3.2.1.

Structures

The GEOStar-3 structure and mechanical subsystem support the SALTUS Payload, provide the primary load path between the bus and LV, and the means of attachment for four required helium inflatant tanks for the ICS. The structure also supports the propulsion subsystem and provides ample mounting locations and radiative surfaces for the flight system and Payload components. At the aft end, a heritage launch adaptor facilitates the interface between the spacecraft and LV.

3.2.2.

Spacecraft to payload interface

A fully composite truss-based structure—the CIM—provides the means of mechanical attachment and thermal separation between the Payload and spacecraft, as shown in Fig. 5(c). This truss structure is based on JWST heritage and is tuned to be near zero coefficient of thermal expansion (CTE) in the cryogenic environment (see Sec. 3.2.8 regarding thermal stability). The CIM not only provides launch load capabilities and on-orbit thermal stability but also includes accessibility during integration and testing (I&T) activities to route and thermally isolate the multiple cables and lines that span the spacecraft and Payload. On top of the CIM, we designed a truss-based instrument deck leveraging similar JWST heritage and materials. The instrument deck provides the primary mechanical means of attachment for the CCM, ACS bench, and deployable PRM. Similar to the CIM, it is also tuned to near zero CTE at cryogenic temperatures to ensure thermal stability and minimize pointing errors to meet pointing requirements. The SM is also a part of the SALTUS Payload and was designed to interface directly with the CIM and spacecraft maintaining an ideal load path with high stiffness and thermal performance.

3.2.3.

Propulsion

The propulsion subsystem performs all maneuvers to reach the Sun–Earth halo $L2$ orbit as well as MUs and orbit maintenance throughout the mission. The subsystem is a mono-propellant hydrazine system sized to 1093.7 kg maximum expected value (MEV) with 5.8 kg MEV of pressurant based on flight-proven components. Six 22 N $\mathrm{\Delta }\mathrm{V}$ thrusters perform all impulsive maneuvers as well as $X/Y$ attitude control during these maneuvers. Six 5 N ACS thrusters operate for $Z$-axis control during impulsive maneuvers and for MUs. All thrusters mount on the aft end of the spacecraft bus with the $\mathrm{\Delta }\mathrm{V}$ thrusters oriented to provide thrust along the $+Z$-axis and ACS thrusters oriented to provide attitude control in all three axes. The propellant budget utilizes the MEV launch mass, $3\sigma$ low thruster performance, and maturity-based contingencies as well as efficiency factors such as misalignments, flight dynamics errors (relating to $\mathrm{\Delta }\mathrm{V}$ requirement uncertainty), and residual fuel.

3.2.4.

Command and data handling

The command and data handling (C&DH) subsystem provides command, control, and data handling capabilities within the SALTUS flight system via block-redundant integrated electronic modules (IEMs) and block-redundant Payload interface electronics (PIEs). The IEM handles all core command, control, and data handling capabilities and provides triple modular redundancy (TMR) storage for the spacecraft and Payload SOH data storage, providing ample data storage margin. The IEM is augmented with block-redundant PIE units that provide command and telemetry interfaces for each element of the Payload and to the redundant Ka-band system for the downlink of stored SOH and science data. The PIE also provides the ICS controller functionality, monitoring analog telemetry from the ICS and maintaining pressure within M1. In addition, the PIE contains sufficient flash storage with TMR capability for science data storage. This storage capability provides ample margin for science data, as shown in Table 4. The SALTUS PIE is derived from similar units flown on Landsat 9 and JPSS-2 that provided mission-unique functionalities to their respective Payloads and redundant interfaces to the IEM.

3.2.5.

Power

The electrical power subsystem (EPS) supplies power to the spacecraft and Payload and consists of a single deployable, non-articulating solar array wing with a power distribution control, charging unit (PDU), and a Li-ion battery. Figure 4 expands on this architecture, Table 5 summarizes the power budget for driving steady-state modes, and Table 7 includes the solar array and battery performance. The battery is sized to support $\mathrm{\Delta }\mathrm{V}$ maneuvers within the allowed depth of discharge (DOD), accounting for all battery usage efficiency factors as well as capacity fading at EOL. The battery provides a greater margin to all other battery-powered modes. The solar array supports all mission phases and modes with margin, and solar cell efficiency losses for the expected 100°C operating temperatures have been captured. Sized for a maximum Sun angle of 25 deg off the solar array surface normal (per Fig. 2), the non-articulating solar array permits pointing across the entire SALTUS FOR. The driving steady-state power mode is a SAFARI-Lite science observation with HiRX in stand-by with a simultaneous Ka-band science data downlink. The solar array provides a greater margin to all other steady-state mission modes. As a conservative approach, MU and slew modes are on battery power because the solar array is not gimbaled but both could be powered by the solar array, if necessary, as noted in Table 5.

3.2.6.

Telecommunications

The telecommunication subsystem provides both low-rate tracking, telemetry, and commanding (TT&C) capabilities throughout all SALTUS mission phases and high-rate science data downlink capabilities and as a result requires two distinct systems: a low-rate S-band system and a high-rate Ka-band system, as shown in Table 6.

The S-band system is designed for compatibility with select ground and space assets during Launch and Early Orbit Phases (LEOPS) and the NASA Near Space Network LEGS network during cruise and science phases. A pair of low-gain, quadrifilar helix antennas (LGAs) mount onto opposite sides of the spacecraft bus, providing near $4\pi$-steradian coverage during uplink and downlink. Both S-band receivers remain powered throughout all flight system modes, whereas the active S-band transmitter is powered only during post-separation and rate capture, safe mode, and scheduled passes. Each S-band channel is augmented with a medium-gain, monofilar antenna (MGA) that points in the same direction as the 0.6-m Ka-band high-gain antenna (HGA) to provide a higher rate S-band capability for TT&C during science data downlink.

The Ka-band system is compatible with the LEGS network for science data downlink and draws heritage from the JPSS-2 mission. The PIE and Ka-band transmitters modulate and condition the data stream for downlink to LEGS. The body-fixed HGA has a beamwidth greater than the angular size of the Earth at the Sun–Earth halo $L2$ orbit and points in the aft direction. The HGA and MGAs are stowed against the spacecraft bus during launch and are boom-deployed during the cruise phase. Throughout a period of 1 week (optimized via CONOP), SALTUS points its HGA toward the Earth to perform ground contact of 7 h total high-rate Ka-band data downlink and command uplink followed by low-bandwidth, low-rate S-band via LEGS ranging. As noted in Sec. 2.2, an additional 7 h per week of S-band coverage has been baselined for SOH monitoring and other communications as required by mission needs.

3.2.7.

Attitude control

The ACS provides pointing stability, control, and knowledge for science observation while providing sufficient agility across the FOR. SALTUS pointing requirements (as shown in Table 1) are readily met with margin on the LEOStar-3 platform as a result of carefully selected star trackers (STs), inertial reference units (IRUs), reaction wheel assembles (RWAs), and Northrop Grumman flight-proven algorithms (e.g., ICESat-2, Landsat 8/9). We show the ACS bench location and approximate ST FOVs in Fig. 5 and provide performance margins for each ACS mode in Table 7. Attitude stability and control are maintained and stabilized throughout all mission phases, beginning with LV separation. During $\mathrm{\Delta }\mathrm{V}$ maneuvers, the ACS utilizes its $\mathrm{\Delta }\mathrm{V}$ mode providing slew capability to and from the maneuver attitude. In spacecraft safe mode, the ACS nulls spacecraft rates and points the $-Z$ axis at the Sun while keeping the solar array Sun pointed with a slow roll induced about the $Z$ axis. Safe mode therefore maintains a power-positive, thermally safe, commandable attitude that avoids SALTUS pointing constraints. This effectively eliminates solar pressure-induced momentum accumulation and keeps the spacecraft approximately Earth-pointed as well.

Attitude sensor and CCM misalignments are minimized by collocating the dedicated, thermally controlled ACS bench with the CCM on the instrument deck, and the ACS sensors are oriented to provide attitude knowledge in all three degrees of freedom (DoF). Six RWAs manage momentum and slew the observatory, and the six 5 N ACS thrusters provide MU as needed. SRP, driven by the sunshield area (see Sec. 4), dominates the environmental torques during science operations and remains nearly constant during observation. Current predictions of environmental torques acting on the SALTUS observatory require an MU event every 5.4 to 8.0 h (BOL–EOL). A typical science observation is estimated to be $\sim 5\mathrm{h}$, and because contiguous observations are not a requirement, this MU cadence is acceptable. We discuss MU in greater detail in Sec. 5.

The LEOStar-3 platform has extensive heritage providing arcsecond-level precision-pointing and pointing stability capabilities and readily meets the SALTUS pointing requirements with margin. Using Northrop Grumman’s flight-validated simulation capabilities for attitude estimation and control errors and allocations based on heritage programs, the predicted performance of the flight system ACS meets SALTUS needs with a robust margin, as shown in Table 7.

3.2.8.

Thermal control

Thermal environments

The observatory has two distinct thermal zones: (i) the warm side on the $-Z$ side of the sunshield and (ii) the cryogenic side on the $+Z$ side of the sunshield, as shown in Fig. 5(a). Notably, for thermal concerns, the warm side contains the WIM and solar array, whereas the cryogenic side contains the CCM, PRM, and ACS bench. As discussed in Sec. 3.2.2, these environments are mechanically and electrically connected via the CIM. Using the same analytical tools as JWST, we predict an equilibrium temperature of $\sim 310\mathrm{K}$ on the warm side, with radiative cooling to deep space allowing all components on the cryogenic side to cool to $\le 45\mathrm{K}$. This is critical for the exceptional IR sensitivity for SALTUS science. To fully understand the stepdown in temperature from $\sim 310$ to $\le 45\mathrm{K}$, we constructed a cryogenic heat map of SALTUS as shown in Fig. 6. The HiRX and SAFARI-Lite focal planes require cooling to $\sim 5.25\mathrm{K}$; therefore, we selected a build-to-print version of the JWST MIRI cryocooler to provide 31 mW of lift at 5.25 K, plus 110 mW of lift at 20 K for the SAFARI-Lite low-noise amplifiers (LNAs). This performance includes a $>30\%$ margin. We used thermal desktop modeling to ensure that the CCM’s design provided 500 mW of cooling at 45 K and included two side-looking radiators, $1.5{\mathrm{m}}^2$ and $0.5{\mathrm{m}}^2$, respectively, to provide additional radiative cooling needs for HiRX. The MIRI compressor system resides on the warm side of the spacecraft (for thermal isolation) and is coupled to the WIM, where gas lines facilitate the pressurized helium to a valve at the 5.25 K stage of the system on the cryogenic side via the CIM. Northrop Grumman has extensive experience with the MIRI cryocooler and therefore understands its power consumption well. Thus, a $2.7{\mathrm{m}}^2$ radiator (coated with specialized paint) radiates to deep space to cool the WIM. In addition, the instrument radiator $4{\mathrm{m}}^2$ surface area provides a 50% margin on these loads and will be placed orthogonally to the spacecraft-Sun axis for a clear view of deep space. All radiator sizing in this work is a first-order approximation and will be refined and fully addressed in phase A.

Fig. 6

SALTUS cryogenic heat map. Acronyms in the figure: ADR, adiabatic demagnetization refrigerator; CCM, cold corrector module; LNA, low-noise amplifier; LO, local oscillator; QCL, quantum cascade laser; WIM, warm instrument module.

3.2.9.

Flight software and in-flight fault management

Nearly all functionality required for SALTUS is already part of the baseline LEOStar-3 C&DH and ACS flight software (FSW). This FSW has a proven flight track record from missions such as ICESat-2, Landsat 8/9, and JPSS-2, where SALTUS can leverage extensive re-use of existing FSW with modifications limited to SALTUS-unique functions. This setup offers modularity and flexibility for additional features/capabilities in which reconfiguration is accessible on-orbit.

Redundancy via in-flight fault management is implemented throughout all subsystems, where standard single-point failure exemptions are acceptable for structure, fuel lines, tanks, and the HGA. Onboard fault detection via FSW as the primary and hardware as the secondary detects and responds to anomalous conditions as identified. Flexible, heritage FSW TLM monitoring (TMON) is used to manage onboard fault detection and correction with the list of critical TLM mission configurables. The FSW maintains a list of failure symptoms for each component. If TLM violates both the threshold and persistence check, the FSW TMON function responds by executing the appropriate command sequence. As noted previously, safe mode sheds all non-critical loads and achieves Sun pointing using the CSSs, and both S-band receivers are always on. Returning SALTUS to science operations is performed by the mission operations staff on the ground.

3.3.1.

Overview

The SALTUS PRM is made up of three key components: the 14-m M1, the AstroMesh segmented boom, and the AstroMesh truss. The L’Garde M1 membrane is discussed in detail in Ref. 10. This section will focus on the AstroMesh boom and truss, which Northrop Grumman is responsible for in this development.

The AstroMesh reflector was developed in the late 1990s¹¹ for the Thuraya constellation and has been refined over the last 30 years with over eleven successful mission deployments (see Table 8). The AstroMesh architecture is therefore well proven and inherently reliable and offers robust deployment kinematics for a wide range of applications. The AstroMesh truss is designed as an efficient drum-like structure with high deployed stiffness, extremely low mass, and high thermal stability, making it a desirable option for the SALTUS use case. Indeed, its high structural stiffness allows for deployed performance measurements in 1G environments, further enhancing I&T capabilities. Crucially, the architecture was designed with scaling in mind, in which an increase in truss outer diameter (aperture size) is implemented without significant change, thus maintaining the heritage and minimizing mission-unique development.

Table 8

SALTUS AstroMesh truss and boom flight heritage.

The basic AstroMesh architecture as shown in Fig. 7 consists of a flat perimeter truss that supports front and rear nets made of high-precision composite tension “webs,” which define a geodesic surface. The depth of the truss structure is defined by the reflector aperture and radius of curvature. Note that, for the SALTUS application, these webs will be used for structure only and are not needed to form the surface geometry of M1—this is driven solely by the ICS. Coupled with the truss are high-precision, high-heritage deployable booms, also scalable to meet increases in the truss diameter without impact on heritage. The AstroMesh truss and boom combination is fully electrostatic damage (ESD) compliant with a demonstrated passive intermodulation (PIM)-free design. The gold-molybdenum mesh shown in Figs. 7(b)–7(d) is normally used for RF applications and is included in the figure for illustrative purposes only.

Fig. 7

(a)–(d) SALTUS AstroMesh architecture showing the stowed and deployed configurations, as well as two key steps in the truss deployment: the release and bloom and the powered deployment. Note that the golden mesh reflective surface shown in this figure is illustrative only; the SALTUS M1 inflatable membrane will instead occupy this volume.

3.3.2.

Heritage and TRL

Table 8 highlights a range of space flight heritage missions that have successfully deployed the AstroMesh system. The SALTUS 14-m truss diameter is easily within the capabilities of the AstroMesh scalable design (as noted in Sec. 3.3.1) in which all previous missions shown in Table 8 have demonstrated $f/D$ ratios in line with the expected SALTUS M1 $f$-number. Previously demonstrated offset (off-axis) aperture deployments and reflector masses also introduce no additional development for SALTUS given the extremely lightweight mass of the M1 membrane. Table 8 also highlights a heritage assessment for the truss and boom with an associated TRL-6 assigned to each component. We expect to comply with all partial heritage elements, with the key required changes identified as (i) a scaling of the truss from demonstrated 12-m to 14-m geometry and adapted to SALTUS optical configuration (for radius of curvature) and (ii) a scaling of the boom to the deployment length to meet the SALTUS M1 $f$-number, as well as spacecraft accommodation for inflatant lines associated with the ICS (see Sec. 5) and specific actuator step size implementation for boom alignment. These are considered low-risk implementations and a standard part of the mission-unique design for the AstroMesh system.

3.3.3.

Deployment sequence

Figure 8 illustrates the SALTUS PRM deployment sequence in six key steps. We note that each hinge will not be required to fully deploy (or fully precision latch) before a subsequent hinge actuator motion—this flexibility provides an option later for positioning optimizations, clearance optimizations, and assessment of moment of inertia impacts. The finer details of the PRM deployment sequence will be refined in phase A; this subsection presents the basic concept. Below, we outlined the six-step sequential process through full PRM deployment:

Fig. 8

(a)–(f) SALTUS Primary Reflector Module (PRM) deployment sequence. The solar array and HGA are deployed prior to PRM deployment. The sunshield is deployed following PRM deployment. This sequence is important for avoiding a cryogenic PRM deployment (which is a more complex proposition). Note that each hinge will not be required to fully deploy or latch before the next hinge actuator in the sequence initiates its motion. This strategy allows for positioning optimizations, clearance optimizations, moment of inertia impacts, etc. and will be refined in phase A. Step 6 (f) (i–iv) shows the four actuator boom locations.

Note that, during deployment, the retention clips manage the PRM soft goods, which include the front and rear nets as well as the uninflated M1 membrane. These are released regularly and passively throughout the deployment. Once deployed and latched, the spoolers are backed off slightly to reduce cable tension, and the truss deployment is completed.

3.4.1.

Structural vibration analysis

The SALTUS-induced vibration analysis leverages prior high-fidelity models of the spacecraft bus. This model was coupled with a low-fidelity model of the Payload. The Payload was modeled with lumped masses to mimic M1, the truss, and boom (the M1-lumped mass) and the CCM (the CCM lumped mass), as well as springs at the connections to the instrument deck to mimic their estimated first structural modes, as shown in Table 9. In total, the finite element model (FEM) used for this effort has $>9$ million nodes and elements. This analysis does not include deployed SM dynamics—we anticipate that the low mass of each SM layer (Table 10) coupled with the low-frequency first structural mode of the deployed layer will have negligible effects on pointing performance. However, the dynamics of each layer will be included in future analyses once the end-to-end observatory architecture is further refined. This assumption will therefore be revisited in phase A.

Table 9

SALTUS observatory deployed structural mode estimates.

Table 10

SALTUS sunshield module component masses.

Northrop Grumman’s induced vibration analysis technique has been used on multiple programs including ICESat-2, Landsat-8/-9, and JPSS-2. This method utilizes high-fidelity RWA models supplied by the vendor that have been correlated to test data. Both structural modes of the model, as well as the disturbance tones including primary wheel imbalance tones and bearing disturbances, have been correlated. Bearing disturbances have been modeled up to 63 times the RWA wheel speed to ensure that any low wheel speed excitations are correctly captured. This model also includes the appropriate rotor gyroscopic coupling as well as a uniform structural damping value for all structural modes at 0.25%. The modeling technique accurately captures the disturbances as an internal force between the bearings and the RWA housing.

The induced vibration analysis considered two configurations: (1) RWAs “hard-mounted” to the bus and (2) RWAs isolated via a passive mechanical isolation system that leverages high TRL systems (e.g., Chandra and JWST). For the induced vibration analysis, the RWAs are spun up at the same wheel speed in 20 RPM increments up to 4000 RPM, providing a conservative estimate with all of the energy concentrated at the same tones. Note that the maximum allowable wheel speed (outside of safe mode) for SALTUS is 3000 RPM, which provides a 50% margin. The frequency range of interest for this analysis was from 0.01 to 500 Hz. Both translational and rotational displacements were reported at the M1-lumped mass node. The response at the M1 lumped mass node was multiplied by a model uncertainty factor (MUF) to account for model uncertainty and program maturity. The selected MUF is consistent with the concept development phase, and the amplitudes will be reduced as the design matures. The residual sum of squares for the scaled three DoF for translations or rotations was calculated to provide a total translational or rotational motion performance of the system. These performance curves were then used to calculate the RMS by finding the area under the curve and reporting $1\sigma$ values. We show these results in Fig. 9.

Fig. 9

(a) SALTUS observatory isolated versus hard-mounted RWAs induced vibration RMS (1 σ) translational displacements of M1. (b) SALTUS observatory isolated versus hard-mounted RWAs induced vibration RMS (1 σ) rotational displacements of M1.

Finally, we note that the influence of the induced vibration from the cryocooler was not included in the initial system-level analysis. This omission was justified on the basis that the cryocooler disturbance frequencies are held constant throughout its operation and are above the ACS controller bandwidth. Active compensation within the CCM also mitigates any influences from the cryocooler in the optical path. In addition, the low first fundamental frequency of the PRM’s boom will act as a mechanical low-pass filter for cryocooler and RWA disturbances, further minimizing the impact on M1 motion. A detailed system-level assessment that includes the cryocooler disturbances is planned for phase A. Adjustments and tuning of the structure stiffness may be implemented at that time to ensure that no interaction between known cryocooler tones and key structural modes occurs, thus maintaining positive margins.

3.4.2.

PRM settling time after maneuvering

A conservative bang-bang slew was performed to estimate the initial M1 pointing error from a maneuver. The max torque from one RWA is applied about the spacecraft $X$ axis for 50 s; then it is applied in the opposite direction to bring the observatory to rest. This maneuver results in an $\sim 6\mathrm{deg}$ slew about the $X$ axis. Once the body is at rest, the motion of M1 is reported, resulting in an initial displacement of 0.00015 radians. From this initial displacement, various structural damping values were assumed, and using the log decrement method, the displacement is propagated forward in time. As seen in Fig. 10, the M1 rotational displacement as a function of time can be seen for various given structural damping values. For bare metallic or composite structures, a value of 0.15% to 0.25% is typically used, and this range is captured in Fig. 10. However, this is highly conservative and does not include the benefits of other components such as thermal blanketing and harnessing, which all contribute to additional damping. Further, we have an on-orbit point of reference for a large deployable structure from SMAP and its 6-m RBA (Table 8). This included a large boom with two locked-out hinges, similar to the SALTUS design. The first two primary mode values were measured at 0.5% and 0.7%, respectively.

Fig. 10

Rotational displacement of M1 as a function of time for a range of structural damping values.

As noted above, we assumed a very conservative approach for the values and assumptions in settling time estimates. Additional methods can be employed to further mitigate settling time and pointing errors and will be investigated later in the program. These include leveraging a slew profile that smooths the transitions to limit the system jerk, a fly-cast maneuver in which the ACS system actively removes energy from the system by counteracting excitation, and other structural methods to introduce damping (such as damping tape or tuned mass dampers). We discuss settling time as it pertains to on-sky efficiency in Sec. 5.2.1.

In the PRM design, we also considered M1’s stability with respect to M2 and found this to be primarily translation-driven by tip/tilt mechanisms in the boom and the wrist hinge—see Fig. 8(f). This stability was assessed via the FEM of the deployed PRM, assuming a 15-m truss diameter (to hold the 14-m M1), three boom sections, and four single-axis actuators. Each boom section and hinge actuator was measured separately and then combined analytically. Note that the analysis captures at minimum one axis of rotation at each of the four hinge locations that are typically not parallel to any other axis of rotation. These locations are the root hinge between the boom and spacecraft, two intermediate hinges on the boom, and the wrist hinge between the reflector and boom. These are labeled in Fig. 8(f) as (i)–(iv). For a total boom length of $\sim 17\mathrm{m}$, the system is required to produce an M1 tip/tilt step size of $2.42\mu \mathrm{rad}$ (0.0001389 deg) and a linear step size of 0.025 mm. Table 8 shows previously demonstrated Northrop Grumman missions on orbit with a similar performance, and the SALTUS phase A design will refine this trade.

3.4.3.

ACS performance

The SALTUS ACS pointing knowledge, accuracy, and stability provides 1.19 arcsec ($1\sigma$), 1.20 arcsec ($1\sigma$), and 0.66 arcsec ($1\sigma$) at the root hinge, respectively, demonstrating a significant margin to meet spacecraft requirements (180%, 736%, and 189%; see Table 7). Once a science target is acquired, the observatory offers $<0.5\mathrm{arcsec}$ pointing stability over a 20-s period in between CCM fast-steering mirror (M7) adjustments. With this performance, the ACS enables M1 to capture the requested SALTUS science target and propagate it to M2 in the CCM.

4.3.1.

Structure

Figure 11 shows the typical features and layout that will be employed for SM layers. Roll widths vary depending on the material selected (ranging from 36 to 60 in.) and will drive the number of seams required to span the 20-m sunshield width. Seams will be parallel to the long edge with $Z$-folding perpendicular to these seams as the sunshield layer membranes are manufactured. Seaming processes are dependent on the material selected; some options include solvent bonding (CP1), thermal spot bonding (Kapton), or the use of silicone adhesives (e.g., Nusil). Seams typically have grounding “jumpers” installed to ensure that the sunshield membrane is electrically continuous and can dissipate any charge that may build up through a ground connection with the spacecraft.

Fig. 11

SALTUS sunshield layer features and layout.

Perpendicular to the seams are ripstops, typically spaced to create a square grid pattern with the seams. Ripstops are to limit the propagation of a tear should one occur. The installation of ripstops is similar to seams. Trimmed edges, found at the outer perimeter and at the circular cutout where the SM’s cylindrical structure resides (red circle in the center of Fig. 11, and labeled as “Central Structure” in the stowed configuration in Fig. 12), are reinforced to prevent tears and to distribute tensioning load across the sunshield membrane. Constructions vary from using Kapton tape (with a cover overlay to mitigate adhesive leach out) to bonding on strips of the base material to “double up” the thickness at this edge. Corners have additional reinforcements where the grommet attachment point is located, typically a stack-up of stainless steel (3 to 5 mil depending on membrane tensioning loads) and additional covers for optical purposes and grounding features. Connection to the mast can be through a small clevis/pin attached to the grommet for ease of installation/removal to the cable/tensioning system at each of the boom tips. Grounding of the membrane to the spacecraft can be done here or at the CIM interface if attached (or both).

Fig. 12

SALTUS sunshield stowed and deployed configurations. The stowed configuration shows a zoomed-in image of the sunshield module (SM), which sits outside and attaches to the CIM. The smaller $\sim 48.5\mathrm{m}\times \sim 19.5\mathrm{m}$ layer 1 deploys first, followed by the larger $50\mathrm{m}\times \sim 20\mathrm{m}$ layer 2, which prevents a cryogenic deployment of the second layer in the deployment sequence. Layer 2 will be the last major deployment of the SALTUS observatory.

Triangular Rollable and Collapsible (TRAC) masts are constructed from two thin gauge high-strain composite (HSC) tape springs that are bonded along one edge to form a triangular cross-section. This geometry can be flattened and then rolled onto a hub for stowage (see Fig. 12). Thirty-meter TRAC masts, the same length required for the SALTUS sunshield, have been demonstrated and matured to TRL-6 as part of the Solar Cruiser solar sail system. The TRAC mast provides very high bending stiffness for a given flattened height. In other words, the mast provides excellent performance for a minimized system mass and stowage volume. Other rollable cross-sectional geometries, such as the Collapsible Tubular Mast (CTM), could also be considered for the masts. Although the CTM cross-section provides substantially higher torsional stiffness and stability than the TRAC variant, further development work would be needed to fabricate and demonstrate the CTM at the required length and stiffness needed for SALTUS. However, the CTM technology is being matured as part of NASA Langley’s ACS3 mission and could certainly be a viable option for SALTUS, given the timeframe for the mission. Apart from cross-sectional geometry, these two mast technologies are very similar with respect to the composite materials used, as well as their spooling and deployment behavior, and are therefore considered interchangeable within the sunshield subsystem. The composite TRAC masts are manufactured from carbon fiber-reinforced polymer (CFRP) composite materials. The high-modulus carbon fibers are predominantly oriented in the longitudinal axis, providing optimized compression and bending stiffness at a minimized mass. Furthermore, the composite laminate architecture is optimized for near-zero CTE to ensure on-orbit shape stability. The SALTUS SM team has extensive experience with composite mast design, and further optimizations can be considered if necessary.

4.3.2.

Deployment sequence

The sunshield is the last major deployment for the observatory; it follows solar array and HGA deployments and PRM deployment. Sunshield layer 1 (the smaller layer closer to the PRM) is deployed first, thus ensuring that no preceding subsystem/component deployment is a cryogenic deployment, which would require additional cryogenic development and testing. Layer 2 is the final deployment in the sequence. We describe the sequence, in reference to Fig. 13, as follows:

Fig. 13

SALTUS sunshield deployment sequence. (a) Stowed configuration showing each sunshield layer and deployment system location. The stowed layers, deployment system, and TRAC masts are co-spooled on a stowage hub. (b) TRAC masts are deployed using a single brushless DC motor mechanism. (c) The deployment mechanism continues to unfurl the layer. (d) Tension is driven into the layer on full deployment. The corners of each layer are mechanically attached to the distal ends of each mast. (e) Steps 1 to 4 are repeated for the second layer, ensuring a non-cryogenic deployment for both layers. Once both layers are deployed, all major observatory deployments are complete, and the cryogenic thermal environment begins to stabilize to $<45\mathrm{K}$.

Fig. 14

(a) Single-deployed quadrant of the Solar Cruiser sail system utilizing 30 m TRAC masts with a side dimension of 40 m. (b) Membrane manufacturing facility at NeXolve, which is capable of producing sunshield layers to meet the SALTUS required area. (c) A four-mast TRAC deployment system that has been matured to TRL-6.

The deployment mechanism that we selected for SALTUS will heavily leverage the Sail Deployment Mechanism (SDM) designed for the Solar Cruiser solar sail.^12,13 The Solar Cruiser SDM was designed to deploy four 30-m long TRAC booms in a square configuration and has been matured to TRL-6 as part of the Solar Cruiser solar sail flight qualification campaign.

5.1.1.

Overview

The SALTUS propellant tank has been sized to carry a total hydrazine propellant mass of 1093.7 kg MEV. Approximately 40% of this budget is required to reach the SALTUS orbit via impulsive burn and X/Y attitude control maneuvers, corresponding to the $\mathrm{\Delta }\mathrm{V}$ budget in Table 2. The remaining $\sim 60\%$ is required for RWA MUs ($\sim 40\%$) and station keeping ($\sim 20\%$). Our calculations show that MU is dominated by SRP and inertia, both unique to SALTUS, and thus, we consider MU maneuvers to be the dominant propulsion mode that limits the mission lifetime when considering the “propellant consumable” in the two-consumable architecture (the other consumable being ICS, discussed in Sec. 5.2). SRP accumulates across the SALTUS FOR over the mission lifetime (including CG migration), which scales directly from the total sunshield larger layer area of $\sim 1000{\mathrm{m}}^2$. In addition, as shown in the table in Fig. 5(b), the observatory produces very large inertial moments ($>\mathrm{80,000}\mathrm{km}·{\mathrm{m}}^2$ in a deployed configuration at BOL), thus requiring large RWAs to meet an acceptable slewing performance across the FOR. These factors require significantly more propellant mass for MUs than other conventional space missions and ultimately define the lifetime.

We consider MU calculations to be a bounding case because the analysis does not capture strategies that minimize momentum accumulation, e.g., strategic observation planning for efficient target selection, which could extend the mission lifetime even further. Currently, the propellant subsystem provides a positive margin at the 5-year mission baseline with a tank capacity of 50.1% (Table 7). In this subsection, we expand on SALTUS on-sky efficiency and MU and discuss migration of the static margin (i.e., the distance between CG and center of pressure, CP) from BOL to EOL.

5.1.2.

Operational on-sky efficiency

The operational efficiency, $\eta$, of SALTUS for the GTO program is the ratio of time observing to the total time (also characterized as one minus the fraction of time spent not observing), given as

This $\eta$ requirement for SALTUS is 60%. Although the GO observing program is currently unknown, the GTO’s performance discussed herein is directly applicable to the GO program. SALTUS will not observe during slew and/or settling time, MUs, ground contacts, or observatory safe modes. We carried out a parametric assessment assuming a slew, MU, and settle before and after an observation, which shows that the determination of observational efficiency, $\eta$, can be reduced to a single parameter—allowable settle time, $\sigma$. Note that, per Table 5, the worst-case power mode does in fact assume an observation during ground contacts for conservatism in power sizing, and although the observatory does have the ability to accommodate this mode, it should not affect the outcome of $\eta$ or $\sigma$. The (time not observing) variable in Eq. (1) can be broken up into the following components: the observatory overhead, $\gamma$; the percentage of time lost to communications, $n(S+\sigma +L_T)/7.24$; the percentage of time lost to MU, $((24/\delta ).t_d)/24$; and the percentage of time lost to spacecraft safe mode from flight actuals, $W_{\mathrm{sc}}=M/365$, where $M$ is the total time the spacecraft spends in safe hold mode over 1 year. Equation (1) now becomes

Solving for the allowable settle time, $\sigma$, we find

Evaluating Eq. (6) for a range of $L$ and potential values of $\eta$, we use the following analysis to determine input parameters. We assume SALTUS targets are uniformly distributed over the FOR. The observatory is capable of slewing 180 deg in $\sim 30\min$ via six RWAs (see Sec. 3.2.7) at 50% max allowable wheel speeds of 3000 RPM (providing a 50% margin), and using the maximum inertia in the deployed configuration from Fig. 5(b), we find a mean slew $S$ of $\sim 15\min$. $L_T$ is a total of 7 h per week distributed per the science CONOP, contributing 4.2% of a day. $\delta$ is expected to occur at the shortest interval, i.e., the worst-case attitude and at BOL when CG-CP is a maximum, which is every $\sim 4\mathrm{h}$ and takes $\sim 5\min$ to complete. This contributes 2.1% of the day. We combine this with a historical average for LEOStar-3 safe mode duration, $W_{\mathrm{SC}}$, at 3 days per year, or 0.8% per day. If the mean observation, $L$, is 4 h (currently, the mean is estimated to be $\sim 5\mathrm{h}$, but we use 4 h for added conservatism), there are six observations per day, thereby resulting in 5.2% of a day in slewing modes. These activities account for 12.3% of the time. We added $\gamma =25\%$ observatory overhead (i.e., $12.3\%\times \gamma$, where $\gamma =1.25$) to capture uncertainties, thereby providing extra contingency to the design process, yielding 15.4% of time lost per day. This results in $\eta =84.6\%$ observational efficiency, well in excess of the $\eta =60\%$ requirement, as shown in Table 12.

Table 12

SALTUS on-sky efficiency key and driving variables.

Solving Eq. (6) with the above inputs yields an allowable settle time, $\sigma$, of 13 min CBE ($\sim 1\%$ of a day), as indicated by the green horizontal line in Fig. 15, which shows the range of compliance against the $\eta =60\%$ efficiency requirement. We note that, to meet $\eta =60\%$, $<27.6\%$ of the day must be spent settling, or 4.6% of a day per observation ($\sim 1.1\mathrm{h}$). Effectively, this means that, with $\sigma =13\min$, $L$ can be as short as 1 h (recall, the current mean $L=5\mathrm{h}$ CBE), indicating that the SALTUS observing efficiency is highly insensitive to the settling time.

Fig. 15

SALTUS maximum allowable settle time to achieve a given observing efficiency as a function of observation time. For shorter observations, the mean efficiency is greater than the science mission requirements. Given the requirement of $\eta =0.60$, the mean observation time can be as short as 1 h (current mean observation time is 5 h CBE), and SALTUS will still meet this requirement.

5.1.3.

Momentum unloading, torque, and static margin

The interval between MUs is dependent on the separation of the CG to the CP, the observatory attitude, and the sunshield properties. The table in Fig. 5(b) shows observatory CG locations, in meters, in $X$, $Y$, and $Z$ with respect to the spacecraft origin, which we define to be the geometric center of the separation ring plane located at the $\sim \mathrm{aft}$ end of the spacecraft [labeled “Sep plane” in Fig. 5(c)]. We designed CG to approximately reside at the $X$- and $Y$-centers of the GEOStar-3 primary structure. CP is located at the geometric center of sunshield layer 2 (on the spacecraft side; the $50\mathrm{m}\times 20\mathrm{m}$ layer) at $XYZ$ coordinates 0 m, 0 m, $+3.88\mathrm{m}$. By assessing the mass and the $XYZ$ location of all major SALTUS spacecraft components, flight system components, and Payload components, we calculated CG at launch and over the full mission lifetime (BOL-EOL). We assessed the initial CG LV dispersions and linear motion as the consumables are expended over mission life. Because we know the momentum storage capacity of the selected RWAs, we can calculate a map of MU intervals as a function of observatory pitch and roll angles, corresponding to different attitudes in the FOR.

We show this result in Fig. 16(a), where we highlighted in a purple rectangular box the MU interval in hours at BOL as a function of the SALTUS FOR, as defined by $\pm 20\mathrm{deg}$ of pitch and $\pm 5\mathrm{deg}$ of roll. By finding the minimum MU interval in the FOR, we can identify the attitude that generates maximum momentum accumulation. As shown, the worst-case MU interval is 4.71 h at BOL, which is less frequent than the observation length of 4 h used in our efficiency calculation in Sec. 5.1.2. Moreover, this worst-case MU interval increases over the mission life as the propellant and inflatant gas are consumed, and therefore, this 4-h MU interval is used to size the propellant mass over the SALTUS mission lifetime. AT EOL, the MU interval minimum is 6.59 h, as shown in Fig. 16(b). To assess the adequacy of the propellant capability of the proposed tank arrangement, we simulated many likely SALTUS missions via Monte Carlo simulation, varying the fuel needed for achieving orbit under launch dispersions. We ran this Monte Carlo for approximately a few hundred cases of nominal properties and compared it with the 95th percentile storage capability to demonstrate the adequate propellant capability of the proposed SALTUS design. Furthermore, we note that, for the MU interval calculations, we used a brute force approach for each case, in which we assume that the RWAs are loaded to capacity in one direction and are required to unload fully before loading in the other direction. We have therefore not considered a low torque accumulation plan, which would reduce propellant even further. This calculation is thus very conservative.

Fig. 16

SALTUS momentum unload interval in hours as a function of observatory pitch and roll at (a) BOL and (b) EOL over the SALTUS FOR. Each number in a colored cell corresponds to the unload interval in units of hours. The smaller the number is, the shorter the unload interval is, which indicates greater momentum accumulation in the RWAs at that attitude. Intervals increase with the mission lifetime as propellant is used, and CG migrates closer to CP. The purple boxes in (a) and (b) highlight the SALTUS FOR defined by $\pm 20\mathrm{deg}$ of pitch and $\pm 5\mathrm{deg}$ of roll and show a BOL minimum unload interval of 4.71 h and an EOL minimum unload interval of 6.59 h.

The conservative approach to estimating the fuel needed for the SALTUS mission gives us confidence that further design analysis, even if it reveals additional propellant needs, will validate that there is enough propellant storage in our design. It is likely that, with time for more simulation and analysis, we can optimize our propellant storage design further.

5.2.1.

M1 inflation

The ICS is a closed-loop control system that regulates the flow of helium inflatant from four high-pressure ($>4000\mathrm{psi}$) storage tanks on the spacecraft’s primary structure through gas lines on the boom to M1 and maintains the pressure within M1 to meet the performance requirements over all attitudes and through the mission during science operations. The ICS maintains M1 pressure during science observations at $5.1\mathrm{Pa}\pm 5.1\mathrm{mPa}$. The ICS consists of a large number of high TRL components from previous Northrop Grumman missions requiring pressure-regulated systems, where all ICS components have been identified by a commercial catalog number. The pressure in M1 is sensed through six strain gauges made from piezo-electric film bonded to M1 and connected in a four-wire arrangement, located outside the optical clear aperture—see the SALTUS block diagram in Fig. 4. The strain gauge produces a voltage proportional to film strain, which is proportional to pressure and forms the basis of the error signal for the control loop. With large gauge functions, it provides sufficient resolution and accuracy for control of M1’s pressure. The baseline approach is to operate this loop as a bang-bang controller, but other control algorithms will be studied in phase A. Also in phase A, the sensor will be exposed to radiation and low temperature to demonstrate that the behavior of the sensor over these environments is suitable for the SALTUS mission.

The control function is housed in the PIE and communicates directly to the “ICS Controller” in the block diagram (Fig. 4). Based on the error signal, helium gas will either flow into M1 or be released from it. The ICS has sufficient resolution that the errors between the continuous loss rate of M1 and the discrete inflow rates have a mean error of a few nmol/sec, allowing the ICS control loop to maintain pressure within requirements. The SALTUS inflatant gas lasts for the required 5 years of mission life under very conservative assumptions (with an allocation of 200 kg; see Table 3) and is likely to last significantly longer under nominal conditions. Factors contributing to inflatant consumption are penetration by MM, creep-induced hole growth, valve and fitting leakage, and cold-to-hot losses,¹⁰ and permeation.¹⁴ The brass board ICS will be part of phase B testing to show that the combined M1 ICS functions together as a system.

5.2.2.

Micrometeoroids (MM)

MM bombardment causes a linearly increasing leak rate with time, resulting in a time-squared dependence on the inflatant mass.^14,15 To make an accurate estimate of the area of penetrations over time, Northrop Grumman carried out a series of hypervelocity impact tests on representative membranes and produced a model of fragmentation, to accurately determine the total area damaged.¹⁴ Reference 14 outlines these tests, and Ref. 10 modifies this approach to give a calculation of the total area damaged from the full MM flux for membranes of thickness and properties appropriate for SALTUS. As an additional contingency factor to create a worst-case expectation of MM penetration, Northrop Grumman also uses an internally developed tool, MADRAT, for the more specific calculations of the system penetration including factors such as orientation, system shielding, orbital velocity, and natural anisotropies in the environmental flux. MADRAT is validated against NASA’s BUMPER code¹⁶ and has specifically included a factor for hole growth due to visco-elastic creep as part of worst-case estimates of gas consumption. Polymeric films as will be used on M1 are expected to have negligible creep as the operating temperature is low and far from the glass transition temperature.¹⁷ The losses due to cold to hot are assumed to be incurred every other observation and taken for the full venting of 0.04 to 0.07 mols, depending on temperature. The vendor-specified valve losses for both internal and external leaks are assumed to be entirely external, then doubled as an additional contingency. Assessment of the inflatant lifetime is performed with worst-case and nominal CBE parameters. The 5-year inflatant lifetime is met under worst-case and nominal conditions.

Finally, we note that, compared with JWST, although the SALTUS lifetime is more sensitive to MM impacts, the wavefront performance of M1 is far less sensitive to these effects. Because JWST has a rigid mirror system, the shape of a damaged rigid mirror element is permanently changed by MM impacts. Crucially, any SALTUS M1 shape change is mitigated by ICS inflation rate adjustments, which reestablish the shape and eliminate residual wavefront error. We refer the reader to Ref. 10 for an in-depth discussion of M1 design, properties, performance, MMs, and lifetime.