• PROJECTS
  • Advocacy and Education
  • Extrasolar Planets
  • Innovative Technologies
    • Carl Sagan Fund for the Future
    • LIFE Experiment: Phobos
    • LightSail - The Future of Solar Sailing
    • Pioneer Anomaly
      • Project Updates
      • Facts
      • Press Room
    • Planetary Microphones
    • SETI Optical Telescope
    • SETI@home
  • International Mission Participation
  • Mars Exploration
  • Near Earth Objects
  • Search for Extraterrestrial Life
  • Past Projects
• THE PLANETARY REPORT
• PLANETARY RADIO
• BLOG
• SOCIETY UPDATES

browse projects: Apophis Mission Design Competition Carl Sagan Fund for the Future Catalog of Exoplanets FINDS Exo-Earths International Year of Astronomy 2009 LIFE Experiment: Phobos LightSail - The Future of Solar Sailing Messages from Earth Observing Earth Pioneer Anomaly Planetary Microphones SETI Optical Telescope SETI Radio Searches SETI@home Shoemaker NEO Grants Space Advocacy Space Information

browse space topics: 2001 Mars Odyssey Asteroids and Comets Cassini-Huygens Chandra X-Ray Observatory Chandrayaan-1 Chang'e 1 Compare the Planets Dawn Deep Impact Earth Extrasolar Planets Genesis Hayabusa (MUSES-C) Hubble Space Telescope Jupiter Kaguya (SELENE) Lunar Reconnaissance Orbiter (LRO) Mars Mars Exploration Rovers Mars Express Mars Global Surveyor Mars Reconnaissance Orbiter Mars Science Laboratory Mercury MESSENGER Moon Near Earth Objects Neptune New Horizons Past Missions Phoenix Planetary Analogs Planetary Exploration Timelines Pluto and Charon Private Missions Rosetta Saturn Search for Extraterrestrial Intelligence SMART-1 SOHO Space Imaging Spitzer Space Telescope Stardust Trans-Neptunian Objects Ulysses Uranus Venus Venus Express Voyager Weekly Planetary Radio Trivia Contest

Projects: Pioneer Anomaly

Pioneer Anomaly Project Update

A Letter From the Project Director

Dear Members of The Planetary Society:

First of all, thank you very much for your continuing support for our efforts to identify the nature of the Pioneer anomaly. In this letter, I would like to provide you with some details on our current work.

As of March 2007, the existence of the anomaly is confirmed by seven independent investigations using different navigational codes -- the signal is present in the Doppler data received from both Pioneers 10 and 11. The most important question now is to identify the cause of this anomalous frequency drift discovered in the Pioneer data.

Previously, we reported on our successful retrieval of a very significant amount of Pioneer data both radio-metric Doppler and telemetry information (that is, all the housekeeping data documenting spacecraft health during their entire missions). Thus, the original 1998-2002 analysis of the Pioneer anomaly was done with just 11.5 years of Pioneer 10, covering heliocentric distances from 40-70 AU (1 AU is the mean distance of Earth from the Sun, about 150 million kilometers or 93 million miles) and with 3.75 years of Pioneer 11 for heliocentric distances from about 20-32 AU.

The new data set is much enlarged: we have almost 30 years of Pioneer 10 data (for distances 4.2-82 AU) and 20 years of Pioneer 11 data (for distances 1-33 AU). These two sets of Doppler data are the primary sources we will use to investigate physical characteristics of the anomaly and to establish its temporal behavior (change of the anomaly's magnitude with heliocentric distance, if any), its true direction (heliocentric, Earth-pointing, along the spin-axis, along the velocity vector, etc.), and the onset of the effect that is apparent in the Pioneer 11 data during the Saturn flyby.

For more details, please take a look at the two recent papers detailing the recovery efforts: The Study of the Pioneer Anomaly: New Data and Objectives for New Investigation and The Pioneer Anomaly: Seeking an explanation in newly recovered data.

The availability of this new and much extended dataset is the most significant change since any other investigation performed to date. However, its analysis is technically complex, as we will have to "re-fly" the 30-year Pioneer missions. In other words, we will have to virtually "launch" the Pioneers, acquire their signals, perform the first orientation maneuver, and go arc-by-arc all the way to Jupiter and Saturn (for Pioneer 11), then fly by outer planets and leave the solar system while still transmitting scientific and engineering information -- and do that until the last data point. Our objective is to obtain the most precise trajectory solutions using the recently recovered data. However, the amount of work and its complexity are overwhelming; furthermore, this work must be done very carefully, with attention given to all pieces of information.

Recovering the Data from the Files

Before initiating the detailed work with Doppler data, we first studied the data itself, which is a standard part of any data analysis. We retrieved about 650 files for each of the two spacecraft from four primary sources. In particular, three sets of files came from our colleagues at the Jet Propulsion Laboratory (JPL); one set came from a data archive at JPL, another set was produced by reading more than 400 tapes of navigational data at JPL, and finally, the last set came from the NSSDC (National Space Science Data Center). This data covers nearly entire durations of the Pioneer missions and is critical for our investigation of the effect.

From the beginning, the data that was received from NSSDC presented us with an issue of corruption. Specifically, each record in the file structure had an extra 8045th bit that was added at the end of a normal 8044-bit word. This corruption was due to the fact that the files were handled by many different computers, each of them affecting the file record. This problem was already known to us from the earlier work of Craig Markwardt of Goddard Space Flight Center (GSFC). Craig's step-by-step description of how to correct the issue helped us to successfully remove that extra bit. After that procedure, our software recognized and handled the data files just fine. We were able to successfully combine the data to initiate the analysis with the new, much-extended set of Pioneer data. So, we excitedly began work on the extended set in September 2006.

We quickly learned that, in general, the entire combined record of the retrieved data looks good; however, it presented us with another set of unexpected challenges that slowed down the progress with data analysis. Specifically, it turned out that the single tracking data file created from the four primary sources was not usable for analysis. We quickly learned that some of the recovered data were incompatible with current navigation software architecture. This, in particular, led to a frequency band misidentification problem that was easily corrected. In addition, we found some data points were duplicated while other points were missing. All of these issues required special investigation and development of countermeasures to correct the issues.

We also noticed that during some periods the data showed the presence of an artifact due to inappropriate data handling. This issue was seen after we started to assemble the data and for a "virtual re-flight" of the spacecraft -- a process that required us to pass the data through JPL's orbit determination program (ODP). Examination of the residuals after several initial data fits with the ODP revealed the expected thick line of residuals with the familiar structures; however, in addition to the anticipated curve with residuals, we also saw also several unusual lines that repeated behavior of the main curve. To investigate the issue, we plotted residual frequency vs time for 1979-1982 and 1984-1986; the residuals had multiple lines with similar behavior separated from each other by the same shift of +/- 1042Hz. After some effort, we confirmed that the problem originated from the NSSDC files. All files retrieved from other primary sources were good and ready for our high-accuracy work.

We decided to look for an approach that would allow us to more effectively deal with the apparent problems with the quality of the recovered data. An alternative approach was to create tracking data files from individual sources (more than 600 files!) instead of working with the single tracking data file created earlier for each spacecraft. This was an unwanted but a necessary step, as some of the files were already suitable for the analysis and did not require any additional processing of the legacy data. We decided to engage experts at JPL to help us in correcting the problems evident in some files. As a result, we identified a strategy that, without going into much technical detail, may be represented as follows: first, we examine each of the more than 600 files and then identify the set of commands necessary to properly interpret the data already present in the file; next, we implement these commands (usually there are more than 250 commands needed to a correct single file); then we pass the corrected orbit data file through the orbit determination program to see if the problems were corrected.

This strategy led to excellent results. After receiving many corrected files, we see that the artifact is not a big problem. In fact, we were able to initiate analysis of the early Pioneer 11 data with the new recovered files and plan to report on our work in updates in the near future.

This complex effort is a part of the overall process of data certification, and it must be done before we can proceed with any serious work on the anomaly. As a result, we already have a significant extension of the valid two-way Doppler data for both Pioneers 10 and 11. Since December 2006, we've been creating usable files from NSSDC archival tracking data files (ATDFs) that were initially corrupted and/or had missing information (ramp frequency, spin-bias, and other calibrations).

After we complete the data certification process, we will need to edit the set to eliminate the data that are clearly erroneous. So far, we see a reasonably clean data set and do not expect to have to do much editing. We expect that we will move onto the editing stage and conclude with data preparation phase of the task by the end of summer. Although, we are very anxious to see the ultimate results sooner, we must work very carefully and explore every piece of new information as we investigate the Pioneer anomaly.

Putting this into perspective, just think about the fact that an effort of assembling the data record of such a length had never been done in the past for any mission we have flown. In terms of computers, during the time since the Pioneer 10 and 11 launches, we moved from punch-cards to main-frame computers and now to PC. Software underwent a major evolution -- progressing from FORTRAN 66 to FORTRAN 99, and now the standard is the object-oriented language C++. Our models for solar system dynamics (solar system ephemerides), and models for and data of monitoring the interplanetary environment have all drastically evolved. The Deep Space Network had seen major developments also -- many stations were built, rebuilt, commissioned, de-commissioned, moved, etc. People -- the main resource -- changed in the most profound way. We are very fortunate that some of the JPL personnel that were present during Pioneer Mission Operations are helping us to re-generate the orbit data files of the desirable quality and completeness.

Click to enlarge > The start of Pioneer 10's data files Just after Pioneer 10 launched from Cape Canaveral, Florida, on March 2, 1972, navigation team members Tony Liu (on the phone) and Phil Laing (reading printout) check out the first navigational data from the spacecraft. Credit: NASA / JPL

At the very beginning of this project, we expected that there would be problems with the data. We now see the problems, but we expect to correct all of them in the near future.

Analysis of the Retrieved Data Begins

The progress with retrieval and initial analysis of the Pioneer Doppler data was reviewed during the Second Pioneer Explorer Collaboration meeting held in Bern, Switzerland during February 19-23, 2007, held at the International Space Science Institute. The meeting was attended by 35 researchers from Germany, France, Italy, Norway, Portugal, Netherlands, Canada, and the U.S. The focus of the meeting was to review our preparations for the upcoming analysis of the newly recovered data with the ultimate goal of determining the origin of the Pioneer anomaly.

The meeting participants reported on the efforts to recover and use the complete set of radio Doppler and telemetry records of both spacecraft. In particular, they discussed the status of the Pioneer Doppler data retrieval and also initial analysis of the new Pioneer Doppler data sets. Several groups have now processed the data previously analyzed and have independently verified the presence of the anomaly.

The team also made progress utilizing the recently recovered Pioneer telemetry data. Researchers discussed initial analysis of the Pioneer telemetry and the status of the auxiliary data (propulsion maneuvers, spin information, etc.), ultimately leading to the discussion of the critical issue of whether or not the Pioneer anomaly is due to anisotropic thermal emission on the spacecraft. The meeting was an important milestone to define a strategy for studying the effect of on-board generated small forces with this telemetry data, especially in conjunction with the analysis of the entire set of the Pioneer Doppler data.

Our thermal modeling of the Pioneer vehicles is progressing very well. We finished the development of the geometric mathematical models of the spacecraft that include geometry and properties of most of the important spacecraft components and surfaces. We are now working on the thermo-dynamical model of the vehicles. At this stage, we have a very good understanding of heat re-distribution within the craft and soon will be ready to compute the heat flow to the outside of the craft. Soon, we will be able to tell whether or not heat contributes to the formation of the anomaly.

Again, I thank you very much for your continuing support and promise to bring more news on the Pioneer anomaly investigation in the near future.