Hot Plasma Composition Analyzer for the Magnetospheric Multiscale Mission

6.1.1 Ion Optics

We will describe the optics in the sense that particles fly through the ESA and TOFA to the detector (Fig. 5). Relative locations will be referenced as though the instrument was sitting vertically. Thus the collimator is “above” the ESA, which is above the TOFA, etc. We defer a description of the RF subsystem to Sect. 6.4 where the problem of dynamic range is addressed.

Figures 9, 10 and 11 show trajectories of 1 keV ions travelling through the optical system in three orthogonal planes (Fig. 9 is in the same plane as Fig. 5). Figure 9 shows pairs of ions entering the collimator over a range of azimuthal angles. They are focused by the ESA on to carbon foils located at the entrance to the TOFA. Trajectories c and f, d and g, etc. in Fig. 10 illustrate how a spread in elevation trajectories is focused in the plane orthogonal to Fig. 9. Ray tracing in Fig. 11 shows the same trajectories seen from “above,” demonstrating the full 3-dimensional aspects of focusing.

Open image in new window

Open image in new window

Open image in new window

Ions enter HPCA via a collimator that consists of two parallel, electrically grounded disks held together by eight posts whose cross sections are designed to prevent scattering of incoming ions in elevation. The collimator has edges that trim the azimuthal FOV and limit trajectories entering with energies outside the ESA passband. A high transmission grid is mounted slightly inboard of the collimator to prevent RF emissions escaping that might cause electrical interference with the spacecraft.

A central disk in the collimator protrudes slightly downwards (Figs. 9 and 10) to shape the tophat electric field. Simulations were used to optimize the diameter and height of the disk in order to obtain maximum transmission while maintaining the central plane of the FOV parallel to the collimator plane and the spacecraft surface. A test of the efficacy of the collimator geometry and its relationship to the ESA and TOF optics is demonstrated in Fig. 12. Here the ions reaching the MCP detector via the collimator and ESA and within the normal energy-angle passband were flown “backwards” through the ESA to the collimator entrance. In Fig. 12 the backwards travelling beam is shown mapped back to the collimator entrance. What the figure demonstrates is that the transmitted beam, which fills the angle-energy passband, also fills the collimator. This indicates nearly ideal coupling between the collimator and ESA optics.

Open image in new window

The ESA is comprised of two concentric mildly toroidal shells. The grounded outer shell supports the collimator assembly while the inner operates at a negative voltage proportional to ion energy/charge. The inner shell is divided into two parts: the upper carries only the ESA DC voltage while the lower carries combined DC and RF voltages. This assembly, including the RF distribution network (Fig. 4), is suspended above the TOFA by a thin conical insulator made of the low-outgassing polymer Ultem 1000. In order to suppress any possible electromagnetic interference, the local RF distribution network resides within the inner ESA shell. High voltage is delivered to the inner shell via hollow spokes that support the inner shell assembly. Posts supporting the collimator, and spokes supporting the inner ESA, were designed to minimize blockage of the elevation pixels. That feature, together with elevation focusing (Fig. 10), resulted in the full theoretical passband of 22.5° being maintained.

Toroidal optics possess two radii of curvature (R ₀ and R ₀+R ₁ in Fig. 13) that focus ions independently in two orthogonal planes. The radii can be adjusted to obtain optimum focusing between the ESA and TOFA (Figs. 9 and 10), which yields some improvement in geometric factor over a spherical analyzer (Young et al. 1988; Gomez 2011). Since the ESA is only mildly toroidal its transmission properties can be estimated analytically to first order. This allowed us to optimize the geometric factor vs. energy-angle resolution while also matching the conditions needed for maximum transmission through the TOFA.

Open image in new window

These first order equations show the dependence of sensitivity and resolution on sensor geometry and suggest design tradeoffs. In particular, sensitivity is proportional to \(R_{0}^{2}\), which quickly drives up instrument size for a given resolution (k=constant). For a given instrument size (fixed R ₀) resolution drives sensitivity even faster: doubling resolution decreases sensitivity by nearly an order of magnitude. Increasing resolution (smaller k) has the advantage of reducing the amount of high voltage on the inner ESA shell for a given particle energy (15).

6.1.2 ESA Electronics

Electronics associated with the ESA consist primarily of a DC high voltage (HV) stepping supply and a novel RF HV supply (Fig. 14). The supplies are described in detail in Sect. 7.2.3. The ESA supply generates a commandable negative voltage −V _j that is applied to the upper and lower sections of the inner ESA toroid to select ions with energy E _j =kV _j . The highest applied voltage on the ESA is −7000 V across the 4.0 mm shell gap giving a maximum electric field of 1.75 kV/mm, which is well within engineering design guidelines limiting electric fields in vacuum to ≤2 kV/mm. The minimum supply voltage is −0.21 V.

Open image in new window

6.2.1 Ion Optics

The TOFA consists of a cylindrical volume topped by 16 equally-spaced carbon foils biased at −15 kV (Fig. 5). Ions exiting the ESA are accelerated by −15 kV across a gap between the ESA and TOFA into carbon foils (∼1 μg/cm²) mounted on 90 % transmissive 333 lines-per-inch grids. The post-acceleration of −15 kV ensures that all ion species, including those with external energies as low as a few eV, are able to penetrate the foils. Ions exit the foils either positively or negatively charged or as neutrals. (In what follows we continue to refer to the particles as “ions.”) The TOFA optics are designed so that charge state doesn’t appreciably affect trajectories inside the analyzer.

Ions exiting the carbon foils eject secondary electrons that are focused in three dimensions (Fig. 15) to form an image of the foil on the outer edge of the top MCP (Fig. 16). The MCP has a 79 mm active diameter and is held at −13.6 kV by a resistor divider network (Fig. 14). A tap off of the −15 kV supply was designed to maintain a nominal bias across the MCP of 900 V. It is important to get the bias voltage on the top MCP correct because once installed the divider resistor that controls the voltage (Fig. 14) cannot be changed. However during tests we found that in going from the maximum allowable instrument operating temperature of 25 C, to the lowest allowable temperature of −25 C, the resistance of the top MCP increased by as much as 35 %. This was large enough to cause arcing across the MCP. Aside from arcing, an increase of this much would also increase MCP gain to unacceptable levels as well as altering the electron deflector voltage by 270 V thereby changing the TOF internal optics. One solution to maintain the correct bias would be to restrict the temperature range (and the instrument operating range) over which the TOF voltage could operate. A better solution, implemented by the MMS project, added heaters to the HPCA to dynamically maintain the HV supply operating temperature at ∼10 C.

Open image in new window

Open image in new window

The foil image on the top MCP is transmitted to the bottom MCP by focused secondary electrons (Figs. 5, 9, 16). Electron focusing in the TOFA, and from the top to the bottom MCP, is critical because the sharpness of elevation passbands depends on preventing electrons emitted by one foil from crossing over onto the adjacent image. Three-dimensional focusing is achieved by carefully shaping the back of the foil holders and by a cylindrical electrode held at −13.8 kV (Fig. 14). Figure 15 shows numerical simulations of electron trajectories leaving the foils and travelling through the flight region to the top MCP.

The bottom MCP is mounted 16.6 mm below the top MCP. The electric field between the two is designed to accelerate and tightly focus electrons on to the bottom MCP. The combined MCPs have a gain of ∼10⁷ at nominal operating voltages of ∼800 V across each plate. Separating the two MCPs in this way solves the problem of decoupling the signal from the top MCP at −13.6 kV to the low-voltage signal electronics associated with the bottom MCP without using large capacitors (Young et al. 2004, 2007). Prior to installation the MCPs are burned in with a UV source until ∼0.1 C of charge is extracted and the gain is stable with respect to the amount of charge extracted.

The cable carrying −15 kV from the HV supply to the TOFA is routed via a 30 kV-rated HV capacitor and HV distribution network located in a small volume inside the TOFA housing (Fig. 14). This area was particularly susceptible to HV breakdown so considerable effort was put into designing the network and surrounding region using field-tracing software as well as a large amount of testing in the flight configuration.

High voltage and signal cables pass through a sealed bulkhead separating the electronics compartment from the sensor. The bulkhead is designed to prevent sensor contamination by outgassing of the electronics. Chemical cleanliness is further ensured by a purge tube that runs along the exterior of the instrument (Fig. 6) and into the sensor at the location of the TOF analyzer. The purge rate is 0.5 to 1.0 liters/minute of high purity N ₂ with the red-tag cover on and 1.0 to 2.0 liters/minute with it off. Purge continues until liftoff.

Below the bottom MCP is the combined anode and front-end signal electronics (FEE) (Fig. 4). The electron cloud from a single event leaves the bottom MCP and is collected on dual delay-line position-encoding anodes (Paschalidis et al. 2008, Byrum et al. 2010). Figure 17 is a schematic of the anodes, delay lines and TOF electronics. The photographs in Fig. 18 show details of the anodes and delay lines. Delay line electronics are discussed in more detail in Sect. 6.2.2.

Open image in new window

Open image in new window

Charge arriving on the start anode splits into two pulses (Start_1 and Start_2) travelling in opposite directions around the start delay line. (Note that we use the convention that an underscore such as ‘Start_1’ indicates a signal. Those without an underscore such as ‘Delay 1’ indicate a circuit component.) The pulses arrive at two amplifiers separated by a time interval proportional to the position of the incident charge on the anode and thus to the ion’s elevation angle. Inductive and capacitive components that couple the delay line elements code for elevation with a resolution of 32 positions although only 16 are reported in telemetry.

Similarly, the charge cloud hitting the stop anode splits into Stop_1 and Stop_2 pulses that travel along the radial delay line made up of concentric electrodes (Fig. 18) that code for 16 radial positions of which 8 are reported. Time separation of the two signals encodes the ion’s radial position. In addition to position information, the average difference between Start and Stop signals gives the ion TOF from which mass is calculated. In what follows we describe in detail the position and TOF measurements.

The transit time T of an ion along a path length L in the TOFA (Fig. 19) is given by

$$\begin{aligned} T = L \bigl( M /2 E ^{*}\bigr)^{1/2}. \end{aligned}$$

(17)

In engineering units (dimensions are in square brackets)

$$\begin{aligned} T\ [\mathrm{ns}] = 22.85 L\ [\mathrm{cm}]\ \bigl(M\ [\mathrm{amu}]/E^{*}\ [\mathrm{keV}]\bigr)^{1/2} \end{aligned}$$

(18)

where E ^∗ is the total ion energy in the TOFA including a correction for energy lost in the foil (ΔE _foil). The total ion energy is

$$\begin{aligned} E^{*} = E_{0} + qV_{\mathrm{acc}} - \Delta E_{\mathrm{foil}}\ \mbox{[keV]} \end{aligned}$$

(19)

Open image in new window

In considering the TOFA design it was important to bound ion flight times since they set requirements on both analyzer geometry and on high-speed TOF electronics. The fastest ion passing through the TOFA is H⁺ travelling at maximum velocity (corresponding to E ₀=40 keV) along the shortest path H (for the purposes of initial estimation H≈2.5 cm). Neglecting the small amount of energy lost in the foil, the shortest H⁺ flight time is 7.7 ns which leads to an acceptable lower limit of 5.0 ns.

The slowest ion through the flight region is O⁺ at E ₀∼1 eV incident on the ESA. Here another operational constraint comes into play. In case there are problems operating at the highest acceleration voltage of −15 kV we want the TOF to be able to function at voltages as low as V _acc=−12 kV. At that voltage we would have degraded but acceptable mass resolution. At −12 kV O⁺ would lose about 5 keV in the foil (Allegrini et al. 2006) so that E ^∗≈7 keV. Then T _max≈138 ns giving L _max≈4.0 cm. It is relatively easy for the TOFA electronics to measure longer times-of-flight so the upper bound was set at 256 ns. This leads to a maximum allowable upper limit on path length across the TOFA of 7.4 cm. In summary, ion times-of-flight between 5 and 256 ns provide acceptable timing limits for the electronics and dimensions for the optical geometry (viz., 2.5 to 7.4 cm from foils to MCP).

One important point about plasma mass spectrometers flown in the Earth’s magnetosphere is that attaining high mass resolution per se is not important. Since the target ion species are well known (H⁺, He⁺⁺, He⁺, and O⁺) only enough resolution is needed to identify these ions. This simplifies the mass spectrometer design and reduces resources significantly compared to a high resolution device designed solely for mass spectrometry.

There is no practical means of reducing the contribution of energy spread to peak broadening. However it is possible to correct for some of the path length differences if the ion position on the MCP is known. Knowing the radial position allows a corrected path length to be calculated. However for this to work the position of each ion must be determined individually and “on the fly” as each event occurs—statistical measurements, which are much easier to make, will not suffice.

6.2.2 TOF Measurement and Position Encoding

With reference to Fig. 17 (Paschalidis et al. 2010), electrons leaving the foils following ion impact produce two signals (Start_1 and Start_2). Some nanoseconds later ions strike the MCP producing Stop_1 and Stop_2 signals. These are delayed 23 ns and 36 ns (Delay 1 and Delay 2 in Fig. 17) respectively in order to ensure that no timing ambiguities arise between starts and stops. Start and stop signals pass through constant fraction discriminators (CFD) to the three ASICS (CFD thresholds can be adjusted in flight). The TOF1 and TOF2 ASICS (Paschalidis et al. 2002) then measure the delays between opposite ends of the start and stop delay lines and pass this information to the FPGA (Fig. 17).

Ion TOF and position calculations then proceed as follows. The uncorrected TOF is measured by the TOF1 and TOF2 ASICs using the delay T ₁ between the time a start pulse reaches the Start_1 end of the anode, and the time the corresponding stop pulse reaches the center of the stop anode (Stop_1) plus Delay_1 (Fig. 17). Similarly T ₂ corresponds to the delay between the time the start pulse reaches the Start_2 end of the anode, and the time the stop pulse reaches the outer ring of the stop anode (Stop_2) plus Delay_2.

The value T ₀ is used to address one of 512 bins which is incremented for each event. The bins are bracketed into five energy-dependent channels, one for each species plus background. Figure 20 is a sample TOF calibration spectrum showing TOF brackets. The flight software allows the TOF correction feature to be turned off for calibration purposes. The bin locations can be moved to detect other species such as O⁺⁺.

Open image in new window

The correction process described above improves TOF peak resolution by 60 % at 1 keV. Figure 21 shows TOF peaks from stop rings 1 (the innermost) and 7 (the outermost) separated by 47 ns edge-to-edge at FWHM. If the counts from each ring (1 through 7) were added together without corrections the peak profile would be smeared out. The corrected peak is only 12 ns wide at FWHM demonstrating the efficacy of the correction process.

Open image in new window

At this point in the data flow ions have been binned according to TOF (512 channels) and elevation (16 angles). Additional counters record the number of raw start and stop events and the number of valid coincident events processed by the FPGA. Start and stop counters have dead times of <250 ns, which allows them to be used to correct for the slower 2 μs dead time associated with coincident TOF signal processing. Processed events are then transmitted to the C&DH system once each spectrum (625 ms). The next steps in data processing take place in the C&DH system discussed in Sect. 8.2.

6.4.1 Introduction

As discussed earlier, intense proton fluxes can potentially produce counting rates as high as 20 MHz at the nominal peak of the energy distribution (Fig. 2). This creates two problems related to dynamic range. The first is potential current saturation of the MCP resulting in reduced gain and lost signal. The second is the potential inability of the TOF processing electronics to keep up with high rates corresponding to coincident rate dead times ∼100 ns. At such high rates uncorrelated (“accidental”) start and stop events can tie up the processing capability of TOF electronics, resulting in high rates of cross-talk between adjacent TOF channels. In particular the high proton signal will completely drown out minor species such as He⁺⁺ and O⁺.

One solution might be to place two different sized apertures at locations around the entrance: Large apertures for low fluxes and small apertures (e.g. ∼1 % of the large) for intense fluxes. However the large apertures would still transmit the same flux per unit area to the detector, causing local saturation, while the smaller apertures would reduce the sensitivity of half (or more) of the instrument, making minor species detection more difficult.

The ideal solution is to reduce proton fluxes to manageable levels while maintaining minor species fluxes close to ambient levels. This approach requires placing what is, in effect, a low-resolution mass spectrometer in front of the primary TOF mass spectrometer in order to separate protons from minor species (He⁺⁺, He⁺ and O⁺). Such an arrangement could attenuate intense proton fluxes while transmitting heavier species. Burch et al. (2005) have developed what is essentially a low-resolution mass filter using a radiofrequency (RF) technique similar to the principle behind quadrupole mass spectrometers.

To get some idea of the requirements for the RF system, assume that the maximum total proton flux reaching the MCP is 2×10⁷ ions/s. If each ion produces on average 2 electrons from a foil then the number of particles striking the MCP is ∼6×10⁷ s⁻¹ (including the incoming ion). In order to have optimal signal amplitudes for the TOF electronics the HPCA MCP is operated at a gain of ∼10⁷. At this rate and gain the signal current exiting the bottom MCP is ∼(6×10⁷ particles/s ×10⁷ electrons/particle) ×1.6×10⁻¹⁹ C/particle ∼10⁻⁴ C/s=100 μA. In order to have a linear output, the MCP signal current should be limited to <10 % of strip current (MCP bias voltage divided by resistance). This amounts to an MCP resistance of ∼10 MΩ for the bottom MCP which is what led us to choose the Hammamatsu Type 1942-04 MCP whose resistance can be specified. For HPCA we chose ∼70 MΩ for the top MCP and 12 MΩ for the bottom.

A second issue is current density on the MCP. The tight focusing of electrons (about 0.25 cm² per pixel in Fig. 16) can lead to current densities ∼100 μA/cm² which the selected MCP is able to support over a few pixels. A third problem area is the TOF electronics processing rate. We are using the best available custom ASICs, designed and built by APL (Paschalidis et al. 2002). The dead time for processing valid TOF and position location events is ∼2 μs which puts a practical limit of ∼0.5 MHz for accurate measurement.

In summary, based on considerations of MCP current saturation and TOF processing speed the RF system needs to reduce proton fluxes by at least a factor of ten.

6.4.2 Attenuation Using Radio Frequency Selection

Our approach to limiting proton fluxes is to turn the ESA into what is effectively a low-resolution RF ion mass spectrometer (Burch et al. 2005) operating independently of the TOF mass spectrometer. The RF principle can be understood with reference to Fig. 22, which shows an idealized case of a sinusoidal RF electric field applied to parallel conducting electrodes placed at ±1.0 and parallel to the x-axis in the figure. In this simple example the RF amplitude and frequency are tuned so that on average the faster protons see only a single RF cycle, causing them to be deflected by a large amount. The heavier, slower O⁺ ions see many oscillations of the electric field which tend to cancel each other out resulting in small deflections and high transmission.

Open image in new window

In the HPCA, as in all ESAs, a DC voltage corresponding to the desired energy is applied to the inner dome (see Eq. (15)). A sinusoidal RF voltage of selectable amplitude and frequency is added to the DC voltage and applied to the lower part of the ESA (Fig. 5). Protons entering with a given energy move through the ESA in a length of time corresponding to about one-half an RF oscillation period. The protons experience a slowly varying field that deflects them to the side of the ESA. The number deflected, and hence the amount of attenuation, depends on the choice of RF amplitude and frequency.

Heavier ions such as O⁺ with the same energy as protons travel more slowly through the ESA, encountering multiple oscillations of the electric field which modify the trajectory slightly but tend to cancel out (Fig. 22), allowing ions to travel through the ESA with minimum deflection. In the ray-trace simulation shown in Fig. 5, H⁺ and O⁺ ions enter the ESA with the same energy. Protons (black trajectories) immediately hit the lower part of the ESA to which RF + DC voltage is applied. Oxygen ions (red trajectories) are transmitted without appreciable losses. Intermediate mass ions (He⁺⁺ and He⁺) are partially attenuated.

The highest proton fluxes found in the magnetosheath extend from approximately 0.5 to 4 keV (Fig. 2). Therefore RF attenuation is designed to operate over this range. Although the choice of frequencies is limited to 16 fixed steps, the amplitude can be set precisely by a 12-bit digital-to-analog converter (DAC). While the ESA uses 63 steps to cover the energy range 1 eV to 40 keV, the RF is applied to only 14 of those steps covering 0.5 keV to 4.0 keV.

Figure 23 shows beam data for several ion species at an energy of 0.995 keV corresponding to the nominal peak where attenuation is most critical. We emphasize that the fraction of flux attenuated has been shown by numerous tests and calibration to be highly repeatable. The RF frequency and amplitude combinations are given in Table 4. (Note that tests of the HPCA prototype used to collect data for Fig. 21 were made at a laboratory beam energy of 1.0 keV. Test and calibration of the Flight Models were carried out at fixed pre-programmed energy levels. The level nearest 1.0 keV is 0.995 keV hence the difference in ion energies between Figs. 21 and 23.)

Open image in new window

Table 4

RF peak-to-peak voltage and frequency

ESA Step No.	DC Voltage (V)	Ion Energy (eV)	RF Frequency (MHz)	RF Amplitude (Vpp)
36	−94.7	515.9	5.1	127
37	−111.7	608.5	5.1	127
38	−131.7	717.8	5.1	127
39	−155.4	846.7	5.1	127
40	−183.3	998.8	5.2	131
41	−216.2	1178.2	5.4	135
42	−255.0	1389.8	5.5	152
43	−300.8	1639.3	5.8	177
44	−354.8	1933.7	5.9	245
45	−418.5	2281.0	6.1	294
46	−493.7	2690.6	6.1	377
47	−582.3	3173.8	6.1	377
48	−686.9	3743.7	6.1	377
49	−810.3	4416.0	6.1	377

Figure 23 demonstrates proton attenuation by factors up to ∼330. The data also indicate that at these settings solar wind He⁺⁺, for which \(\mathrm{H}_{2}^{+}\) is a stand-in, is attenuated by a factor of 10 or less at ∼1 keV. The peak of the He⁺⁺ distribution is roughly four times higher where attenuation will be considerably reduced. In any case the loss in counting rate of He⁺⁺ is compensated by an improvement in signal-to-noise ratio (SNR). Finally, as expected from theory and ray-tracing, heavy ions such as N⁺ (a stand-in for O⁺) and \(\mathrm{N}_{2}^{+}\) are transmitted without any attenuation.

One important feature of the attenuation process calling for careful calibration is apparent in Fig. 24. The attenuated passbands (Figs. 24b and 24d) are shifted relative to the nominal passbands (Figs. 24a and 24c) but maintain their shape, i.e., resolution remains the same. Calibration data such as this allows the correct energy and flux of the attenuated ions to be recovered during analysis. The shift in the attenuated distributions can be attributed to ion trajectories just entering the ESA that are deflected by the RF field through angles and energies not ordinarily transmitted through the ESA in the DC mode (e.g., rays resembling f and c in Fig. 9).

Open image in new window

6.4.3 High Counting Rate Capability

The RF system will be operated continuously in Fast Survey mode over the pre-selected parts of the orbit where reconnection is judged likely to occur (see Sect. 8.1.2 for a full discussion of HPCA modes). On other parts of the orbit HPCA will be operated in Slow Survey mode. In this operational scenario there are three ways in which proton fluxes might exceed the planned maximum rates. The first is where the flux maximum is above model rates in Fig. 2. The second is when reconnection occurs outside the pre-planned regions—during a strong magnetic storm, for example, when the magnetosphere collapses and the instrument is not in an RF mode. The third is when large fluxes are encountered in Slow Survey mode. Although RF is by far the best way to increase dynamic range in planned scenarios, it is also important to have some back-up capability to detect and process events at as high rates as possible.

The design goals for the start and stop counters were 200 ns and 3.0 μs dead times respectively for correlated events in the TOF1, TOF2, and TOF3 ASICs. Figure 25 shows linear fits to FM1 calibration data for rate in vs. rate out at relatively low rates while Fig. 26 shows data taken over a wider range of rates. Dead times are the same in both plots: 100 ns for Start_1 and Start_2, 1 μs for TOF1, and 2 μs for TOF3. At low rates the TOF electronics are non-paralyzable, i.e., if a second event arrives while the first is being processed the second event is ignored. The measured count rate C _M as a function of the true rate C _T (measured by the Faraday cup), for a dead time τ is

$$\begin{aligned} C_{ M } = C_{ T } /(1 - \tau C_{ T }). \end{aligned}$$

(33)

Equation (33) was used to fit the linear portion of data plotted in Fig. 25 where dead time effects are not important. Figure 26 shows the entire data set including non-linear portions where dead time effects are important. Data were not taken at beam currents above 1.3 pA at which point the instrument would be completely saturated.

Open image in new window

Open image in new window

7.2.1 Command and Data Handling

HPCA’s command and data handling (C&DH) subsystem (Fig. 28) provides interfaces to the Central Instrument Data Processor (CIDP). The primary tasks of the HPCA C&DH, which is based on an ACTEL RTAX2000 FPGA, are to control science modes, acquire data from the TOF FPGA, compress and format that data, and transmit the result to the CIDP. Once a command is received ESA energy sweep cycles are executed by commanding HV power supply voltages step-by-step through a pre-loaded 64-step table. Nominally the steps are spaced 625 ms apart. However, in order to deal with possible changes in spin rate, the table also controls the duration of each energy step. The second major task is collection of raw TOF data once every energy step in synchronization with HV sweeps. Data are stored in two ping-pong memories (Fig. 28) and then decimated as required using a process discussed in detail in Sect. 8.3.

Open image in new window

Partially decimated data are transmitted to an Atmel SPARC-8 micro-controller that further decimates and compresses science data before passing it back to the C&DH FPGA. The amount of decimation depends on the data mode and is controlled by command. During calibration raw data rates are ∼27 Mbps. In Fast Survey Burst Mode this is compressed to 180 kbps (10³ bits/s) and in Fast Survey the data are further decimated to 5.6 kbps. In Slow Survey data are heavily decimated to 0.8 kbps. Compression modes and their application are discussed in more detail in Sect. 8.3. After formatting the data to CCSDS (Consultative Committee for Space Data Systems) standards, the C&DH transmits it to the CIDP. Figure 29 shows a schematic diagram of the raw data products leaving the HPCA headed for the CIDP or, during ground tests, the instrument EGSE (Electrical Ground Support Equipment).

Open image in new window

7.2.2 Time-of-Flight

Much of the TOFA functionality was discussed in Sect. 6.2.2. This section presents details about implementation of the electronics (refer to the block diagram in Fig. 17).

Twenty-four discrete amplifiers and discriminators would have been needed for conventional position encoding methods, which would run up against both volume and power limitations. Our delay line solution discussed earlier requires only 4 low-power amplifiers and discriminators. The former are mounted on the anode board while the latter are located in the TOF ASICs (Figures 17 and 18).

Aliveness and functionality tests of the TOF board are carried out with four built-in pulsers capable of stimulating the anodes and signal chain at rates between 24 kHz and 6 MHz. Pulse amplitudes can be varied as can delays between pulses at intervals of 40 ns to produce an artificial TOF spectrum. Stimulation of all start and stop positions constitutes a complete test of the integrated TOF system.

In addition to science data in the form of ion TOF and elevation and radial positions, the TOF FPGA transmits the number of single events (Start_1, Start_2, Stop_1, Stop_2), valid events for each of the three TOF chips, and the number of times the FPGA state machine was initiated. “Ground truth” data used to check TOF processing is provided by recording the last 1024 valid “direct events” per sample. Direct-event data consist of TOF1, TOF2 and TOF3 values for individual events. These can be checked against the processed position and TOF measurements.

7.2.3 Power System

Figure 30 is a block diagram of the entire power system. Figure 14 is a diagram of the HV distribution network inside the sensor. The power system is comprised of five circuit boards. Moving from top to bottom of the electronics compartment (see also Fig. 4).

• PCB 1: dual-range ESA HV stepping supply; TOF, MCP1 and MCP2 HV supplies

• PCB 2: DC low voltage power supply

• PCB 3: RF low voltage power supply

• PCB 4: RF generator

• PCB 5: RF network (located inside the ESA structure)

The RF coupling circuit (RFCC) is a small board that carries RF control and monitor lines. It also acts as a front plane between the C&DH board and the RF-LVPS. Some of the signals in the RFCC circuit run between the C&DH board and RF Generator. For that reason they are routed through the RF_LVPS and RF backplane to the RF Generator.

Open image in new window

8.1.1 Sample Timing

Data sampling is designed to obtain even coverage of the entire sky (4π sr) in 10 s. Timing, pixel resolution and coverage were discussed in Sect. 4 and summarized in Table 2. Those considerations also determine the rate at which full resolution data are generated [16 (azimuthal) × 16 (elevation) × 64 (energy) × 5 (ion species + background) = 81,920 data words per 10 s]. In fact the actual internal data rates, which include transfer of three sets of TOF spectra binned into 512 samples every 625 ms, are over a thousand times faster.

Acquisition of azimuthal samples is not synced to the spacecraft rotation period. HPCA stepping is allowed to free-run and, since the spacecraft spin rate varies little from 3 rpm, there is little effect on velocity distribution sampling. The slight shift in look directions caused by variation in spacecraft spin rate can be removed easily during data analysis.

Each voltage (or equivalently energy) step involves a number of actions by the C&DH FPGA which (1) sets ESA and RF power supply voltage levels and RF frequency based on pre-programmed tables, (2) waits for ESA voltages to settle, and (3) directs the TOFA electronics to collect TOF and singles data for each step. At each step TOF data from the previous step are sent from the TOFA FPGA to the C&DH FPGA. During Step 0 the ESA HV slews from high to low voltage and any data generated during that step is not saved.

During each energy scan a data set consisting of 63 TOF spectra × 512 TOF bins × 16 elevations is accumulated and histogrammed. The resulting TOF spectra are then parsed into five bins that define the ion species H⁺, He⁺⁺, He⁺, O⁺ and background (Fig. 20). The red portions of the spectrum in Fig. 20 indicate typical species boundaries. Since ion times-of-flight are both mass and energy dependent the range of TOF limits for each species changes with energy (Fig. 31).

Open image in new window

8.1.2 Operational Modes

Figure 32 is a high-level schematic of HPCA flight software (FSW) and operations including the ground system discussed in Sect. 8.4.

Open image in new window

When power is first applied, the HPCA goes into a “Boot” state using software stored in PROM. If the instrument passes software self-tests then it proceeds into a “Safe” state running under science mode software stored in EEPROM. At this point the instrument can be commanded out of Safe mode and into a mode in which high voltages can be enabled. With high voltage enabled and applied, if a self-protect mode is triggered by an incident such as excessively high counting rates, or a watchdog reset event occurs (e.g., no command arrives from the CIDP for 30 s), then the instrument will immediately return to Safe State and protect itself by setting all HV supplies to zero volts in a few seconds. It then reverts to the Boot State to await further instructions.

The instrument can only enter the HV-enabled state by telecommand. From this point on the HPCA is fully functional and ESA, TOFA, MCP and RF supplies can be operated using a selection of tables stored in memory. The TOFA and C&DH subsystems are also operational and science and housekeeping data can be collected, formatted and transmitted. The source of low voltage power to the RF supply is controlled separately by the CIDP. HPCA cannot command on the RF LVPS power directly, but does control the LVPS output levels once the CIDP switch is on.

Science modes are defined by setting the range of ESA voltages to be scanned and the science data products to be returned. The voltage scans are table-driven and just about any combination of voltages (i.e., ion energies) can be selected. For example, instead of using the standard 63 log-spaced stepping intervals, linear scans or scans in which only every other log-spaced step is scanned are all possible subject only to limitations imposed by the ESA power supply slew rates.

On orbit HPCA collects and transmits data according to spacecraft position in the orbit and the modes agreed to by the MMS science team. As discussed in Sect. 7.2.1, the HPCA generates three distinct data sets: Slow Survey, Fast Survey, and Burst. Slow Survey data are highly compressed, Fast Survey Data are much less compressed, and Burst Data are minimally compressed. Survey data products are automatically transmitted to the ground. Burst data (only collected during Fast Survey mode) are stored in the CIDP.

An important feature of the software is that it will autonomously safe the instrument in case of higher than expected detector counting rates. This situation can occur because of failure of the dynamic range control system (e.g., RF operation) or to arcing in the HV electronics. When high rates are detected the software averages the counting rates to make sure the incident is not due to anomalous data, and then commands the MCP voltage down by one or more steps depending on how the safing procedure is set up. Figure 33 demonstrates the operation of this feature during beam tests. In this case the beam current was raised to a level such that the auto turn-down software detected the high rates. At that point the voltage on the MCP was lowered within a few seconds. After a commanded delay period to let the situation settle (for example to let the spacecraft exit a high flux region) the voltage was automatically raised to normal operating levels and then the test was repeated.

Open image in new window

There are several other operational modes designed for specific events that do not involve science, including calibration and maneuver modes. The standard calibration mode generally operates following spacecraft exit from the Fast Survey portion of the orbit in coordination with other particle instruments such as DIS.

8.4.1 Commanding

During flight operations instrument commands are generated at SwRI and then verified by running the command sequences on the HPCA Engineering Model (Fig. 34). Commands are then sent to the SOC for uplink and executed out of a time-tagged sequence run on board the spacecraft. Other than during commissioning no real-time operations are required.

Open image in new window

8.4.2 Data Processing

Figure 34 shows more detail of the ground system and operations. Raw (Level 0) science data arrive at the MMS SOC from NASA ground stations and are streamed from there to SwRI within 30 minutes of receipt. Within another 30 minutes these data are processed automatically by HPCA ground system software into a form that provides the scientists located at the SOC with Level 1B′ data suitable for quick perusal that allows choice of the burst periods to be down-linked on the next spacecraft pass. More complete, but still uncalibrated “quick-look” data (Level 1B), are delivered by SwRI to the SOC within 23 hours of the first receipt of data. Intermediate products, termed Level 1A, are used to produce higher order data products but are not themselves delivered outside SwRI. All Level 1 data are produced with a cadence of 10 s.

Calibration factors are applied to Level 1B data to generate Level 2 products that consist of physical parameters such as velocity distribution functions and moments such as density, flow velocity and temperature for all four ion species. These are provided on the highest available time resolution: 10 s for 3-dimensional parameters such as velocity and 625 ms for 2-D snapshots of elevation, energy, and mass. The final step in data production is Level 3 where the in situ magnetic field and similar auxiliary data are used to derive parameters such as pitch angle distributions, plasma beta, and the Alfven velocity. Additional products such as synthesized energy spectra can be obtained by combining HPCA data (at energies <40 keV) with Energetic Particle Detector (EPD) data above that energy. Both Level 2 and 3 data sets will be delivered to the SOC within 30 days.

The bottom of Fig. 34 indicates that additional calibration data may be taken with the HPCA Engineering Qualification Model (EQM), which has been reworked to be as nearly identical to the flight models as possible—particularly in the critical area of the RF generator, the anodes, and the TOF electronics. The reason for upgrading the EQM to flight performance levels is that it could well be required in case unusual phenomena are encountered outside the anticipated normal range of performance. One example would be unusually high and rapidly varying ion fluxes that exceed dynamic range capabilities.

9.4.1 Faraday Cup

At low beam currents the CRU’s ESA CEM has to be used to monitor current instead of the FC. Figure 45 is a calibration between FC current and the CEM counting rate, C _ESA, giving the relationship

$$\begin{aligned} I_{0} = 1.37 \times 10^{-3} C_{\mathrm{ESA}} - 2.07 \times 10^{-2}\ [\mbox{pA/cm}^{2}]. \end{aligned}$$

(42)

Figure 46 is a plot of ESA CEM counting rate vs. beam current measured by the FC over a wide range of ion species that corresponds to the mass spectrum in Fig. 40. This also demonstrates that the CEM efficiency is very nearly constant with ion mass.

Open image in new window

Open image in new window

9.4.2 High Resolution CRU ESA

9.4.3 Beam Imager

Beam profiling in the Y–Z plane (orthogonal to the beam direction) is made with a commercial Beam Imaging Solutions Beam Observation System Model 40 (BOS-40). The imager’s active diameter is 40.0 mm and its center is offset 368 mm from the rotation center of the motion platform. Rotation is not required for measurement since the profile is taken in a plane perpendicular to the beam direction. The imager combines an MCP backed by a phosphor screen. This creats an image that is relayed to the control computer using an image intensified camera located outside the vacuum system. When projected on the EPIC computer monitor the image can be captured to give a detailed record of beam extent and uniformity. An example of a beam image and profile is shown in Fig. 47. The beam is flat across the top with FWFM of 16.0 mm and FWHM of 22 mm. The HPCA collimator entrance is 14.0 mm high, which means that the beam entirely fills the aperture at FWFM.

Open image in new window