Deep Space 1

from Wikipedia, the free encyclopedia
Deep Space 1 (NASA)
Deep Space 1 is being prepared for tests
Close-up of the spacecraft's NSTAR ion engine

The US American spacecraft Deep Space 1 (or DS1 for short) was launched on October 24, 1998 to test new technology and programs and operated until December 18, 2001. DS1 was the first mission as part of the so-called New Millennium Program to test new ones Space technologies. The mission also enabled the scientific observation of the asteroid (9969) Braille and the comet 19P / Borrelly . NASA chose the name of the probe because of the techniques tested for missions in deep space, not in reference to the Star Trek series Deep Space Nine .

The mission cost a total of less than US $ 150 million for the primary mission, including all development costs for the techniques, and a further US $ 10 million was the cost of extending the mission including the costs for the "hyperextended mission" after the encounter with Borrelly. According to another source, the total mission cost was US $ 152.3 million, including $ 94.8 million for development, 43.5 million for launch, 10.3 million for operations and 3.7 million for science.

Mission objectives

The primary objective of the mission was to test a dozen new techniques, some of them in an experimental stage. One wanted to gain experience with the reliability of the technology before it was to be used in an expensive scientific mission. The first use of new technology in a space mission always involves a certain risk, despite all the tests. As a so-called “high-risk, high-benefit mission”, this mission took on the risks of all twelve technologies at the same time. A failure of a technology would not in this case be synonymous with the failure of a mission objective, if the data enable the analysis of the errors and the possible uses and limits of the technology for later missions can be determined.

In general, the objectives of the mission fell into two categories: first, technologies that make spacecraft smaller or lighter and cheaper, then technologies that make spacecraft more autonomous and thus save staff and time on large antenna systems such as the Deep Space Network (DSN) and ground stations.

In addition to the technology itself, other goals were tested: The limits of time-limited development time for deep-space missions were tested. The initial study for DS1 with only very vague specifications came just 39 months before the start, an unprecedented short time for such a mission. The actual mission goals were set 26 months before the start. The start was originally planned for July 1998 in an extremely ambitious schedule, but it became clear that the integration of untested components and software represented an uncontrollable risk and so the start was postponed to October and the original flight plans and flight destinations were changed. The initial planning included Comet 76P / West – Kohoutek – Ikemura and Asteroid (3352) McAuliffe as targets . Part of the software was not ready even when it was started and had to be added later.

The control and monitoring of the mission from the ground with the Advanced Multi-Mission Operations System (AMMOS) was also new. This form of ground control allows many different missions to share facilities and personnel on one schedule. For this reason, there are standardized basic functions for missions that can be adapted to the respective mission and the mission is given certain specifications so that it can be operated effectively with AMMOS. The engineers can plan missions and complex flight maneuvers in advance. AMMOS can send preplanned commands to the spacecraft at a given time and process and archive all kinds of received data, calculate telemetry data and monitor technical data of the spacecraft. DS1 was the first mission that was controlled exclusively with the capabilities of AMMOS. Most of NASA's scientific missions now work with AMMOS for ground control and even older missions have now been taken over by AMMOS.

technology

The probe body has the dimensions 1.1 × 1.1 × 1.5 m, with the attachments and foils 2.1 × 1.7 × 2.5 m. When the solar modules are extended, the span is 11.8 The take-off weight was 490 kg, of which 31 kg were hydrazine for the attitude control and 82 kg xenon for the ion drive. The solar modules have an output of 2400 watts.

The main processor was the IBM RAD6000 , on which the real-time operating system VxWorks was used. Unlike other missions, there was no backup computer that could take control during a computer failure or restart. Many components such as the battery, hydrazine engines, inertial measurement system, star tracker and sun sensor were inexpensive standard parts, some leftover parts from other missions were installed. The remaining high-gain antenna came from Mars Pathfinder , the control electronics for the control nozzles were similar, the computer was identical and the software was also based on Pathfinder. Much of Pathfinder’s test stands and test equipment was also re-used. A 30 cm high-gain antenna for X-band , three low -gain antennas and a comparatively small Ka-band antenna were used for communication .

In addition to the standard operational functions, twelve new technologies were installed and tested in the probe. For various reasons, redundant systems were largely dispensed with, although some systems were able to take over the functions of other systems. Compared to the technology tests, the scientific results were only a by-product.

NSTAR ion propulsion

The most important goal of the mission was the first use of the ion drive "NASA Solar Technology Application Readiness" (NSTAR) under real conditions. The xenon ion drive has a diameter of 30 cm and would have required a maximum of 2500 watts of electrical power at full thrust, 500 watts at minimum power. Since the solar modules could not produce as much power and additional power was needed to operate the other systems, the ion drive could not be tested at maximum power. The engine could be operated in 112 levels, the highest level reached was 90. The thrust was 0.09 Newtons at maximum and 0.02 Newtons at minimum.

When testing the ion drive, the prognoses from the tests were checked in the laboratory for performance, continuous operation, wear and tear and efficiency. Apart from the difficulties at the beginning of the mission, the drive worked as expected and proved its suitability. Before it was deactivated on December 18, 2001, the drive had been started over 200 times and was in operation for a total of 16,246 hours, during which time it consumed 72 kg of xenon. The reliability of the ion propulsion system was a key result for the Dawn spacecraft , which was equipped with three NSTAR ion propulsion systems.

On board was the so-called "Ion Propulsion System Diagnostic Subsystem" (IDS), which was primarily intended to check the function of the ion drive. Since the effects of the “Ion Propulsion System” (IPS) were unproblematic, this instrument was reprogrammed in order to be able to obtain scientific data. It consists of twelve sensors, including an antenna for plasma waves and two magnetometers with search coils. A magnetometer did not work and was probably damaged by a strong alternating current magnetic field before it was started. The second magnetometer worked in three axes in the frequency range between 10 Hz and 50 kHz. The measurement range was 100 nT with a resolution of 1 pT.

Solar modules

The new solar modules were equipped with "Refractive Linear Element Technology" (SCARLET), a form of solar concentrators. The modules have 720 Fresnel lenses made of silicon, which cast light on 3600 highly efficient multi-layer solar cells made of indium gallium phosphide , gallium arsenide and germanium . They have 2400 watts of power at 100 volts, whereby the power decreases with increasing age and distance from the sun. They have around 15–20% more power than conventional solar modules of the same size up to that point. Since the ion drive requires a lot of energy, the performance of the new solar modules first had to be checked under various conditions. The aging process under space radiation was also tested.

Autonomous Navigation (AutoNav)

Previously, spacecraft were controlled by tracking radio signals from the ground and calculating position and course from them. Occasionally, recordings are made of the target object in order to determine the position more precisely. Commands from the ground then ignite the engines for the fine correction. To do this, regular radio communication must be maintained. AutoNav now took on this role of ground team. Conventional spacecraft with chemical propulsion are only strongly accelerated in the launch phase. Most of the time passes with an unpowered trajectory that corresponds to a ballistic object, interrupted only by brief jolts of the hydrazine position control nozzles or by individual uses of the main drive for a few minutes or seconds. Deep Space 1, on the other hand, was driven for days and weeks and could not turn the antenna to earth during this time. The permanent drive was constantly changing the trajectory and therefore a different type of navigation was required.

AutoNav was able to independently recognize the orientation of the probe in space, align the probe and control the ion drive. It recognized the distance to the sun and the amount of electrical power available for propulsion. At the start, the orbits of 250 asteroids and the positions of 250,000 stars were stored in the computer. AutoNav knew the orbits of asteroids and the positions of the fixed stars and was able to determine its own position based on the parallax . In the beginning, images of four to five asteroids were taken three times a week, and later once a week of seven asteroids. During the recording, the ion engine was switched off so that the probe could turn the camera in the direction of the asteroids. The recordings were evaluated by AutoNav and commands put the ion drive or the hydrazine drives into operation in a targeted manner. AutoNav relied on good recordings to function correctly, but the camera was affected by unexpected stray light and the optics distorted the recordings, especially in the edge area, both of which had negative effects. Overall, it became apparent in the course of the mission that the quality of the recordings was a critical factor.

Autonomous remote agent

The “remote agent” was a kind of autopilot for a predetermined course; it could work through a complex, predetermined plan for the experiments without requiring commands or monitoring from the ground station. The remote agent set the targets for AutoNav. The software allowed the spacecraft to make independent decisions, to automatically switch components or backup systems on or off, whereby only general specifications were made from the ground. The ground team relies on the agent to find a way and make appropriate decisions to meet these requirements, even in the event that systems fail or unplanned events occur.

The software also included a roadmap that specifies what to do at a specific time or event. The decisions were made based on the system status, the restrictions a mission is subject to, and the general mission specifications. Accordingly, the system issues a series of commands to the corresponding subsystem. It monitors how the systems react to the commands and repeats them or gives changed commands if the result is different from what was anticipated. The remote agent was not operational during the entire mission and software updates were uploaded during the mission.

Beacon monitor

The beacon monitor operations experiment was a simple method of communication between the probe and the ground station. The probe works as a “beacon” and only emits a single signal. The previous missions relied on the regular transmission of telemetry data. This requires the frequent and cost-intensive use of the DSN as well as personnel to evaluate the data in the Mission Control Center. The probe was equipped with enough intelligence that it was informed about its condition and could decide whether an intervention from the ground station was necessary. The beacon monitor only emitted four different simple signals that give the ground station general information about the state of the probe. The simple signal was not coded and simple antennas 3 to 10 meters in diameter were sufficient for reception, so that the DSN was not absolutely necessary. A “green” signal showed normal progress, an “orange” signal showed something unexpected, but that the probe was able to solve the problem and all values ​​are acceptable, or that contact is required within four weeks. A “yellow” signal indicated that the probe wants to send data or that an existing development over a long period of time could lead to a problem in the future, or that contact is required within a week. A “red” tone, on the other hand, indicated a serious problem that the electronics could not control and required rapid intervention from the ground station. The Beacon Monitor works with activity-dependent limits. A value of a measurement can be outside the limits when there is little activity, but it can be completely within the normal range during activity. The beacon monitor did not control the mission and was not in operation all the time, but was only operated on a test basis.

Miniature Integrated Camera Spectrometer (MICAS)

This twelve kilogram instrument can fulfill several tasks at the same time: It works as a camera, as an ultraviolet image spectrometer and as an infrared image spectrometer. MICAS also provided images for the AutoNav. It has two black-and-white cameras, a UV and an IR image spectrometer, all of which use a 10-centimeter telescope with mirrors made of silicon carbide . Of the two cameras in the visible area, one is a CCD pixel camera, the other has a CMOS active pixel sensor . The spectrometers have to scan individual points on the target object in order to obtain data. MICAS should use the UV spectrometer to detect hydrogen distributed in the solar system. The UV channel should work in the range between 80 nm and 185 nm wavelength, but could not obtain any useful data. The fault was somewhere in the chain after the photon detectors. In practice, stray light in the camera reduced the scientific value of the data and made evaluation by AutoNav more difficult. Some design changes and a different mounting of the camera could fix this problem on future missions.

Plasma Experiment for Planetary Exploration (PEPE)

The “Plasma Experiment for Planetary Exploration” (PEPE) is a six kilogram multi-purpose instrument for studying plasma and charged particles. The device can recognize electrons and ions. It was tested how the ion drive affects the measurement results. It worked like several physical observation devices, besides studying the effects of ion propulsion on the surface of the probe and on the instruments, and studying how the ion propulsion interacts with the solar wind. It was also able to obtain scientifically interesting data from the flyby on the asteroid.

In January 1999, DS1 and Cassini were conveniently positioned to each other and jointly took measurements of the solar wind for 36 hours, with both probes approximately 0.5 AU apart. For full functionality, the device worked with a voltage of 15,000 volts. Most of the time, the solar panels could not generate enough energy, so in this case it had to be operated at voltages of around 8,000 volts. In this case, complex and heavy ions could not be detected. The measuring range of PEPE for electrons is 10 eV to 10 keV and for ions from 3 eV to 30 keV.

Small deep-space transponder

The three kilogram "Small Deep-Space Transponder" is intended to improve the telecommunications hardware. It contains a command detector, telemetry modulation, a tone generator for the "beacon" mode, as well as control functions. It can send and receive in the X-band as well as send in the Ka-band. Low weight and dimensions are possible through the use of integrated microwave chips made of single crystal gallium arsenide, a dense arrangement and the use of application-specific silicon ICs . Various tests were carried out simultaneously in the X-band and in the Ka-band in order to be able to compare the results. 2001 Mars Odyssey and other Mars missions later used this transmitter because it has proven itself.

Ka-band solid-state power amplifier

This very small and 0.7 kg light amplifier with a transmission power of 2.3 W allows higher data rates due to the higher frequency in the Ka-band compared to the communication in the X-band that was common up until then. The Ka-band allows the same data rate with a smaller antenna, but is more susceptible to weather influences with terrestrial reception. The transmitter was not only used for communication, but also for general experiments on communication in the Ka-band. At the time of the mission, only the Goldstone complex was equipped with the appropriate technology for Ka-Band by the DSN , so that all experiments were carried out with the station in Goldstone.

Low power electronics

It was extremely energy-saving microelectronics, insensitive to radiation. The experiment worked with low voltages, had low-activity logic, energy-saving architecture, and power management. A ring oscillator , transistors and a multiplier with minimal power consumption were tested . A dosimeter showed at the end of the mission an overall radiation exposure of 450 Gray .

Multifunctional structure

The multifunctional structure is a further step towards weight savings, fewer components and more reliability. So far, load-bearing functions, temperature control functions and electronic functions have been developed separately and housed in different parts. All parts were then connected with large connectors and cable harnesses for power supply and data transmission. The multifunctional structure combines temperature control and electronics and at the same time replaces one of the panels of the probe body. It has copper- polyimide foil on one side and built-in heat transfer devices. The surface is used for heat radiation and the wiring is created with the polyimide film. Flexible connections between the foils enable power supply and data distribution.

Power activation and switching module

This module consists of eight very small electrical switches that are arranged in redundant pairs so that four electrical consumers can be monitored. The switches register voltage and current and can limit the current.

course

  • Deep Space 1 was launched on October 24, 1998 on a Delta II 7326 launcher. The launcher was the first of its kind. She had SEDSAT-1 on board as an additional payload . Three new technologies were used right from the start.
  • The ion drive was tested for the first time on November 10th; the drive switched itself off after 4.5 minutes and could not be restarted at first. On November 24th, the ion drive was restarted. A temporary short circuit is suspected to be the cause of the problem.
  • On November 12th, the Star Tracker was malfunctioning. It is a purchased standard model. A problem for the AutoNav software was unexpected stray light in the MICAS camera, which made it difficult to evaluate the images. A first software update enabled continued operation. AutoNav was able to pinpoint the position to within 2,000 kilometers.
  • At the beginning of February 1999 the software of the onboard computer was updated for the first time in order to carry out all tests, to be able to use the first results of the tests and to eliminate programming errors.
  • On March 15, 1999, after a long unpowered flight, ion propulsion was resumed. The drive worked for six and a half days at a time; for half a day a week the probe performed other tasks and directed the antenna towards the earth.
  • The remote agent was tested in May 1999, and various problems were simulated for the remote agent by the ground team. One of the first problems simulated was that the camera wouldn't turn off. The remote agent gave several shutdown commands and then devised an alternative plan. The remote agent turned the probe so that pictures of asteroids could be made and then turned the probe in the direction of flight and started the ion drive. During the subsequent drive phase, the remote agent and the drive stopped unexpectedly due to a software error. The remaining simulated system errors were correctly identified and a suitable solution found.
  • At the beginning of June 1999 the probe received a second software update. It took three days for the DSN to upload 4 megabytes of data, after which the computer was shut down and restarted.
  • On June 14, 1999, AutoNav made the first completely independent course correction. There was no predetermined plan, the system had to start from scratch. When calculating the correct thrust direction, direct sunlight would have fallen into the camera and the Star Tracker. The system therefore divided the direction into two burning phases in different directions, which together give the desired direction. This process is called vectoring burn. As the mission progressed, it became evident that the limits of the camera were a limiting factor for AutoNav.
  • At the end of July 1999, a few days before the encounter, the target asteroid 1992 KD was renamed (9969) Braille . However, the asteroid is so small and dark that it was not visible with the camera even three days before the flyby and AutoNav could not detect it in the recordings. When the discovery was finally made with the help of the ground team and specialized software on recordings, the asteroid was more than 400 km away from the previously calculated location. The probe was immediately set on the new course.
  • A little more than half a day before the flyby, DS1 went into save mode due to a software error. The ground team worked feverishly to fix the problem and put the probe back into flight mode. The flyby happened on July 29, 1999 at a speed of 15.5 km / s at a distance of only 26 km, the plan was 15 km, but at the time it was the closest flyby of a probe to an object. At the time of the encounter, Deep Space 1 was 1.25 AU or 188 million kilometers from Earth. Due to several problems, AutoNav could not find and identify the asteroid with its very irregular shape until the end and the camera could not be aimed precisely enough. However, it was possible to make infrared spectrograms and the plasma detector also provided data. Fifteen minutes after the flyby, the probe turned and took another picture from a distance, showing the asteroid.
  • It took the probe a day to transmit all data to Earth. The spectrographic evaluation shows a profile that corresponds to the basalt of Vesta.
  • At the beginning of August 1999 Nasa extended the mission and the planned new target was (4015) Wilson-Harrington and then 19P / Borrelly , the targets changed accordingly from a mission to test components to a scientific mission.
  • On November 11, 1999, the Star Tracker failed completely after the device had inexplicable interference during the entire mission. Attempts to reactivate it failed. With the help of the DSN, the antenna of the probe could be aligned with the earth in order to load new software. The following months passed with finding a solution to the problem and developing new software and testing it in the simulator. During this time the probe could not be propelled, (4015) Wilson-Harrington thus became inaccessible.
  • On May 30, 2000, a fifth version of the software could be loaded, which could use the MICAS camera instead of the Star Tracker for position control. The DSN gave the mission additional communication time for software updates and tests. MICAS has a much smaller observation field and works completely differently than the Star Tracker. The position control was now carried out by aiming at a star with the camera: One star serves as the target during the propulsion phase, another for aligning the antenna with the earth. Towards the end of June 2000, the engine was able to be used again after various tests.
  • On October 11th, DS1 and Earth were in opposite positions, with the sun in between. The probe disappeared behind the sun for two days during the conjunction , during which time no communication was possible.
Core of the comet Borrelly
  • In March 2001 the probe received its sixth and final update. It consisted of over 4MB of data, divided into 267 files. It took four days to receive all of the data via the probe's main antenna, which is only 30 cm in size, and the distance to earth was 2.1 AU. The software has been specially adapted to better distinguish the comet's core from the tail.
  • On September 22, 2001, Deep Space 1 flew past comet 19P / Borrelly at a distance of about 2200 km. This time it was also possible to take pictures and spectrograms, measure the angle and energy of electrons and ions, examine the ion composition and analyze the magnetic field. The probe was not designed for an encounter with a comet, it lacked a shield against dust particles etc. so that this observation carried a high risk of damage; however, there was no major damage and all data could be sent.
  • After the flyby, the goals were set again in a “hyperextended mission”. All systems were tested again in order to obtain comparative data and to be able to measure the wear and tear or degradation caused by space radiation. There were also risky tests carried out and tests that tested the limits of technology. All tests were successfully completed. The probe was thus in operation for almost three years, during which time it circled the sun twice, while the earth circled the sun three times.
  • On December 18, 2001, Deep Space 1 was deactivated. There was no other mission goal that could have been achieved with the remaining fuels. At this point in time, the ion thruster was in operation for a total of 16,265 hours or 677 days and during this time it achieved a cumulative acceleration of ∆v = 4.3 km / s. 73.4 kg, i.e. more than 90% of the xenon stocks, were used for this. The software has been modified so that the transmitter will no longer operate, and data storage has been blocked to prevent memory overflow. All systems no longer needed and the main antenna were switched off, only the three emergency antennas remained in operation. At the end of the mission, only a small amount of hydrazine remained, which was sufficient for a few months. As soon as the supplies were used up, the probe could no longer orient the solar panels to the sun and thus had no more electricity for operation. However, the probe will continue to orbit the sun on its orbit.
  • Contrary to expectations, there was one more occasion when the probe's Ka transmitter could have been useful. In early 2002, new tests were developed to examine the effects of stormy weather on Ka frequency reception. At the time, there was no spacecraft apart from DS1 that could have delivered the required signals in the Ka-band. Assuming that the probe had still been aligned with the sun, the probe's antenna would also have been aligned with the earth for a while during the opposition on March 10, 2002. On March 2nd and 6th an attempt was made to contact two DSN stations, but the attempts were unsuccessful as predicted and no signal could be found.

Results

Deep Space 1 was a complete success from a technical point of view. All mission objectives were met or exceeded by the end of the 11 months of the primary mission up to September 1999. The NSTAR ion propulsion proved its worth and cleared the way for the Dawn mission, which used three such propulsion systems. Prior to the test, there were concerns that the ejection of the ion thruster could possibly affect the radio link or the scientific instruments. The PEPE instrument was on board to identify and quantify these effects, but the plasma output did not cause any problems. The Small Deep-Space Transponder has proven itself and has been used in several missions since then.

The Star Tracker was not part of the test program, but a purchased component that was actually considered to be very reliable. The failure almost led to the end of the mission, but the problem-solving with new software was a success story in its own right. In contrast to Deep Space 1, scientific missions usually have several redundant star trackers.

From a scientific point of view, Deep Space 1 was able to record quite a lot, including the first measurement of an asteroid magnetic field in Braille. While it failed to get close-up braille images, the Borrelly flyby was a complete success, which provided some new and surprising insights into comets.

The Dawn, New Horizons and some Martian missions were able to directly leverage the experiences of DS1 by using these technologies in science missions.

See also

Web links

Commons : Deep Space 1  - collection of images, videos and audio files

Individual evidence

  1. ^ A b Marc D. Rayman: The Successful Conclusion of the Deep Space 1 Mission: Important Results without a flashy title . In: Space Technology . tape 23 , no. 2-3 , 2003, pp. 185 (English, online [PDF]).
  2. ↑ a b NASA - NSSDCA - Spacecraft - Details. Retrieved June 15, 2017 .
  3. ↑ NASA (Ed.): Deep Space 1, Launch, Press Kit October 1998 . S. 3 ( online [PDF]).
  4. ↑ a b c Marc D. Rayman, Philip Varghese, David H. Lehman, Leslie L. Livesay: Results from the Deep Space 1 technology validation mission . In: Jet Propulsion Laboratory (ed.): Acta Astronautica . tape 47 , 2000, pp. 475 ff . ( Online [PDF]).
  5. ↑ Comet Space Missions . In: SEDS.org . Retrieved November 20, 2016.
  6. ↑ AMMOS. (No longer available online.) NASA, archived from the original on November 14, 2016 ; accessed on June 16, 2017 . Info: The archive link was inserted automatically and has not yet been checked. Please check the original and archive link according to the instructions and then remove this notice. @1@ 2Template: Webachiv / IABot / ammos.jpl.nasa.gov
  7. ↑ NASA (Ed.): Deep Space 1, Launch, Press Kit October 1998 . S. 7 ( online [PDF]).
  8. ↑ Deep Space 1 (DS-1) . www.bernd-leitenberger.de. Retrieved July 16, 2012.
  9. ↑ NASA (Ed.): Deep Space 1, Launch, Press Kit October 1998 . S. 31 ( online [PDF]).
  10. ↑ NASA (Ed.): Deep Space 1, Launch, Press Kit October 1998 . S. 24 ( online [PDF]).
  11. ↑ GRC - NSTAR Ion Thruster. Retrieved June 3, 2017 .
  12. ↑ David DeFelice: NASA - Deep Space 1 Ion Propulsion System Operation Sequence and Status. Retrieved June 3, 2017 .
  13. ^ The Ion Propulsion System (IPS) Diagnostic Subsystem (IDS). Retrieved July 17, 2017 .
  14. ↑ NASA (ed.): Contributions to Deep Space 1 . April 14, 2015 ( online [accessed June 3, 2017]).
  15. ↑ NASA (Ed.): Deep Space 1, Launch, Press Kit October 1998 . S. 26 ( online [PDF]).
  16. ↑ NASA Space Science Data Coordinated Archive Header Plasma Experiment for Planetary Exploration (PEPE). Retrieved July 17, 2017 .
  17. ↑ Dr. Marc Rayman's Mission Log, December 9, 1998. Retrieved May 25, 2017 .
  18. ↑ Marc D. Rayman: The Successful Conclusion of the Deep Space 1 Mission: Important Results without a flashy title . In: Space Technology . tape 23 , no. 2-3 , 2003, pp. 185 (English, online [PDF]).
  19. ↑ Dr. Marc Rayman's Mission Log October 24th. Retrieved May 25, 2017 .
  20. ↑ Dr. Marc Rayman's Mission Log, November 10, 1998. Retrieved May 25, 2017 .
  21. ↑ Dr. Marc Rayman's Mission Log, November 24, 1998. Retrieved May 25, 2017 .
  22. ↑ Dr. Marc Rayman's Mission Log, November 12th. Retrieved May 25, 2017 .
  23. ↑ Dr. Marc Rayman's Mission Log, February 13, 1999. Retrieved May 25, 2017 .
  24. ↑ Dr. Marc Rayman's Mission Log, May 19, 1999. Retrieved May 28, 2017 .
  25. ↑ Dr. Marc Rayman's Mission Log, May 23, 1999. Retrieved May 28, 2017 .
  26. ↑ Dr. Marc Rayman's Mission Log, June 12, 1999. Retrieved May 28, 2017 .
  27. ↑ Dr. Marc Rayman's Mission Log, June 20, 1999. Retrieved May 28, 2017 .
  28. ↑ Dr. Marc Rayman's Mission Log, July 11, 1999. Retrieved May 28, 2017 .
  29. ↑ Dr. Marc Rayman's Mission Log, July 29, 1999. Retrieved May 28, 2017 .
  30. ↑ Dr. Marc Rayman's Mission Log, August 1, 1999. Retrieved June 7, 2017 .
  31. ↑ a b Dr. Marc Rayman's Mission Log, August 8, 1999. Retrieved June 7, 2017 .
  32. ↑ Dr. Marc Rayman's Mission Log, January 16, 2000. Retrieved June 7, 2017 .
  33. ↑ Dr. Marc Rayman's Mission Log, July 4, 2000. Retrieved June 7, 2017 .
  34. ↑ Dr. Marc Rayman's Mission Log; October 29, 2000. Retrieved June 8, 2017 .
  35. ↑ Dr. Marc Rayman's Mission Log, March 18, 2001. Retrieved June 8, 2017 .
  36. ↑ Susan Reichley: 2001 News Releases - NASA Spacecraft Captures Best-Ever View of Comet's Core. Retrieved June 8, 2017 .
  37. ↑ Dr. Marc Rayman's Mission Log, September 11, 2001. Retrieved June 8, 2017 .
  38. ↑ Dr. Marc Rayman's Mission Log, December 18, 2001. Retrieved June 8, 2017 .