ephemeris. In addition, the cometary ephemeris was derived from the much denser and longer set of ground-based observations. To determine whether this discrepancy might have been a result of errors in knowledge of the spacecraft’s trajectory, the Doppler and range data were supplemented with another data type. Delta differential one-way range data were acquired on the fourteenth and fifteenth of September. Delta differential one-way range achieved very high accuracy in the determination of a spacecraft’s angular position using simultaneous measurements from two Deep Space Network (DSN) stations. Each DSN station alternated between observing the spacecraft and observing a quasar within five degrees to remove or reduce the effects of many of the uncalibrated error sources. The observations proved that the spacecraft’s trajectory was not the source of the ephemeris discrepancy. Despite the risks from the environment, the spacecraft already being handicapped with a very complex design involving 685 commands and more than 3,000 parameters in 44 sequences, the encounter was essentially flawless and exceeded expectations. On approach to Borrelly, Deep Space 1 viewed the comet near the south ecliptic pole. The spacecraft’s closest approach of 2,171 kilometers was at 22:29:33 UTC on September 22, 2001 with a velocity of 16.58 kilometers per second. The encounter was placed 1.36 AU from the Sun, eight days after the comet’s perihelion. Science data were acquired with three instrument suites. All were body-fixed, so pointing required spacecraft maneuvers. Micas collected “panchromatic images and also obtained spectra with a sampling interval of seven nanometers” (Encyclopedia Austronautica). Ion and electron energy and angle spectra and ion mass/charge measurements were made with another instrument included on the flight as a technology experiment for the primary mission, the PEPE. PEPE had a resolution of five percent in energy and in mass/charge. Magnetic field and plasma wave measurements were made with sensors that had been carried as part of the assessment of the IPS. These IPS diagnostic sensors measured the effects of the IPS on the spacecraft and space environment during the primary mission and were reprogrammed in flight to collect science data at the comet. Some of the optical navigation observations proved to be of scientific interest as well. Dedicated scienctific data acquisition, however, began with PEPE and IDS measurements twelve hours before closest approach. MICAS acquired some visible images for coma science and some to initialize RSEN; a subset of the images served both purposes. The operation of MICAS as an attitude sensor precluded its uninterrupted use as a science instrument during the approach to Borrelly.
Auto Tracking System: The auto tracking system was designed to accommodate the complicated scene with a partially illuminated nucleus, jets, the coma, cosmic ray tracks, background stars that could produce streaks as the spacecraft tracked the nucleus, and stray light. Deep Space 1 kept MICAS pointed at the nucleus long enough for the image to be in the field of view. By the time the next CCD image was taken thirty seconds later, the nucleus was no longer in the field. In fact, RSEN continued to predict the position of the nucleus to an accuracy smaller than the camera’s field of view but, ACS, “not designed to track a body through such a flyby, was not able to keep up with the predicted position” (Deep Space 1 Extended Mission). This was not a limitation of RCS control authority; rather, it apparently was the result of a lag in ACS that was manifested only in the case of a significant angular acceleration of the target. This was of no importance for the primary mission’s technology testing requirements. There were insufficient resources to address this limitation in the extended mission, particularly because it was not an obstacle to achieving the scientific goals. By the time the probability of the nucleus being out of the CCD field of view was very high, the slit of the infrared spectrometer was swept across a range predicted to include the nucleus. This maneuver also was used to begin achieving the attitude required for PEPE measurements through closest approach. In addition to attaining the optimal orientation for PEPE, the spacecraft stopped attempting to track the nucleus and, instead, assumed a constant angular rate. This greatly reduced RCS firing, thus minimizing the possible interference of hydrazine decomposition products with PEPE’s measurements. It also served to reduce the amount of RCS solenoid activity that could register in IDS magnetometer measurements during this important portion of the encounter.
Results:
All MICAS, PEPE, and IDS science measurements worked as planned, and more science data were returned that had been expected. Because the tracking software located the nucleus in all but one of the images that contained it, fifty-two images of the nucleus were returned, many also showing details of the coma and dust jets. In the final view, the nucleus spanned about 175 pixels, or 3.5 times the target for the best image. No spatially resolved infrared spectra were required as their acquisition initially had been expected to place too much risk on the image data. However, the scan of the IR slit across the nucleus succeeded, yielding forty-five spatially separated swaths across the nucleus, each 165 meters in its shorter dimension. Several collimated jets and broader fans of dust are observed emanating from the nucleus. The strongest jet is at least 100 kilometers long and is directed thirty degrees from the Sun-nucleus axis. The jet originated from a broad basin near the center of the nucleus. Optical navigation images taken during the thirty-four hours prior to the encounter show evidence that the direction of the jet was stable over times longer than the nucleus rotation period. Borrelly was at a solar elongation of sixty-three degrees at the time of Deep Space 1’s encounter, permitting complementary observations from Earth. In addition to ground-based measurements, “Hubble Space Telescope obtained visible images and ultraviolet spectra and Odin acquired spectra at 557 gigahertz; observations planned from Chandra X-ray Observatory were missed because of a temporary spacecraft problem. In addition to the direct scientific return, Deep Space 1’s data are of engineering value to other missions planning to visit comets” (Remote Agent). Stardust, already on its way to collect samples of the coma of comet 81P/Wild 2 in January 2004, will benefit in several ways from these data. Some of Deep Space 1’s operational experience of encountering a comet had been transferred to the Stardust project by having some people working on both missions as well as selected Stardust team members participating as guests in the Borrelly encounter. Deep Impact will launch in 2004 for an encounter with comet 9P/Tempel 1. Deep Space 1’s science data have led to several changes in Deep Impact’s plans.
Summary/Conclusion:
The end of Deep Space 1 mission marked the end of a project that overcame numerous daunting obstacles and returned many important results, from development through the hyperextended mission. Deep Space 1 was inherently risky, even before launch, as it used technologies that were chosen in part of high risk they presented to the first user. Indeed, if a technology did not pose some important risk, its testing in an NMP mission would not be needed. Furthermore, Deep Space 1 probed the limits of schedule and costs for development and operations. From the beginning of the pre-phase A study to launch was thirty-nine months. The total cost for “development, launch service, and operations through the conclusion of the primary mission in 1999 was less than $150 million” (DS1 Technology Validation Reports). This included the development cost of some of the technologies in Deep Space 1’s payload and the integration costs of all of them. Despite its very aggressive schedule and small budget, Deep Space 1 met or exceeded all of the primary mission success criteria. The knowledge gained during development and operations will be significant help to many future missions, as the costs and risks of using the technologies that formed its payload had been significantly reduced. The benefits accrue not only from the quantification of their performance during flight, but also from the insight derived from incorporating the new capabilities into the spacecraft, ground segment, and mission design, thus illuminating implementation issues that would not have arisen in typical technology development or conceptual mission studies. Many missions are now being enhanced or enabled through the results of Deep Space 1’s intensive testing of technologies in the primary and hyperextended missions.
References

1. “Deep Space 1 Extended Mission”. NASA, N.d. Web 29 Feb. 2013. < http://nmp.jpl.nasa.gov/ds1/DS1_Extended_Mission.pdf>.
2. "DS1 Technology Validation Reports". NASA, N.d. Web. 26 Feb. 2013. < http://nmp-techval-reports.jpl.nasa.gov/>.
3. Encyclopedia Astronautica. N.p.: n.p., n.d. astronautix. Web. 26 Feb. 2013. .
4. “Remote Agent”. NASA, n.d. Web. 26 Feb. 2013. .
5. “Solar System Exploration”. NASA, n.d. Web. 26 Feb. 2013.
< http://solarsystem.nasa.gov/missions/profile.cfm?MCode=DS1>.