By late 1998, JPL had settled on an MSR mission design based on the Mars Orbit Rendezvous (MOR) mode. This was not surprising, since the Pasadena laboratory had staunchly advocated MOR since the early 1970s. At that time, JPL was responsible for building the Viking Orbiter, and MSR missions in the late 1970s/early 1980s were expected to be based on Viking hardware. The MOR mode would require use of an Orbiter; MOR's chief rival, Direct-Ascent, would not because it would launch samples directly from the surface of Mars back to Earth. No Orbiter meant no role for JPL in Viking-based Direct-Ascent MSR; JPL thus supported MOR. This institutional preference became thoroughly ingrained by the early 1980s.
In MOR MSR's basic form, samples would reach Mars orbit in an ascent vehicle. An orbiter would perform rendezvous and collect the samples, then would depart Mars orbit for Earth. Splitting the Mars ascent and Earth return functions between ascent and orbiter vehicles would enable a smaller, lighter Mars lander than in Direct-Ascent, and thus would trim overall mission mass. The reduced mass of MOR MSR would mean that the mission could leave Earth on a smaller, cheaper launch vehicle or could include more science payload - for example, a rover for sample collection beyond the immediate landing site.
One can argue, however, that MOR increases mission complexity and thus risk. JPL's 1998-1999 MOR MSR plan aimed to reduce risk by collecting samples from two different martian sites using landers launched from Earth during two Earth-Mars transfer opportunities (specifically, in 2003 and 2005). After completing its 90-day sample collection mission, each lander would launch to Mars orbit a Mars Ascent Vehicle (MAV) bearing an Orbiting Sample (OS) canister. To help keep its MSR mission under a strict cost cap, NASA invited the French space agency, Centre National d'Etudes Spatiales (CNES), to provide the MSR orbiter.
At the August 1999 AAS/AIAA Astrodynamics Specialist Conference in Girdwood, Alaska, a team of engineers from JPL and another from JPL contractor Charles Stark Draper Laboratory (CSDL) presented papers in which they examined how the CNES orbiter might perform rendezvous with the 2003 and 2005 OSs. They proposed a complex three-phase MOR strategy for each OS consisting of preliminary, intermediate, and terminal rendezvous phases.
In 2003, OS preliminary rendezvous would begin with MAV liftoff. The 2003 MSR lander would be rated to function on Mars for 90 days, so its MAV would need to launch from Mars within 90 days of touchdown. The 2003 OS would thus reach Mars orbit no later than April 2004. To save money and ensure adequate development time, the JPL MSR mission would employ a simplified solid-propellant MAV with a spin-stabilized first stage and a second stage with only a simple guidance system.
In their paper, the JPL engineers noted that even small OS orbit dispersions could place significant rendezvous propulsion demands on the CNES orbiter. An OS dispersion of only 1° in inclination, for example, would require that the orbiter alter its velocity by an additional 60 meters per second to match orbits, which would require an additional 48 kilograms of propellants.
For their MOR calculations, they assumed that a MAV capable of reliably placing the OS into a circular orbit 600 kilometers above Mars (plus or minus 100 kilometers) and inclined 45° to the planet's equator (plus or minus 1°) could be developed. They assume that the OS would take the form of a 14-to-16-centimeter sphere covered with solar cells which would power a radio beacon. The OS power system would include no batteries, so the beacon would operate only when the cells were in sunlight.
Between July 24 and August 26, 2006, the CNES orbiter would arrive in 250-by-1400-kilometer Mars orbit inclined 45° to Mars's equator. Once there, it would activate its Radio Direction Finder (RDF) to begin a four-week hunt for the 2003 OS. The RDF, which would collect OS data for relay to controllers on Earth, would have a range of 3000 kilometers. The JPL engineers suggested that other spacecraft in Mars orbit (Europe's Mars Express, the U.S. Mars Surveyor 2001 orbiter, or a specialized U.S. navigation & communications orbiter proposed for launch in 2003) might augment data from the CNES orbiter's RDF.
On September 24, 2006, controllers on Earth would begin the intermediate rendezvous phase by commanding the CNES orbiter to perform the Nodal Phasing Initiation (NPI) maneuver, the first in a series of maneuvers over 19 weeks designed to nearly match orbits with the 2003 OS. Radio signal roundtrip travel time would gradually increase from 23 to 43 minutes over the 19 weeks as Mars and Earth moved apart in their Sun-centered orbits.
At the start of this phase, both OS and orbiter would be in orbits inclined about 45° to Mars's equator; however, their orbits would have different ascending and descending nodes (that is, they would cross the equator at different places) and thus different orbital planes. In the planned 2003 OS orbit, the nodes would shift along the equator at the rate of 6.09° per day. This shifting, called regression of the nodes, would occur because of irregularities in the martian gravity field. The NPI would adjust the CNES orbiter's orbit so that its nodes would shift at a slightly different rate, enabling it to gradually match nodes with the 2003 OS.
Between October 8 and November 5, 2006, Mars would be behind the Sun as viewed from Earth and largely out of radio contact. No maneuvers would occur during this solar conjunction period, though nodal phasing would continue.
The Nodal Phasing Termination maneuver on January 7, 2007, would see the 2003 OS and CNES orbiter in nearly the same orbital plane. At the end of the intermediate rendezvous phase (February 4, 2007), the orbiter would be two kilometers below and 400 kilometers behind the OS. In its slightly lower (thus slightly faster) orbit, the orbiter would close with the OS at a rate of 200 kilometers per day (about 8.3 kilometers per hour).
In their paper, the CSDL engineers proposed a "double coelliptic" rendezvous strategy for the week-long terminal rendezvous phase. The CNES orbiter would fire its rocket motor about two days before planned OS capture to place itself in an orbit only 0.2 kilometers lower than that of the OS. This would slows the closing rate to about 20 kilometers per day (about 0.8 kilometers per hour).
The orbiter would acquire the OS with its twin Light Detection and Ranging (LIDAR) lasers as it closed to within five kilometers. At a distance of 0.4 kilometers, the orbiter would perform several maneuvers to intersect the OS's orbit 80 meters ahead of the OS. As it crossed the OS's path, it would fire its motor again to precisely match orbits.
The orbiter would then keep station with the OS for four hours. During this period, controllers on the ground would check the orbiter's systems. If everything checked out as normal, they would then give it the go-ahead to perform OS capture. If all went as planned, the CNES orbiter would automatically capture the 2003 OS on February 11, 2007.
The 2005 OS preliminary rendezvous would overlap the 2003 OS intermediate rendezvous. For purposes of their study, the JPL engineers assumed that the 2005 MAV would deliver its OS to Mars orbit on October 8, 2006, the last possible day before the start of solar conjunction. The 2005 OS would be targeted to an orbit matching as closely as possible that planned for the CNES orbiter at the time it captured the 2003 OS.
Intermediate rendezvous in 2005 would begin immediately after 2003 OS capture (that is, at the end of the 2003 OS Terminal Rendezvous phase) on February 11, 2007. Nodal phasing would end after 13 weeks, on May 13, 2007, and the 2005 OS Intermediate Rendezvous phase would end on June 10, 2007.
The 2005 OS terminal rendezvous would resemble its 2003 counterpart. The CNES orbiter would capture the 2005 OS on June 17, 2007, then would begin a series of maneuvers over four weeks to place itself into the proper orbital plane for departure for Earth on July 21, 2007.
The JPL engineers calculated that each 10-meter-per-second velocity change during intermediate rendezvous would require about 8 additional kilograms of orbiter propellant and subsystem mass at launch from Earth, and that the orbiter would need to make velocity changes totaling 478 meters per second during intermediate rendezvous if it were to have a 99% probability of successfully capturing both the 2003 and 2005 OSs. This would imply a rendezvous propellant mass of 382.4 kilograms. They noted that the MSR Project required only a 99% probability of retrieving one OS, and that this level of reliability could be achieved with an orbiter capable of velocity changes totaling 349 meters per second (which implied a propellant mass of 279.2 kilograms).
The CSDL engineers added that a 99% probability of successfully retrieving one OS meant a 60% probability of retrieving both. They calculated that terminal rendezvous using the propellant-saving double coelliptic rendezvous strategy would require velocity changes totaling only a little more than one meter per second up to the 80-meter stationkeeping point, and no more than 4.6 meters per second from the 80-meter point up to capture.