- Email: [email protected]
Status of the relativity mission superfluid helium flight dewar1
Cryogenics 39 (1999) 369–379
Status of the relativity mission superfluid helium flight dewar1 D.C. Read a
a,*
, R.T. Parmley b, M.A. Taber c, D.J. Frank a, D.O. Murray
a
Lockheed Martin Missiles and Space Company, Advanced Technology Center, 3251 Hanover Street, Palo Alto, CA 94304, USA b Consultant, USA c W.W. Hansen Experimental Physics Laboratory, Stanford University, Stanford, CA 94305, USA
Abstract The world’s largest capacity helium flight dewar has been assembled for use on the Relativity Mission, also known as Gravity Probe-B (GP-B). Acceptance tests have been successfully performed and the dewar has been delivered to Stanford University. The science mission dewar (SMD) is the first piece of flight hardware delivered for GP-B. It will be used in testing to demonstrate payload functionality with a prototypical Science Instrument Assembly, the flight spare probe (Probe-B), and prototypical electronics. Since delivery to Stanford, the flight dewar has undergone preparations for integration of Probe-B with the dewar. This involved filling both the main tank and the well with liquid helium and installing a superconducting ultralow magnetic field shield. This shield consists of 63-m-thick lead foil which lines the inside of the dewar well and provides an ambient magnetic field environment of ⬍ 10 pT (0.1 G) in the region of the science instrument assembly. The installation of the shield is an iterative process which will be described briefly. Once the shield is installed, the dewar and the shield must be kept cold through to the end of the science mission. The ambient field inside the shield that has just been installed has been measured to be 5 pT or less in the gyro region. After completion of testing with Probe-B, the flight probe (Probe-C) will be integrated in early 1998 in preparation for testing the flight payload. This paper reviews the key design features of the dewar, especially those driven by the ultralow magnetic field requirements. The measured mass, helium boil-off, and ground hold performance are compared to those of the previously launched superfluid helium cryostats (IRAS, COBE, and ISO). 1999 Elsevier Science Ltd. All rights reserved. Keywords: Space cryogenics; Superfluid helium dewar; Gravity Probe-B
1. Introduction The Relativity Mission is an orbital test of Einstein’s general theory of relativity using gyroscopes. The precession of the gyroscopes will measure both the geodetic effect (6.6 arc-s/year) through the curved space–time surrounding the Earth and the motional effect (0.042 arcs/year) due to the rotating Earth dragging space–time around with it. To achieve the extraordinary accuracy needed to measure these small processions, it is necessary to have the gyroscopes operating in the following environments: a vacuum of ⬍ 10−11 torr, an acceleration level of ⬍ 10−10 g, a magnetic field of ⬍ 10 pT, and a temperature near 2 K. The flight dewar is critical in
meeting the acceleration, magnetic field, and temperature requirements for proper operation of the science instrument assembly (SIA), and it must provide a minimum orbital lifetime of 16.5 months in order to achieve the required experimental accuracy. The SIA consists of four gyroscopes, superconducting quantum interference device (SQUID) readouts for measuring the gyroscope precession, and a telescope that, in conjunction with a guide star, provides a frame of reference against which the precession can be measured. The SIA fits inside a long cylindrical container called the Probe (Fig. 1); the sealed probe in turn fits inside the well of the dewar (Fig. 2). 2. Dewar overview 2.1. Dewar description
* Corresponding author. 1 Paper presented at the 1997 Space Cryogenics Workshop, Eugene, OR, USA, 4–5 August 1997.
The dewar shown in Fig. 3 is 3.0 m high, 2.2 m in diameter, and weighs 810 kg dry. The main tank con-
0011-2275/99/$ - see front matter 1999 Elsevier Science Ltd. All rights reserved. PII: S 0 0 1 1 - 2 2 7 5 ( 9 9 ) 0 0 0 4 4 - 2
370
D.C. Read et al. / Cryogenics 39 (1999) 369–379
Fig. 1.
The probe houses the science instrument assembly and provides all its service cables and lines.
Fig. 2.
The probe is inserted into a cold dewar through a helium-purged air lock.
tains 2328 l of He II at 1.8 K, weighing 338 kg with a vapor ullage of 5%. The guard tank used to simplify ground servicing contains a maximum of 99 l of He I at 4.2 K, which is reduced to 30 l prior to launch by boiling off the liquid, using heaters. The guard tank allows the He II main tank to be kept nonvented for months, while the guard tank only has to be filled once a week. Consequently, all of the helium from the main tank is available for use on orbit, as opposed to a system that pumps on the subatmospheric main tank right up until shortly before launch. The 30 l left in the guard tank allows for two launch aborts plus launch separated by 24-h intervals with no dewar servicing required. The main tank is supported by 12 passive orbital disconnect strut (PODS-V) high-efficiency supports off the
vacuum shell, while the guard tank is supported off the composite neck tube. The ␥ alumina neck tube with a 0.03-mm TiV15Cr3Al3Sn3 foil liner provides a low-heatleak, vacuum-tight cylinder connecting the main tank to the vacuum shell. This cylinder allows the well to be filled with He I during lead bag expansions and probe insertion. Four perforated 1100 aluminum face sheets with 5052 aluminum-honeycomb-core vapor-cooled shields plus double-aluminized Mylar/silk net spacers provide thermal protection inside the evacuated vacuum shell. 2.2. Design drivers The design of the GP-B dewar is driven by the demanding requirements on the gyroscope environment.
D.C. Read et al. / Cryogenics 39 (1999) 369–379
Fig. 3.
371
Cutaway of science mission dewar (SMD).
The requirements that have the most profound influence on the GP-B dewar design are the ultralow magnetic field requirement, cold insertion of the probe, meeting the mass and lifetime requirements, and providing a stable mounting platform for the star trackers. The ultralow magnetic field is achieved by the successive expansions of superconducting lead magnetic shields. This technique will be described briefly in this paper and has been discussed extensively in previous papers [1]. This expansion must be performed when the shield is below the superconducting transition temperature ( ⬍ 7.2 K) of the lead and in a very stable thermal environment. Once the expansions are completed and the desired magnetic field is achieved, the remaining lead shield must be maintained in a superconducting state or the low field is lost. Therefore, the gyroscopes must be inserted into the low field environment while it is at helium temperatures without increasing the temperature of the lead shield above its transition temperature. To perform these operations, the dewar must provide a separate volume, called the well, for the expansion of the lead bags. This volume must be accessible from the ambient environment and must be capable of being filled with liquid helium. The well volume cannot contain superfluid helium in a zero-g environment since it communicates thermally with the warm vacuum shell, hence it must be filled and evacuated separately from the main tank. Since it is filled with helium during lead bag operations, it must be isolated from the guard vacuum even though the well volume is evacuated once the probe is installed. This vacuum isolation is accomplished with a titanium-foil-lined composite neck tube. The composite neck tube also provides thermal isolation, since it must seal at the main tank well (1.8 K) and at the vacuum
shell (230 K). The probe must be inserted through a helium-purged air lock into the liquid helium filled well containing a lead bag magnetic shield. As part of the insertion process the probe is thermally/structurally attached to the helium tank, thermally grounded at four stations in the composite neck tube and sealed to the vacuum shell. Axial lock clamps [2] provide the necessary clamping force for good thermal contact and structural support between the probe and main tank at 2 K. All this must be done without increasing the temperature of the magnetic shield above 6.5 K Mass and lifetime are always design drivers for a space cryogenic system, and significant effort was expended to meet the lifetime while minimizing mass. The helium lifetime is dominated by parasitic heat loads so an efficient thermal isolation system is critical. Twelve passive orbital disconnect struts (PODS) [3] are used to support the main tank and minimize the support heat leak. The gamma alumina/epoxy neck tube with a 0.03-mm 15-3-3-3 titanium alloy foil liner provides a low-heat-leak, vacuum-tight cylinder connecting the dewar well to the ambient environment. This neck tube has the thermal stop rings [4] which are the thermal/mechanical interface to the probe and allow the probe to be vapor cooled, improving lifetime performance. The insulation system of alternating layers of double aluminized Mylar and up to three silk net spacers was selected based on subscale test results showing this construction to have the minimum heat leak compared to other spacers, such as polyester. Dewar mass was minimized by a highly optimized structural design on the main tank and vacuum shell and the use of honeycomb vapor cooled shields [5]. The vacuum structures were optimized for both internal and
372
D.C. Read et al. / Cryogenics 39 (1999) 369–379
external pressure requirements to achieve a minimum mass which resulted in waffled domes and ring stiffened cylindrical sections, shown in Fig. 4. Honeycomb vaporcooled shields (Fig. 5) have a 38% lower mass with equivalent stiffness to the more traditional monocoque vapor cooled shield design. Table 1 shows the mass and mass savings associated with these components. The alignment and stability of the externally mounted star trackers relative to the science instrument telescope is critical for mission success. The dewar must provide a mounting platform for the star trackers that is referenced to the telescope and is stable to less than 10 arcs for the orbital thermal environment where the dewar vacuum shell temperature is 230 ⫾ 5 K. The normal number of six PODS supports was increased to 12 to meet this requirement (Fig. 6). The six forward supports are located in a plane and are attached to a graphite/epoxy ring at the warm end of the struts. The graphite/epoxy ring is attached to the 2219 aluminum vacuum shell with titanium flexures. As the tank cools down, the thermal loads put into the struts are low with this arrangement. In orbit, changes in vacuum-shell dimension due to temperature changes also put minimal thermal strains on the struts because of the ultralow expansion characteristics of the graphite/epoxy support ring. This graphite ring is used to mount the star trackers, providing a very stable mounting surface that is referenced to the science instrument telescope.
3. Unique features The porous plug is unique in that it must operate over a large range of flowrates, since the helium boil-off is used for attitude control of the spacecraft. The primary requirements that affected the design of the porous plug were to operate over a flowrate range of 4–16 mg/s without choking while maintaining a constant bath temperature of 1.8 K and to operate with sufficient back pressure
Fig. 4.
Light-weighted aft vacuum shell.
so that the vented gas can be used for spacecraft attitude control. The nominal flowrate during the science phase of the mission is 6.8 mg/s, which is sufficient for attitude control. Higher flowrates are needed earlier in the mission to perform attitude control. During most operations, the dewar is vented at its nominal flowrate, and a portion of the flow is “nulldumped” if not needed for attitude control. If additional thrust is needed, the downstream pressure is decreased to increase the vent rate. The bath temperature is maintained during these periods of nonequilibrium flowrate by turning on a heater in the helium bath. As flowrate is increased by decreasing the back pressure, the temperature gradient across the plug increases until a choked flow condition is achieved and the flowrate vs. temperature drop is no longer linear. The flowrate at the end of this linear range of operation is referred to as the critical flowrate. As the flowrate is increased and subsequently decreased, the temperature gradient goes through a hysteresis loop. Fig. 7 shows typical temperature gradient and back-pressure data from the acceptance testing of the porous plug. This shows that the slope of the flowrate versus temperature drop in the linear region is decreasing with lower bath temperatures and the critical flowrate is also decreasing. A unique design was used to facilitate safe venting of the helium during a catastrophic loss-of-vacuum. If the guard vacuum is compromised, the heat rates to the helium tank are very high due to the condensation of air on the tank. These high heat rates cause the helium tank pressure to increase until a burst disc ruptures to relieve the tank pressure. For GP-B, a 5-cm-diameter orifice is required to vent the helium and ensure that the pressure does not exceed the design pressure of 0.26 Mpa. If the helium is vented into the vacuum space, the vacuum shell internal pressure will increase until the burst discs on the vacuum shell rupture and the helium is vented into the ambient environment. This internal pressure can become large, depending on the assumptions about the warm-up (and hence the volume increase) of the gas in the vacuum shell. A conventional vent line with a 5-cm diameter would have an unacceptable increase in parasitic heat leak to the tank. A flexible vent line was developed to safely vent the GP-B dewar in an emergency situation. The vent line is constructed with laminated Dacron fiber/Mylar sheets that minimize heat conduction into the dewar and can withstand 1.1 MPa internal pressure. The vent line is installed in the dewar in a folded configuration (Fig. 8) to eliminate radiation tunneling down the vent line. If there is an emergency vent situation, the vent line expands to 5 cm in diameter and vents the helium outside the dewar without increasing the vacuum shell pressure. An inflated flexible vent line immersed in liquid nitrogen during development testing is shown in Fig. 9. Another advantage of the flexible vent line is
D.C. Read et al. / Cryogenics 39 (1999) 369–379
Fig. 5.
373
(a) Cone section of first vapor cooled shield. (b) First (of four) vapor cooled shields.
that refurbishment of the dewar after a loss-of-vacuum will be easier, since the multilayer insulation (MLI) and vapor-cooled shields will not be damaged. Venting directly into the vacuum space would require a rebuild of the thermal protection system.
4. Performance summary
Acceptance test results have been reported previously [6]. The critical performance parameters of mass, para-
374
D.C. Read et al. / Cryogenics 39 (1999) 369–379
Table 1 Weight savings Assembly
Flight weight (kg)
Main tank (including a 51-kg proton shield) Guard tank Vapor cooled shields Vacuum shell
Fig. 6.
Savings (kg)
262 15 72 345
80 4 44 100
Comment Waffle domes, ring-stiffened cylinder, 2219 aluminum Lightening holes in support cone Perforated aluminum honeycomb shields Ring stiffened and waffle construction, 2219 aluminum
Main tank with cold plumbing, guard tank, composite neck tube, PODS-V supports, graphite rings, and support ring.
sitic heat leak, and ground hold will be summarized and compared with previous flight cryostats. The dewar, dry, weighed 810 kg. The measured net parasitic heat rate to the 4.2 K main tank was 136 mW (compared to a predicted value of 112–136 mW), with a vacuum shell temperature of 294 K. The measured value falls within the 30% margin added to the predicted value. This margin has been used throughout the program. The 1.8-K nonvented main tank temperature rose at a rate of 0.002 K/day with the guard tank filled. The 4.2-K guard tank boiled off at a rate of 5%/day (probably too low a value because thermal equilibrium was not achieved for this test). All these test results meet the requirements for weight, lifetime, and launch pad servicing.
4.1. Performance comparison of GP-B to previous flight cryostats Performance comparisons were made in three major areas based on the acceptance tests data. Fig. 10 shows the 30% weight savings achieved compared to the similarly sized ISO cryostat. Even larger savings are achieved if you eliminate the 51-kg proton shield dead weight. The thermal performance improvement of the GP-B cryostat with the SIA and probe installed (Table 2) is due to the use of passive orbital disconnect struts (PODS-V), the carefully tailored double-aluminized Mylar radiation shields (149 each), the silk net low-conductance spacers (three each at the tank reduced progressively to one each
D.C. Read et al. / Cryogenics 39 (1999) 369–379
Fig. 7.
Fig. 8.
375
Effect of bath temperature on temperature drop and back pressure.
Cryogen tank emergency vent line is shown in both nonventing and venting configurations.
at the vacuum shell) plus the use of four vapor-cooled shields. The remarkable ground-hold performance of keeping the GP-B main tank nonvented for 90 days (Table 3) is due to the guard tank driving the inner vapor-cooled shield to 4.2 K (top) to 5 K (bottom); the main tank tem-
perature rises from 1.6 to 1.85 K in this time period. The He I guard tank is filled once a week and no helium is lost from the main tank. Both IRAS and COBE pumped on the main tank up to 26 h before launch. Consequently, they had less helium for use in orbit.
376
D.C. Read et al. / Cryogenics 39 (1999) 369–379
Fig. 9.
Fig. 10.
Development tests of the flexible emergency vent line verified design requirements were satisfied.
Dry weight/He II weight ratio comparison of dewars.
Table 2 Thermal performance comparison of He II flight cryostats with science instruments turned off (TH ⫽ 294 K) Flight payload
Tank volume (l)
COBE
660
ISOa IRAS GP-Ba,b
2300 550 2441
% Boil-off per Normalized day boil-off rate 1.25 0.68c 0.83 0.83d 0.27
4.6 2.5 3.1 3.1 1.0
a
Guard tank not used. b Dewar boiloff measured. Probe/SIA heat leak calculated. 30% margin. c Aperture cover actively cooled to 5 K with separate He I flow. d Aperture cover bottle filled with supercritical helium at 5 K.
5. GP-B integrated system test program Early in the GP-B program, it was recognized that the unconventional requirements imposed by the supercon-
ducting shield would require significant development effort. Integration of a warm cryostat probe with a cold dewar while maintaining a lead shield in the superconducting state had never been attempted on this scale before. Of particular concern was the fact that the lead shield has a radial clearance of only 4 mm from the outer shell of the probe. As a result of the need to develop this technique as well the need to resolve a number of other integrated-system development issues, one development dewar (the engineering development dewar, EDD), and two development probes (Probes A and B) were built. This hardware has been used in a series of three full-scale integrated-system tests. We are now preparing for a fourth and final test (GTU-2) prior to the integration of the flight payload. GTU-2 will involve the integration of Probe-B in the SMD. Key objectives for GTU-2 are to verify integration hardware and procedures with the SMD, to verify the performance and compatibility of the SMD with Probe-B, and to resolve open issues remaining from previous tests. Upon completion of GTU-2 near the end of 1997, Probe-B will be removed and the flight probe, Probe-C will be installed in preparation for final payload tests. After completion of the payload test, the payload will be shipped to Lockheed Martin for integration and testing with the spacecraft. 6. Installation and verification of the superconducting shield Installation of the superconducting lead-foil shield needed to provide the ultralow magnetic field environment required by the GP-B experiment is a lengthy iterative process. A folded cylindrical lead foil shield which is closed at one end of the cylinder is mounted from a cloth sleeve and installed into a vacuum-tight glass cooling tube. The cooling tube with the new folded lead
D.C. Read et al. / Cryogenics 39 (1999) 369–379
377
Table 3 Cryostats ground hold performance at the launch pad Cryostat
Helium loaded
Percentage ullage
(kg) IRAS COBE ISO GP-Bb
72 89 331 338
7 7 1 5 max to 2 min
Guard tank
Initial orbital vent rate (compared to orbit equilibrium value)
Used?
Fill cycle (days)
No No Yesa Yesc
– – 4 7
Main tank nonvented time
(days) Above Above Above Below
1.1 2 to 3 Up to 4.2 Up to 90
Mounted on main tank. Guard tank vapor cools inner vapor cooled shield to ⫺ 30 K. Based on dewar acceptance test data. Includes calculated probe heat leak with 30% margin. c Guard tank, inner vapor cooled shield run at 4.2–5 K. a
b
shield is then installed in the dewar well inside of a previously installed shield. The new shield can then be warmed and cooled through the superconducting transition temperature of lead (7.2 K) by controlling the partial pressure of helium exchange gas in the cooling tube. Once the new shield is cooled in this manner, it will essentially pin the ambient magnetic flux. The previous shield is now removed from the dewar, the bottom of the cooling tube is opened, and the cooling tube is then removed, leaving the new shield hanging in the dewar well. The new shield is mechanically unfolded using an expansion tool so that the shield assumes its cylindrical shape. Since the magnetic flux is pinned in the lead foil, it now becomes spread over a much larger volume thereby reducing the magnitude of the field. The magnitude of the field can be measured by means of a SQUIDbased magnetometer. Field reductions by a factor as large as 100 can be achieved by this technique, and it can be repeated as many times as necessary to achieve our requirement (10 pT). This process is depicted schematically in Fig. 11. The shield that was recently installed in the science mission dewar was obtained after four iterations. After the internal field is determined to be acceptable in the final shield, it is detached from its cloth sleeve and left in the bottom of the dewar well (see Fig. 12). Measurements of the field in the flight lead shield after completion of installation have indicated that the radial component of the field is ⱕ 5 pT (0.05 G) in the region that will be occupied by the science gyroscopes. In addition, the axial component, which in the absence of physical defects in the shield will be much smaller than the radial components, has been measured to be constant to within 5 pT over the same region. These results are consistent with the field magnitude being well below the requirement of 10 pT. Future measurements will also be made to verify that the shield meets its ac field attenuation requirements. Before a probe can be integrated into the dewar, a
Fig. 11. Lead Shield (LS) field reduction technique: (a) previous LS positioned in dewar well; (b) next LS enclosed in cooling tube under controlled helium atmosphere and installed inside the previous LS; (c) cooling tube diaphragm cut out; (d) removal of previous LS; (e) removal of cooling tube; (f) LS ready for expanding; (a) LS expanded.
device called a lead bag retainer is installed inside the shield (Fig. 13). This device consists of a number of linear leaf springs (much like Venetian blinds) made of titanium copper alloy which hang inside the shield. It serves to protect the foil shield and to center the probe during probe insertion, and it also insures good thermal contact between the shield and the cold well wall while providing thermal isolation to the probe shell. In addition to the leaf springs, there is a basket-shaped structure to support the bottom of the shield during launch accelerations.
7. Conclusions The SMD acceptance test results show that the dewar meets all relativity mission requirements. The use of
378
D.C. Read et al. / Cryogenics 39 (1999) 369–379
Fig. 12. View of the interior of the dewar well with the flight lead shield installed. Also visible are the composite necktube with thermal stop rings and the primary probe-dewar thermal interface at Station 200 (the intermediate annular region). Station 200 is 0.87 m from the top of the dewar and the portion of the well below Station 200 is 1.85 m in length.
Fig. 13. View of the interior of the dewar after installation of a lead bag retainer. The retainer shown here is one used for testing purposes. The flight retainer will be very similar except it will have modifications; to protect the shield from being heated above its superconducting transition temperature during launch.
D.C. Read et al. / Cryogenics 39 (1999) 369–379
advanced technology in weight reduction techniques, PODS-V supports, multilayer insulation, vapor-cooled shields, and guard tank design make this dewar the lightest, per unit mass of He II stored, the most thermally efficient, and the easiest to service at the launch site (compared to the three previous free flight cryostats that have been launched, IRAS, COBE, and ISO). The magnetic field achieved meets the program requirements for the SIA.
Acknowledgements This work was performed for Stanford University under subcontract PR8709 and funded by the NASA Marshall Space Flight Center via contract NAS 8-39225.
379
References [1] Taber MA, et al. Production of ultralow magnetic field for Gravity Probe B (GP-B). In: Advances in cryogenic engineering, vol. 39. New York: Plenum Press, 1994. [2] Parmley RT, et al. Installing scientific instruments into a cold Lhe dewar—the Gravity Probe-B approach. In: Advances in cryogenic engineering, vol. 35. New York: Plenum Press, 1990. [3] Spradley IE, Parmley RT. Design and test of a modified passive orbital disconnect strut (PODS-IV). In: Advances in cryogenic engineering, vol. 33. New York: Plenum Press, 1988. [4] Parmley RT. Unique cryogenic features of the Gravity Probe B (GP-B) experiment. In: Advances in cryogenic engineering, vol. 33. New York: Plenum Press, 1988. [5] Abrahamson TE, et al. Weight reduction approaches for helium flight dewars. In: Advances in cryogenic engineering, vol. 39. New York: Plenum Press, 1994. [6] Parmley RT, et al. Performance of the relativity mission superfluid helium flight dewar. Presented at the COSPAR Conference, Birmingham, UK, 1996.
Status of the relativity mission superfluid helium flight dewar1 D.C. Read a
a,*
, R.T. Parmley b, M.A. Taber c, D.J. Frank a, D.O. Murray
a
Lockheed Martin Missiles and Space Company, Advanced Technology Center, 3251 Hanover Street, Palo Alto, CA 94304, USA b Consultant, USA c W.W. Hansen Experimental Physics Laboratory, Stanford University, Stanford, CA 94305, USA
Abstract The world’s largest capacity helium flight dewar has been assembled for use on the Relativity Mission, also known as Gravity Probe-B (GP-B). Acceptance tests have been successfully performed and the dewar has been delivered to Stanford University. The science mission dewar (SMD) is the first piece of flight hardware delivered for GP-B. It will be used in testing to demonstrate payload functionality with a prototypical Science Instrument Assembly, the flight spare probe (Probe-B), and prototypical electronics. Since delivery to Stanford, the flight dewar has undergone preparations for integration of Probe-B with the dewar. This involved filling both the main tank and the well with liquid helium and installing a superconducting ultralow magnetic field shield. This shield consists of 63-m-thick lead foil which lines the inside of the dewar well and provides an ambient magnetic field environment of ⬍ 10 pT (0.1 G) in the region of the science instrument assembly. The installation of the shield is an iterative process which will be described briefly. Once the shield is installed, the dewar and the shield must be kept cold through to the end of the science mission. The ambient field inside the shield that has just been installed has been measured to be 5 pT or less in the gyro region. After completion of testing with Probe-B, the flight probe (Probe-C) will be integrated in early 1998 in preparation for testing the flight payload. This paper reviews the key design features of the dewar, especially those driven by the ultralow magnetic field requirements. The measured mass, helium boil-off, and ground hold performance are compared to those of the previously launched superfluid helium cryostats (IRAS, COBE, and ISO). 1999 Elsevier Science Ltd. All rights reserved. Keywords: Space cryogenics; Superfluid helium dewar; Gravity Probe-B
1. Introduction The Relativity Mission is an orbital test of Einstein’s general theory of relativity using gyroscopes. The precession of the gyroscopes will measure both the geodetic effect (6.6 arc-s/year) through the curved space–time surrounding the Earth and the motional effect (0.042 arcs/year) due to the rotating Earth dragging space–time around with it. To achieve the extraordinary accuracy needed to measure these small processions, it is necessary to have the gyroscopes operating in the following environments: a vacuum of ⬍ 10−11 torr, an acceleration level of ⬍ 10−10 g, a magnetic field of ⬍ 10 pT, and a temperature near 2 K. The flight dewar is critical in
meeting the acceleration, magnetic field, and temperature requirements for proper operation of the science instrument assembly (SIA), and it must provide a minimum orbital lifetime of 16.5 months in order to achieve the required experimental accuracy. The SIA consists of four gyroscopes, superconducting quantum interference device (SQUID) readouts for measuring the gyroscope precession, and a telescope that, in conjunction with a guide star, provides a frame of reference against which the precession can be measured. The SIA fits inside a long cylindrical container called the Probe (Fig. 1); the sealed probe in turn fits inside the well of the dewar (Fig. 2). 2. Dewar overview 2.1. Dewar description
* Corresponding author. 1 Paper presented at the 1997 Space Cryogenics Workshop, Eugene, OR, USA, 4–5 August 1997.
The dewar shown in Fig. 3 is 3.0 m high, 2.2 m in diameter, and weighs 810 kg dry. The main tank con-
0011-2275/99/$ - see front matter 1999 Elsevier Science Ltd. All rights reserved. PII: S 0 0 1 1 - 2 2 7 5 ( 9 9 ) 0 0 0 4 4 - 2
370
D.C. Read et al. / Cryogenics 39 (1999) 369–379
Fig. 1.
The probe houses the science instrument assembly and provides all its service cables and lines.
Fig. 2.
The probe is inserted into a cold dewar through a helium-purged air lock.
tains 2328 l of He II at 1.8 K, weighing 338 kg with a vapor ullage of 5%. The guard tank used to simplify ground servicing contains a maximum of 99 l of He I at 4.2 K, which is reduced to 30 l prior to launch by boiling off the liquid, using heaters. The guard tank allows the He II main tank to be kept nonvented for months, while the guard tank only has to be filled once a week. Consequently, all of the helium from the main tank is available for use on orbit, as opposed to a system that pumps on the subatmospheric main tank right up until shortly before launch. The 30 l left in the guard tank allows for two launch aborts plus launch separated by 24-h intervals with no dewar servicing required. The main tank is supported by 12 passive orbital disconnect strut (PODS-V) high-efficiency supports off the
vacuum shell, while the guard tank is supported off the composite neck tube. The ␥ alumina neck tube with a 0.03-mm TiV15Cr3Al3Sn3 foil liner provides a low-heatleak, vacuum-tight cylinder connecting the main tank to the vacuum shell. This cylinder allows the well to be filled with He I during lead bag expansions and probe insertion. Four perforated 1100 aluminum face sheets with 5052 aluminum-honeycomb-core vapor-cooled shields plus double-aluminized Mylar/silk net spacers provide thermal protection inside the evacuated vacuum shell. 2.2. Design drivers The design of the GP-B dewar is driven by the demanding requirements on the gyroscope environment.
D.C. Read et al. / Cryogenics 39 (1999) 369–379
Fig. 3.
371
Cutaway of science mission dewar (SMD).
The requirements that have the most profound influence on the GP-B dewar design are the ultralow magnetic field requirement, cold insertion of the probe, meeting the mass and lifetime requirements, and providing a stable mounting platform for the star trackers. The ultralow magnetic field is achieved by the successive expansions of superconducting lead magnetic shields. This technique will be described briefly in this paper and has been discussed extensively in previous papers [1]. This expansion must be performed when the shield is below the superconducting transition temperature ( ⬍ 7.2 K) of the lead and in a very stable thermal environment. Once the expansions are completed and the desired magnetic field is achieved, the remaining lead shield must be maintained in a superconducting state or the low field is lost. Therefore, the gyroscopes must be inserted into the low field environment while it is at helium temperatures without increasing the temperature of the lead shield above its transition temperature. To perform these operations, the dewar must provide a separate volume, called the well, for the expansion of the lead bags. This volume must be accessible from the ambient environment and must be capable of being filled with liquid helium. The well volume cannot contain superfluid helium in a zero-g environment since it communicates thermally with the warm vacuum shell, hence it must be filled and evacuated separately from the main tank. Since it is filled with helium during lead bag operations, it must be isolated from the guard vacuum even though the well volume is evacuated once the probe is installed. This vacuum isolation is accomplished with a titanium-foil-lined composite neck tube. The composite neck tube also provides thermal isolation, since it must seal at the main tank well (1.8 K) and at the vacuum
shell (230 K). The probe must be inserted through a helium-purged air lock into the liquid helium filled well containing a lead bag magnetic shield. As part of the insertion process the probe is thermally/structurally attached to the helium tank, thermally grounded at four stations in the composite neck tube and sealed to the vacuum shell. Axial lock clamps [2] provide the necessary clamping force for good thermal contact and structural support between the probe and main tank at 2 K. All this must be done without increasing the temperature of the magnetic shield above 6.5 K Mass and lifetime are always design drivers for a space cryogenic system, and significant effort was expended to meet the lifetime while minimizing mass. The helium lifetime is dominated by parasitic heat loads so an efficient thermal isolation system is critical. Twelve passive orbital disconnect struts (PODS) [3] are used to support the main tank and minimize the support heat leak. The gamma alumina/epoxy neck tube with a 0.03-mm 15-3-3-3 titanium alloy foil liner provides a low-heat-leak, vacuum-tight cylinder connecting the dewar well to the ambient environment. This neck tube has the thermal stop rings [4] which are the thermal/mechanical interface to the probe and allow the probe to be vapor cooled, improving lifetime performance. The insulation system of alternating layers of double aluminized Mylar and up to three silk net spacers was selected based on subscale test results showing this construction to have the minimum heat leak compared to other spacers, such as polyester. Dewar mass was minimized by a highly optimized structural design on the main tank and vacuum shell and the use of honeycomb vapor cooled shields [5]. The vacuum structures were optimized for both internal and
372
D.C. Read et al. / Cryogenics 39 (1999) 369–379
external pressure requirements to achieve a minimum mass which resulted in waffled domes and ring stiffened cylindrical sections, shown in Fig. 4. Honeycomb vaporcooled shields (Fig. 5) have a 38% lower mass with equivalent stiffness to the more traditional monocoque vapor cooled shield design. Table 1 shows the mass and mass savings associated with these components. The alignment and stability of the externally mounted star trackers relative to the science instrument telescope is critical for mission success. The dewar must provide a mounting platform for the star trackers that is referenced to the telescope and is stable to less than 10 arcs for the orbital thermal environment where the dewar vacuum shell temperature is 230 ⫾ 5 K. The normal number of six PODS supports was increased to 12 to meet this requirement (Fig. 6). The six forward supports are located in a plane and are attached to a graphite/epoxy ring at the warm end of the struts. The graphite/epoxy ring is attached to the 2219 aluminum vacuum shell with titanium flexures. As the tank cools down, the thermal loads put into the struts are low with this arrangement. In orbit, changes in vacuum-shell dimension due to temperature changes also put minimal thermal strains on the struts because of the ultralow expansion characteristics of the graphite/epoxy support ring. This graphite ring is used to mount the star trackers, providing a very stable mounting surface that is referenced to the science instrument telescope.
3. Unique features The porous plug is unique in that it must operate over a large range of flowrates, since the helium boil-off is used for attitude control of the spacecraft. The primary requirements that affected the design of the porous plug were to operate over a flowrate range of 4–16 mg/s without choking while maintaining a constant bath temperature of 1.8 K and to operate with sufficient back pressure
Fig. 4.
Light-weighted aft vacuum shell.
so that the vented gas can be used for spacecraft attitude control. The nominal flowrate during the science phase of the mission is 6.8 mg/s, which is sufficient for attitude control. Higher flowrates are needed earlier in the mission to perform attitude control. During most operations, the dewar is vented at its nominal flowrate, and a portion of the flow is “nulldumped” if not needed for attitude control. If additional thrust is needed, the downstream pressure is decreased to increase the vent rate. The bath temperature is maintained during these periods of nonequilibrium flowrate by turning on a heater in the helium bath. As flowrate is increased by decreasing the back pressure, the temperature gradient across the plug increases until a choked flow condition is achieved and the flowrate vs. temperature drop is no longer linear. The flowrate at the end of this linear range of operation is referred to as the critical flowrate. As the flowrate is increased and subsequently decreased, the temperature gradient goes through a hysteresis loop. Fig. 7 shows typical temperature gradient and back-pressure data from the acceptance testing of the porous plug. This shows that the slope of the flowrate versus temperature drop in the linear region is decreasing with lower bath temperatures and the critical flowrate is also decreasing. A unique design was used to facilitate safe venting of the helium during a catastrophic loss-of-vacuum. If the guard vacuum is compromised, the heat rates to the helium tank are very high due to the condensation of air on the tank. These high heat rates cause the helium tank pressure to increase until a burst disc ruptures to relieve the tank pressure. For GP-B, a 5-cm-diameter orifice is required to vent the helium and ensure that the pressure does not exceed the design pressure of 0.26 Mpa. If the helium is vented into the vacuum space, the vacuum shell internal pressure will increase until the burst discs on the vacuum shell rupture and the helium is vented into the ambient environment. This internal pressure can become large, depending on the assumptions about the warm-up (and hence the volume increase) of the gas in the vacuum shell. A conventional vent line with a 5-cm diameter would have an unacceptable increase in parasitic heat leak to the tank. A flexible vent line was developed to safely vent the GP-B dewar in an emergency situation. The vent line is constructed with laminated Dacron fiber/Mylar sheets that minimize heat conduction into the dewar and can withstand 1.1 MPa internal pressure. The vent line is installed in the dewar in a folded configuration (Fig. 8) to eliminate radiation tunneling down the vent line. If there is an emergency vent situation, the vent line expands to 5 cm in diameter and vents the helium outside the dewar without increasing the vacuum shell pressure. An inflated flexible vent line immersed in liquid nitrogen during development testing is shown in Fig. 9. Another advantage of the flexible vent line is
D.C. Read et al. / Cryogenics 39 (1999) 369–379
Fig. 5.
373
(a) Cone section of first vapor cooled shield. (b) First (of four) vapor cooled shields.
that refurbishment of the dewar after a loss-of-vacuum will be easier, since the multilayer insulation (MLI) and vapor-cooled shields will not be damaged. Venting directly into the vacuum space would require a rebuild of the thermal protection system.
4. Performance summary
Acceptance test results have been reported previously [6]. The critical performance parameters of mass, para-
374
D.C. Read et al. / Cryogenics 39 (1999) 369–379
Table 1 Weight savings Assembly
Flight weight (kg)
Main tank (including a 51-kg proton shield) Guard tank Vapor cooled shields Vacuum shell
Fig. 6.
Savings (kg)
262 15 72 345
80 4 44 100
Comment Waffle domes, ring-stiffened cylinder, 2219 aluminum Lightening holes in support cone Perforated aluminum honeycomb shields Ring stiffened and waffle construction, 2219 aluminum
Main tank with cold plumbing, guard tank, composite neck tube, PODS-V supports, graphite rings, and support ring.
sitic heat leak, and ground hold will be summarized and compared with previous flight cryostats. The dewar, dry, weighed 810 kg. The measured net parasitic heat rate to the 4.2 K main tank was 136 mW (compared to a predicted value of 112–136 mW), with a vacuum shell temperature of 294 K. The measured value falls within the 30% margin added to the predicted value. This margin has been used throughout the program. The 1.8-K nonvented main tank temperature rose at a rate of 0.002 K/day with the guard tank filled. The 4.2-K guard tank boiled off at a rate of 5%/day (probably too low a value because thermal equilibrium was not achieved for this test). All these test results meet the requirements for weight, lifetime, and launch pad servicing.
4.1. Performance comparison of GP-B to previous flight cryostats Performance comparisons were made in three major areas based on the acceptance tests data. Fig. 10 shows the 30% weight savings achieved compared to the similarly sized ISO cryostat. Even larger savings are achieved if you eliminate the 51-kg proton shield dead weight. The thermal performance improvement of the GP-B cryostat with the SIA and probe installed (Table 2) is due to the use of passive orbital disconnect struts (PODS-V), the carefully tailored double-aluminized Mylar radiation shields (149 each), the silk net low-conductance spacers (three each at the tank reduced progressively to one each
D.C. Read et al. / Cryogenics 39 (1999) 369–379
Fig. 7.
Fig. 8.
375
Effect of bath temperature on temperature drop and back pressure.
Cryogen tank emergency vent line is shown in both nonventing and venting configurations.
at the vacuum shell) plus the use of four vapor-cooled shields. The remarkable ground-hold performance of keeping the GP-B main tank nonvented for 90 days (Table 3) is due to the guard tank driving the inner vapor-cooled shield to 4.2 K (top) to 5 K (bottom); the main tank tem-
perature rises from 1.6 to 1.85 K in this time period. The He I guard tank is filled once a week and no helium is lost from the main tank. Both IRAS and COBE pumped on the main tank up to 26 h before launch. Consequently, they had less helium for use in orbit.
376
D.C. Read et al. / Cryogenics 39 (1999) 369–379
Fig. 9.
Fig. 10.
Development tests of the flexible emergency vent line verified design requirements were satisfied.
Dry weight/He II weight ratio comparison of dewars.
Table 2 Thermal performance comparison of He II flight cryostats with science instruments turned off (TH ⫽ 294 K) Flight payload
Tank volume (l)
COBE
660
ISOa IRAS GP-Ba,b
2300 550 2441
% Boil-off per Normalized day boil-off rate 1.25 0.68c 0.83 0.83d 0.27
4.6 2.5 3.1 3.1 1.0
a
Guard tank not used. b Dewar boiloff measured. Probe/SIA heat leak calculated. 30% margin. c Aperture cover actively cooled to 5 K with separate He I flow. d Aperture cover bottle filled with supercritical helium at 5 K.
5. GP-B integrated system test program Early in the GP-B program, it was recognized that the unconventional requirements imposed by the supercon-
ducting shield would require significant development effort. Integration of a warm cryostat probe with a cold dewar while maintaining a lead shield in the superconducting state had never been attempted on this scale before. Of particular concern was the fact that the lead shield has a radial clearance of only 4 mm from the outer shell of the probe. As a result of the need to develop this technique as well the need to resolve a number of other integrated-system development issues, one development dewar (the engineering development dewar, EDD), and two development probes (Probes A and B) were built. This hardware has been used in a series of three full-scale integrated-system tests. We are now preparing for a fourth and final test (GTU-2) prior to the integration of the flight payload. GTU-2 will involve the integration of Probe-B in the SMD. Key objectives for GTU-2 are to verify integration hardware and procedures with the SMD, to verify the performance and compatibility of the SMD with Probe-B, and to resolve open issues remaining from previous tests. Upon completion of GTU-2 near the end of 1997, Probe-B will be removed and the flight probe, Probe-C will be installed in preparation for final payload tests. After completion of the payload test, the payload will be shipped to Lockheed Martin for integration and testing with the spacecraft. 6. Installation and verification of the superconducting shield Installation of the superconducting lead-foil shield needed to provide the ultralow magnetic field environment required by the GP-B experiment is a lengthy iterative process. A folded cylindrical lead foil shield which is closed at one end of the cylinder is mounted from a cloth sleeve and installed into a vacuum-tight glass cooling tube. The cooling tube with the new folded lead
D.C. Read et al. / Cryogenics 39 (1999) 369–379
377
Table 3 Cryostats ground hold performance at the launch pad Cryostat
Helium loaded
Percentage ullage
(kg) IRAS COBE ISO GP-Bb
72 89 331 338
7 7 1 5 max to 2 min
Guard tank
Initial orbital vent rate (compared to orbit equilibrium value)
Used?
Fill cycle (days)
No No Yesa Yesc
– – 4 7
Main tank nonvented time
(days) Above Above Above Below
1.1 2 to 3 Up to 4.2 Up to 90
Mounted on main tank. Guard tank vapor cools inner vapor cooled shield to ⫺ 30 K. Based on dewar acceptance test data. Includes calculated probe heat leak with 30% margin. c Guard tank, inner vapor cooled shield run at 4.2–5 K. a
b
shield is then installed in the dewar well inside of a previously installed shield. The new shield can then be warmed and cooled through the superconducting transition temperature of lead (7.2 K) by controlling the partial pressure of helium exchange gas in the cooling tube. Once the new shield is cooled in this manner, it will essentially pin the ambient magnetic flux. The previous shield is now removed from the dewar, the bottom of the cooling tube is opened, and the cooling tube is then removed, leaving the new shield hanging in the dewar well. The new shield is mechanically unfolded using an expansion tool so that the shield assumes its cylindrical shape. Since the magnetic flux is pinned in the lead foil, it now becomes spread over a much larger volume thereby reducing the magnitude of the field. The magnitude of the field can be measured by means of a SQUIDbased magnetometer. Field reductions by a factor as large as 100 can be achieved by this technique, and it can be repeated as many times as necessary to achieve our requirement (10 pT). This process is depicted schematically in Fig. 11. The shield that was recently installed in the science mission dewar was obtained after four iterations. After the internal field is determined to be acceptable in the final shield, it is detached from its cloth sleeve and left in the bottom of the dewar well (see Fig. 12). Measurements of the field in the flight lead shield after completion of installation have indicated that the radial component of the field is ⱕ 5 pT (0.05 G) in the region that will be occupied by the science gyroscopes. In addition, the axial component, which in the absence of physical defects in the shield will be much smaller than the radial components, has been measured to be constant to within 5 pT over the same region. These results are consistent with the field magnitude being well below the requirement of 10 pT. Future measurements will also be made to verify that the shield meets its ac field attenuation requirements. Before a probe can be integrated into the dewar, a
Fig. 11. Lead Shield (LS) field reduction technique: (a) previous LS positioned in dewar well; (b) next LS enclosed in cooling tube under controlled helium atmosphere and installed inside the previous LS; (c) cooling tube diaphragm cut out; (d) removal of previous LS; (e) removal of cooling tube; (f) LS ready for expanding; (a) LS expanded.
device called a lead bag retainer is installed inside the shield (Fig. 13). This device consists of a number of linear leaf springs (much like Venetian blinds) made of titanium copper alloy which hang inside the shield. It serves to protect the foil shield and to center the probe during probe insertion, and it also insures good thermal contact between the shield and the cold well wall while providing thermal isolation to the probe shell. In addition to the leaf springs, there is a basket-shaped structure to support the bottom of the shield during launch accelerations.
7. Conclusions The SMD acceptance test results show that the dewar meets all relativity mission requirements. The use of
378
D.C. Read et al. / Cryogenics 39 (1999) 369–379
Fig. 12. View of the interior of the dewar well with the flight lead shield installed. Also visible are the composite necktube with thermal stop rings and the primary probe-dewar thermal interface at Station 200 (the intermediate annular region). Station 200 is 0.87 m from the top of the dewar and the portion of the well below Station 200 is 1.85 m in length.
Fig. 13. View of the interior of the dewar after installation of a lead bag retainer. The retainer shown here is one used for testing purposes. The flight retainer will be very similar except it will have modifications; to protect the shield from being heated above its superconducting transition temperature during launch.
D.C. Read et al. / Cryogenics 39 (1999) 369–379
advanced technology in weight reduction techniques, PODS-V supports, multilayer insulation, vapor-cooled shields, and guard tank design make this dewar the lightest, per unit mass of He II stored, the most thermally efficient, and the easiest to service at the launch site (compared to the three previous free flight cryostats that have been launched, IRAS, COBE, and ISO). The magnetic field achieved meets the program requirements for the SIA.
Acknowledgements This work was performed for Stanford University under subcontract PR8709 and funded by the NASA Marshall Space Flight Center via contract NAS 8-39225.
379
References [1] Taber MA, et al. Production of ultralow magnetic field for Gravity Probe B (GP-B). In: Advances in cryogenic engineering, vol. 39. New York: Plenum Press, 1994. [2] Parmley RT, et al. Installing scientific instruments into a cold Lhe dewar—the Gravity Probe-B approach. In: Advances in cryogenic engineering, vol. 35. New York: Plenum Press, 1990. [3] Spradley IE, Parmley RT. Design and test of a modified passive orbital disconnect strut (PODS-IV). In: Advances in cryogenic engineering, vol. 33. New York: Plenum Press, 1988. [4] Parmley RT. Unique cryogenic features of the Gravity Probe B (GP-B) experiment. In: Advances in cryogenic engineering, vol. 33. New York: Plenum Press, 1988. [5] Abrahamson TE, et al. Weight reduction approaches for helium flight dewars. In: Advances in cryogenic engineering, vol. 39. New York: Plenum Press, 1994. [6] Parmley RT, et al. Performance of the relativity mission superfluid helium flight dewar. Presented at the COSPAR Conference, Birmingham, UK, 1996.