Table 1. Investigations to be performed on Cluster
Cluster's payload is an advanced set of experiments to measure electric and magnetic fields, plasmas and energetic particles. They have been designed primarily to study the small-scale structure of the Earth's magnetosphere. In addition to the very sophisticated measurement techniques that have been incorporated, the engineering challenges of manufacturing the state-of-the-art instruments and accommodating them successfully on board the spacecraft have been immense. The unique opportunity to build,integrate and test four identical sets of eleven instruments has also created problems not only of a technical, but also of a management nature, with a complexity of interfaces and logistics never previously experienced in an ESA scientific programme.
Table 1. Investigations to be performed on Cluster
Accurate measurement of the cold plasma population of the magnetospere demands that the electrostatic potential of the spacecraft be very low with respect to the ambient plasma. Cluster is the first mission that will be equipped with an ion emitter to routinely control this potential (ASPOC).
Cluster is equipped with a suite of five instruments - EFW, STAFF, WHISPER, WBD and DWP - to measure the electric fields and waves around each spacecraft. They have been combined together to make up the so-called 'Wave Experiment Consortium', which not only saves resources by using the same elements - sensors, data-handling unit and power supplies - but also enhances the scientific output.
The EFW instrument has been specifically designed for the investigation of fast time- and space-varying electric fields, but it will also cover the DC to low-frequency range, by using two pairs of thin, multiconductor wire booms per spacecraft. These booms deploy to a tip-to-tip distance of 100 m and on the end of each is a spherical sensor.
STAFF is a three-axis search-coil magnetometer which will measure the magnetic components of electromagnetic fluctuations in a frequency range up to 4 kHz. To put it beyond the influence of the spacecraft, the STAFF instrument is mounted on a 5 m rigid radial boom which will be deployed in orbit.
Figure 1. Top view of a Cluster spacecraft, showing experiment accommodation around the periphery and on the stowed radial booms
The relaxation sounder WHISPER is an intermittent transmitter/receiver instrument that can also be operated in a passive (receive-only) mode. The transmitter emits a short pulse to stimulate plasma resonances, thereby allowing the measurement of plasma density under widely varying plasma conditions. For both its transmit and receive modes, it will use the long wire booms of the EFW instrument.
The objective of the Wide-Band Data receiver system (WBD) is to provide high-resolution electric-field waveforms in order to analyse the highly structured and complex waves that occur in the Earth's magnetosphere.
The Digital Wave Processing (DWP) system contains a wave/particle correlator to correlate the time series of counts in the electron and ion analysers or the cross-correlation function between particle counts and wave measurements.
The Flux Gate Magnetometer (FGM) will perform important investigations by exploiting Cluster's four-point measurement capabilities in space. The resulting abilities to infer from true measurements in space the current density vectors, wave propagation characteristics and field discontinuities will represent 'firsts' in space plasma physics. The DC Magnetometer consists of two sensors mounted on a similar symmetric boom to that of STAFF, at distances of 3.5 m and 5 m from the spacecraft's body.
Cluster also carries another instrument to measure electric fields, namely EDI. The EDI instrument is based on the emission and subsequent detection of tracer electrons to derive the ambient electric field. It is a highly complex instrument, final calibration of which will only be possible in the true magnetospheric environment of space.
The plasma package, i.e. CIS and PEACE, is designed for the three-dimensional measurement of the distribution functions of both electrons and major ion species. CIS, the plasma ion spectrometer, employs two sensors to obtain the full three-dimensional ion distribution of the major species with high time resolution, and the mass-per-charge plasma composition. PEACE will perform the plasma electron measurements using two separate sensors covering the very cold electrons (LEAA) and the medium and higher energies (HEAA), respectively. PEACE will provide the three-dimensional electron distribution with high time resolution.
The final instrument aboard the spacecraft is RAPID, which consists of two spectrometers each containing position-sensitive solid-state detectors providing high angular and temporal resolution.
Installation of the payload has involved accommodating a total of 24 units on each of the four spacecraft. More than half of these units require unobstructed fields of view in space, something that has been difficult to accommodate given the presence of the many other experiments and the deployable rigid and the ultra-long wire booms Table 2.
Table 2. Major payload requirements on the Cluster spacecraft Number of units per spacecraft ~24 Total mass of payload <=72 Kg Total power available for payload (end-of-file) <=47 W Residual magnetism of spacecraft at magnetic sensors <=0.25 nT Stability of residual magnetic field <=0.1 nT Spacecraft skin potential: point-to-point <=1 V Accomodation of 100 m tip-to-tip wire booms High EMC cleanliness High cleanliness requirements for AIV programme and thruster positions
In addition, the detectors are highly sensitive to contamination, which could have occurred on the ground during the very long integration and test programme or could come from the thrusters and main engine of the spacecraft itself after launch. The first problem was solved by imposing a rigorous cleanliness programme during the integration and test phases, and only integrating the flight sensors at the last possible moment. In designing the spacecraft, the thrusters and main engine have been kept as far away as possible from the sensitive instruments, which are mounted around Cluster's upper periphery. The thrusters and main engine are thus on the opposite side of the spacecraft.
At the beginning of the programme, the payload was allocated 72 kg of mass and 47 W of power in the system budgets. The optimisation efforts to keep within these allocations have been extremely demanding. Cluster's various operating modes and the areas of operation within the elliptic orbit mean that the payload has a requirement for data rates varying between 17 and 220 kbit/s, whilst the overall volume of data to be handled per orbit can be as high as 2 Gbytes.
One of the major technical requirements on both the payload and the spacecraft has been to maintain electromagnetic cleanliness (EMC), as electric-wave and magnetic-field measurements are both adversely affected by spacecraft residual values. This resulted in a unique EMC programme for Cluster, and the mounting of the AC and DC magnetometers (STAFF and FGM) on rigid booms.
The planned detection of cold electrons by PEACE has called for a very careful design to eliminate the possibility of spurious effects being introduced by photo-electrons, which it is known will be abundant near the spacecraft's skin. The elimination of voltage potential variations over the surface has meant that the outer skin of the spacecraft must be as conductive as possible.
The major error sources had to be quantified for each instrument and major efforts were required to reduce such errors by careful design, stringent control of materials and manufacturing processes, and extensive pre-flight (and later in-flight) calibration of the individual instruments, as well as careful intercomparison of different measurements from the different instruments.
For the particle detectors - CIS, PEACE and RAPID - proper sensor calibration is especially important. It is also known that some of the long-term drifts in such instruments, mainly involving the micro-channel plates are difficult to correct entirely.
PEACE is a good example of how differences between the instruments on the four spacecraft were assessed during the development programme. It is necessary to maintain comparability between the analysers on the four flight spacecraft to within 1%, including the in-flight intercalibration.
The PEACE analyser consists of two concentric hemispherical shells between which the electrons are deflected. In order to maintain the 1% goal, the concentricity of the hemispheres had to be maintained to better than 40 microns. When tested, the initial design (Fig. 2a) was found to have a value of 150 microns. Even by improving the manufacturing tolerances to state-of-the-art values using specialist facilities, it could only be reduced to 80 microns. A major redesign was therefore undertaken both to increase the rigidity of the hemispheres and to connect them by a shorter load path using materials with similar coefficients of expansion (thereby reducing the effects of minor temperature differences between detectors). By these means, and by using a Kapton instead of a ceramic anode and changing the mechanical design (Fig. 2b), the tolerances could be reduced to 37 microns even using the normal flight-standard manufacturing tolerances.
Figure 2a. The original mounting of the PEACE hemispheres
Figure 2b. The final mounting of the PEACE hemispheres
For the wave experiments, the question of similarity on the four spacecraft was not a central issue. The stability of the oscillators is far better than the required frequency resolution, which is mainly dictated by telemetry constraints. In order to ensure correspondence between measurements on the four spacecraft, the relative timing between the measurements on each spacecraft is very important, which has meant that operational constraints have had to be rigorously defined.
FGM requires an overall single instrument accuracy of 0.1%. This has been demonstrated to be possible, and can also be checked in orbit using a four-point intercalibration technique (Div B must always be zero).
Payload resources were tight from the outset, with the baseline 72 kg/47 W/17 kbit/s being oversubscribed already at the proposal stage. A de-scoping and rationalisation exercise had to take place immediately, which resulted in the formation of the Wave Consortium and the provision of the power-supply and data-handling functions via boxes common to all five of those experiments.
During the development phase, the payload mass again began to grow and every additional gramme had to be accounted for. Very thin walled structures were developed for some electronic boxes (e.g. FGM) and some instruments switched to using magnesium as their primary structural material (e.g. WBD; Fig. 3).
Figure 3. The WBD experiment's outer structure, weight-optimised by the extensive use of magnesium
Power for driving the sophisticated instruments and their computers was also critical from an early stage. Various operating modes were therefore developed for each instrument to conserve power by, for example, running the processors at different speeds during the two-year mission. The data rates have also been optimised by defining different operational modes and time-sharing between the individual instruments.
The payloads to be launched on the flight spacecraft are still (just!) within their allocated resources, thanks to the continual attention that has been paid to the problem by both the Experimenters and the ESA Project Team.
Before making any scientific measurements, Cluster's instruments will have to survive the launch-induced vibrations. Once in orbit, they must work in an environment varying between cold deep space and hot sunlight, and also endure a constant shower of strong radiation.
The Cluster launch vibration environment has been assumed to be slightly harsher than usual, to take into account the inherent uncertainties due to it being the first Ariane-5 launch. This has had design repercussions for some instrument units. The CIS experiment, for example, contains very thin carbon foils (density of order 3 g/cm 2 ). It was very soon realised that these would have problems surviving the vibration environment and so the experiment was mounted on rubber anti-vibration mounts. During testing, however, these foils still broke due to the high induced displacements causing the foil to hit the surrounding structure. A design modification and further testing has ensured that CIS will survive the launch (Fig. 4).
Figure 4. CIS 1 sensor integrated on its anti-vibration mounts
The in-orbit thermal environment, varying from hot sunlight to long cold eclipses with a minimum of heater power provided by the spacecraft, has presented a severe challenge for all protruding sensors. These need an environment close to room temperature for their sensitive components like the micro-channel plates when operating, and not colder than a freezer when non-operational.
A few sensors required further design optimisation after the spacecraft thermal-balance test. In particular, their coatings and thermal blanket interfaces had to be changed. Figure 5 shows the complex way in which the thermal blankets are now 'hand-tailored' around the protruding instruments.
Figure 5.Tailoring of the thermal blankets around the protruding experiments
The harsh in-orbit radiation environment of typically 20 krad that Cluster must endure has required the use of radiation-hardened or shielded components. Special circuitry to survive the latch-up that could be caused by particle intrusion has also been necessary, especially for the CMOS technologies.
The instruments aboard each new generation of scientific satellite endeavour to exploit the latest sensor and electronic technology to achieve the highest possible sensitivity and resolution. For a plasma mission like Cluster, this ideally means that the spacecraft itself should be 'invisible', with no apparent interaction with the space environment.
Reality is rather different; the spacecraft does become charged under the cyclic influence of sunlight and shadow. In addition, certain of the spacecraft's subsystems contain relays and valves, which generate DC or slowly-varying magnetic fields. It has an electrical support system for power and data-handling, which also constitutes a source of electromagnetic interference. Last but not least, the scientific payload itself may generate signals that could influence the performance of other instruments on board. Most of these phenomena are either impossible to simulate, or the levels involved are orders of magnitude too small to be measured on the ground.
How then can a suitable spacecraft and its payload be built and tested to the satisfaction of the scientific community? Well, the process begins with a careful selection of design concepts and materials, which are then translated into an electromagnetic design and test specification. The necessary electrostatic cleanliness is achieved by using a conductive coating on the spacecraft's external surfaces, including the solar arrays. Thermal insulation blankets and foils are coated with indium tin oxide and, along with all other external parts, locally grounded to the spacecraft structure to avoid the build-up of electrostatic potential. DC magnetic cleanliness imposes the selection of non-magnetic materials wherever possible. In addition, the magnetic sensors are mounted on deployable booms, as far as possible from the spacecraft body.
Electromagnetic interference is usually controlled by the synchronisation of clock signals onboard the spacecraft. This, together with an optimised electrical harness configuration and grounding scheme, ensures minimum disturbance of the frequency bands being observed by the scientific instruments.
However, when these design optimisations came to be translated into hardware for Cluster, some of the above concepts could not be realised: either the ideal technology could not support the original purpose of a certain element or instrument, or it was not available within the given constraints of mass, power or schedule. An EMC Review Board was therefore established, made up of scientists from each particular area of Cluster plasma science, from Industry and from ESA, to analyse the problems and to agree on acceptable solutions.
The verification phase involved various EMC analyses and tests not only state-of-the-art conductivity and susceptibility tests, but also dedicated experiment tests for the particularly sensitive WEC and PEACE instruments. After several iterations and optimisation of the grounding schemes of these experiments and the spacecraft interface, these two experiment teams were able to confirm a satifactorily low noise environment on the spacecraft. Radiated emission and susceptibility testing at IABG in Munich (Fig. 6) with the EFW wire booms partially deployed finally confirmed that the Cluster spacecraft generates levels very close to the instruments own background noise.
Figure 6. Radiated emission and susceptibility testing of Cluster at IABG in Munich (D)
In summary, the way to EMC cleanliness was not straightforward, but the close cooperation of all parties involved in addressing the problem has resulted in Cluster being the most electromagnetically clean spacecraft that ESA has launched.
A feature of modern-day instrumentation and the result of the great advances in micro-technology is the fact that each experiment has almost as much onboard computational capability as the spacecraft's own on-board computer (CDMU). In fact, the Cluster payload has the processing capability of almost seven personal computers (PCs), the Wave Consortium being served by one processing unit (DWP).
This onboard computational power not only provides the interfaces with the spacecraft bus in terms of telemetry and telecommand handling, but also the interface with the various experiment units. It also performs a certain amount of onboard scientific processing, which reduces the data rates needed to the ground.
The DWP (Fig. 7), perhaps the most performant computer, consists of three T222 transputers, each with 32 kbytes of external RAM (internal memory disabled to provide increased radiation tolerance) and 32 kbytes of PROM. The DWP's design permits the transputers to be operated at input clock frequencies of 2.5 or 5 MHz, the slower rate requiring less power.
Figure 7. The three printed-circuit boards of the DWP, including the three T222 transputers
The DWP performs two major software-driven tasks. The first is particle correlation, with a novel diagnostic technique based on forming auto-correlation functions of the time series of particle-detector counts as a function of energy and pitch angle. Secondly, it performs data compression in order to cope optimally with the restrictions on the available telemetry bandwidth. Various data-compression methods are implemented within the DWP to remove any redundant information from the data stream.
Management of the Cluster experiments has provided some unique challenges for the spacecraft payload groups. To begin with, getting together the best payload suite in Europe to satisfy the mission requirements meant the involvement of 11 Principal Investigators (PIs) and interfacing with the 250 main investigators who will study the Cluster data. This is not a unique scenario in terms of ESA missions, but it did mean that rationalisation of the instrument groups was required early in the programme. One result of this was the setting up of the Wave Consortium, mentioned earlier.
The next major problem was that, because the integration, testing and preparation for launch of the four flight spacecraft would take over two years, engineering breadboards were required very early in the programme compared to the launch date. The experimenter groups performed this difficult task in a very satisfactory way for the hardware, but the software for ground testing and on-board processing always lagged behind.
For a nominal one-off spacecraft, once the flight hardware has been built and tested at the home institute, the payload team is available to assist the Project with the system-level Assembly, Integration and Verification (AIV) programme. In Cluster's case, once the first flight instruments had been delivered another four (including flight spare) were needed. In addition, system-level testing had to be performed simultaneously on up to three flight spacecraft (e.g. functional testing, EMC, and thermal vacuum), all of which required payload support. The short-term schedule for such system testing was also varying on a day-to-day basis. The logistical problems for the payload groups have therefore been immense and a successful outcome hinged on the trust built up between the Project and the payload groups in the early days of the programme, and on the flexibility shown by both sides. This has been supplemented by technological advances such as the remote links established to both the integration and test sites (Dornier and IABG), which gave the payload teams remote access to their test data from their home institutes.
The engineering challenges faced with the Cluster payload have been some of the most demanding of any spacecraft hitherto launched. They have been faced by the instrument group and the Cluster Project together, in order not only to launch the most sophisticated instrument set possible but also to launch four identical sets. Their combined efforts over the past eight years will ensure that ESA's 'space fleet to the magnetosphere' will provide the scientific community with a wealth of unique data on that region of space around the Earth known as the magnetosphere and its interaction with the constant stream of particles known as the solar wind, emitted by our mother star, the Sun.
ESA Bulletin Nr. 84.