3LA02 1 FERMILAB-CONF-04-273-TD R&D of Nb3Sn Accelerator Magnets at Fermilab A.V. Zlobin, G. Ambrosio, N. Andreev, E. Barzi, B. Bordini, R. Bossert, R. Carcagno, D.R. Chichili, J. DiMarco, L. Elementi, S. Feher, V.S. Kashikhin, V.V. Kashikhin, R. Kephart, M. Lamm, P.J. Limon, I. Novitski, D. Orris, Yu. Pischalnikov, P. Schlabach, R. Stanek, J. Strait, C. Sylvester, M. Tartaglia, J.C. Tompkins, D. Turrioni, G. Velev, R. Yamada, V. Yarba Abstract—Fermilab is developing and investigating different high-field magnet designs for present and future accelerators. The magnet R&D program was focused on the 10-12 T accelerator magnets based on Nb3Sn superconductor and explored both basic magnet technologies for brittle superconductors - wind-and-react and react-and-wind. Magnet design studies in support of LHC upgrades and VLHC are being performed. A series of 1-m long single-bore models of cos-theta Nb3Sn dipoles based on wind-andreact technique was fabricated and tested. Three 1-m long flat racetracks and the common coil dipole model, based on a singlelayer coil and wide reacted Nb3Sn cable, have also been fabricated and tested. Extensive theoretical studies of magnetic instabilities in Nb3Sn strands, cable and magnet were performed which led to successful 10 T dipole model. This paper presents the details of the Fermilab’s high field accelerator magnet program, reports its status and major results, and formulates the program next steps. Index Terms—Superconducting accelerator magnets, dipoles, quadrupoles, Nb3Sn strands and cables I. INTRODUCTION T he program goal is the development of new generation SC accelerator magnets with high operation fields and large operation margins for different applications. A short list of possible applications includes SC magnets for the Tevatron, particularly to replace some present dipoles in order to provide space for special devices or to replace existing low-β quadrupoles or to create a new interaction region (IR), 2nd generation LHC IR dipoles and quadrupoles for luminosity upgrade [1], SC magnets for a future Very Large Hadron Collider (VLHC) [2], etc. Similar (complimentary) programs are being performed in U.S. at BNL, LBNL, TAMU [3] and in Europe at CERN, Twente University, Saclay, RAL, etc. [4]. In many cases magnet requirements for upgrading existing and future machines push accelerator magnet technology to limits exceeding the present level based on NbTi superconductor. Our present SC Magnet R&D program is focused on accelerator magnets based on Nb3Sn superconductor and explores two basic technologies - windand-react and react-and-wind. Fermilab has all necessary equipment and infrastructure to perform short and full-scale model magnet R&D programs. Manuscript received October 4, 2004. This work was supported by the U.S. Department of Energy. Authors are with the Fermi National Accelerator Laboratory (Fermilab), P.O. Box 500, Batavia, IL 60510 USA (phone: 630-840-8192; fax: 630-8403369; e-mail: zlobin@fnal.gov). The fabrication and test infrastructures include cable fabrication and insulation machines, 2-m long winding tables and 15-m long automatic winding machine, 1-m long coil HT oven, 6-m long epoxy impregnation facility, collaring/yoking presses, magnet test facilities with vertical (up to 4-m long) and horizontal (up to 15 m) cryostats with 1.8-4.5 K operation temperature, 30 kA power supply. The development of new generation accelerator magnets requires advanced superconductors, structural materials and components. Fermilab is participating in a national programs sponsored by DOE promoting the development of advanced superconductors and magnet structural materials in U.S. industry. Fermilab also has developed the adequate infrastructure to perform extensive superconductor, cable and material R&D in support of the magnet R&D program. Studies include strand and cable characterization, optimization of HT, cable development, etc. The review of our main activities and achievements in this area are summarized in [5]. This paper summarizes the main results of Fermilab’s high field magnet program. II. NB3SN ACCELERATOR MAGNET R&D Magnet R&D program includes magnet design studies, material and technology development, and model magnet R&D. All three activities are tightly connected, crossfertilizing each other. A. Magnet Design Studies At the present time we are investigating two types of highfield dipole designs for VLHC. One design is based on shelltype coils. This design approach is implemented in almost all NbTi SC magnets used in present high-energy accelerators. The other design is based on flat block-type coils arranged in the common-coil configuration. In this design the coil radii are set by the aperture separation, not the aperture size, and hence, conductor bends are relatively gentle and friendly to brittle conductors such as A15 or HTS. Based on these basic design approaches we have developed several innovative dipole magnets for VLHC [2]. Fermilab participates in the U.S. LHC Accelerator Research Program (LARP) [6]. One of the LARP goals is to develop 2nd generation IR magnets for LHC to replace the 1st generation magnets which have limited lifetime and will be one of the machine limiting systems. Contributions of Fermilab to LARP Magnet R&D include conceptual designs studies of various magnet types for 2nd generation LHC IRs. We performed 3LA02 B. Technology Development We develop accelerator magnets based on Nb3Sn superconductor. Critical parameters of Nb3Sn (Bc2=27-28T, Tc=18K,and Jc(12T,4.2K)~2.5-3 kA/mm2) are much higher than NbTi parameters. High-performance Nb3Sn strands are commercially available in long lengths at affordable price. Since Nb3Sn is brittle it requires new magnet technologies and materials for accelerator magnets based on brittle superconductors. Two basic approaches are used at present time, Wind-and-React and React-and-Wind, and both have been studied at Fermilab. Wind-and-react techniques impose demanding requirements on the magnet insulation which must withstand a long hightemperature heat treatment under compression. Ceramic or S2glass insulation with a liquid ceramic binder or pre-preg, which meets these requirements, are being studied and optimized at Fermilab in collaboration with industry [10]. Complicated end parts, used in traditional cos-theta coils, in case of wind-and-react techniques have to withstand the heat treatment and match the cable shape in the ends to avoid shorts. An optimization method for metallic end parts was developed and successfully used at Fermilab [11] together with the rapid prototyping technique reducing the time and the cost of end part development processes. Water jet machining was first time used for end part fabrication resulting in the reduction of their costs by a factor of 3 and manufacturing time by a factor of 10 while providing acceptable part quality. Traditionally Nb3Sn coils are impregnated with epoxy to improve their mechanical and electrical properties. However, the radiation limit for epoxy is quite low which reduces the lifetime of the magnet. Various commercially available polyimide solutions are being investigated to replace epoxy as an impregnation material for Nb3Sn coils [12]. These studies will be continued on practice coils and in model magnets. C. W&R Cos-theta Dipole Models The goal of this work is to develop 11-12 T Nb3Sn accelerator quality magnets based on the W&R technique. The main design features of our cos-theta Nb3Sn dipole models (HFDA) are 28 strand cable based on high-Jc 1-mm Nb3Sn strands, 2-layer coil with 43.5-mm diameter bore and cold, vertically-split iron yoke [13]. The maximum design field is 12 T at 4.5 K. A 3D view of the HFDA dipole and a cross-section of 28-strand cable are shown in Fig.1. This design rests on the designs of the first 50 mm Nb3Sn dipole models developed in 1990s: 10 T dipole model (CERN/ELIN, 1989) [14]; 11 T MSUT (Twente University, 1995) [15] and 13 T D20 (LBNL, 1997) [16]. From the beginning the focus was on the practical and inexpensive design, and robust technology acceptable in the future for industrialization of full-scale magnets. Significant coil area reduction has been achieved in this design with respect to previous magnets with similar design fields. The coil crosssection area is smaller than the one in MSUT by a factor of 2 and is smaller than the D20 coil cross-section by a factor of 3. Fig.1. HFDA dipole and 28-strand cable developed and fabricated at Fermilab. New features have been implemented in the fabrication process to simplify the technology, to reduce the fabrication cost and time, and to make the design robust: • thick ceramic or S2-glass cable insulation impregnated with liquid ceramic binder or ceramic pre-preg • half-coil curing at 120C after winding before reaction to obtain a solid coil structure • simultaneous reaction and impregnation of two half-coils This technology allowed coil handling and size control after curing, eliminated mid-plane shimming during assembly, avoided expensive collars and delicate collaring procedure. To study and optimize the quench performance issues for cos-theta magnets we adjusted the design for using half-coils with a magnetic mirror (HFDM) [17]. The main advantages of this approach are the same mechanical structure and assembly procedures, possibility of advanced coil instrumentation, the shorter turnaround time and the lower cost. 0.6 0.5 0.4 Iq/Iss studies of 90-mm IRQ based on shell-type coil geometry [7], aperture limitation studies for LHC IRQ [8], comparison of large-aperture quadrupoles based on shell-type and racetrack coils [9]. We will also participate in short and long model magnet R&D as well as in the design, fabrication and tests of full-scale prototypes of the LHC IR magnets. 2 hfda02 hfda03 hfda04 hfdm02 0.3 0.2 0.1 0.0 0 10 20 30 Quench number Fig.2. Quench performance summary of the first Fermilab’s cos-theta models. HFDA02-04 are dipoles models and HFDM02 is mirror configuration. Three 1-m long dipole models (HFDA02-04) and two magnetic mirror configurations (HFDM01-02) were fabricated and tested at Fermilab during 2001-2003. We were able to achieve good, well-understood field quality including geometrical harmonics and coil magnetization effects [18]. The use of Nb3Sn conductor typically results in significant coil magnetization effects in high field magnets due to large effective filament diameters. We developed and tested a 3LA02 D. R&W Technology In parallel with W&R technology we have been working on the alternative R&W approach. The goals of this work were to study the possibilities and limitations of the R&W technique for Nb3Sn magnets and develop a 10 T accelerator quality common coil dipole based on this approach. An advantage of R&W is the possibility to use the traditional SC magnet technologies well established for NbTi accelerator magnets during past 30 years. The sensitivity of brittle Nb3Sn to stress and strain impose serious restrictions on magnet design. After comprehensive design studies [22] we developed common coil dipole (HFDC) which meets accelerator requirements and tolerant to the R&W technology. The main design features of our common coil magnet are wide 60-strand cable based on high-Jc 0.7 mm Nb3Sn strand, single-layer coil confined in an innovative mechanical structure with two bores of 40 mm and surrounded with a cold iron yoke. The maximum design field is 10 T at 4.5 K. A 3D view of the HFDC dipole and a cross-section of 60strand cable are shown in Fig.3. were fabricated and tested during 2001-2003 [22,24]. All magnets survived the complicate fabrication process and reached 60-75% of the expected Short Sample Limit (see Fig.4). In the common coil magnet we also reached good, well understood field quality [25]. However, the critical current degradation of reacted cable was much larger than expected. 0.8 0.7 0.6 Iq/Iss simple and effective passive correction system to correct large coil magnetization effect in Nb3Sn accelerator magnets [19]. However, the maximum quench current in all those magnets was only 50-60% of expected short sample limit (Bmax~6-7 T) [20, 21] (see Fig.2). Detailed studies of possible causes of the limited magnet quench performance allowed correlating it with the properties of superconductor used and the magnet design parameters. 3 0.5 HFDB01 HFDB02 HFDB03 HFDC01 0.4 0.3 0.2 0.1 0 20 40 60 80 100 Quench number Fig.4. Quench performance summary of the Fermilab’s Nb3Sn common coil models based on R&W approach. HFDB01-03 are simple racetrack models and HFDC01 is the dipole models. The works on R&W approach demonstrated that in order to realize the potential of this technology and use it in accelerator magnets, significant further efforts are required including conductor and magnet mechanics improvement. Due to the limited resources we have focused since last year on W&R cos-theta magnets since this design approach and technology are the baseline for the 2nd generation LHC IR magnets [6]. III. QUENCH PERFORMANCE STUDIES AND OPTIMIZATION To address the quench performance problem in our model magnets, thorough analysis of the experimental data including quench origin and quench propagation velocity, critical current and critical temperature measurements in the magnet coil, was done which initiated the instability studies in superconducting strands and cables used in those magnets. Fig. 3. FNAL Common coil dipole and 60-strand cable fabricated at LBNL. The design of the common coil magnet required the development of new technologies, in particular, simultaneous winding of both coils into support structure using brittle wide cable, impregnation of whole collared coil, etc. The optimization of different R&W techniques was performed using 1-m long racetrack coils (HFDB) made of sub-sized 41strand cable made of the same 0.7 mm strand [23]. Simple coil geometry based on two flat racetrack coils separated by 5 mm gap and bolted, solid stainless steel mechanical structure allowed reducing the cost and the turnaround time. Three 1-m long racetracks (HFDB01-03) and one common coil dipole model (HFDC01) based on reacted Nb3Sn cable A. Strand Stability Studies The calculation performed using the developed adiabatic flux jump model [26] showed significant reduction of strand current carrying capability at low fields (see Fig.5). The key parameters of the model are the strand Jc and the effective filament size. For strands with large deff and high Jc the maximum transport current, that strand can carry at low fields, becomes smaller than the transport current at high fields. Due to the non-uniform field distribution in the magnet, different turns in the coil at the same current are exposed to different magnetic fields from 0 to Bmax. The two most important magnet load lines are shown in Fig.5. The first one corresponds to the point in the coil with the maximum field. This line is usually used for the calculation of magnet short sample limit. The second one is the load line which cross the minimum on the Iq(B) curve. Depending on the relative value of Iq at low and high field, the magnet could be high-field or low-field limited. This model allows explaining the limited quench performance of our first models. 3LA02 3000 35000 Ic(B) Iq(B) Bmax(I) Blf(I) 25000 Current, A 2000 30000 1500 1000 20000 15000 Ic (B,4.33K) PIT-28-1.0_CERN 10000 500 5000 0 0 0 2 4 6 8 Magnetic field, T 10 The model stimulated measuring the strand critical current both at high and low fields at Fermilab [27] and elsewhere [2829]. At low fields the dependence of critical current vs. field was measured using the traditional V-I method at constant external field and the V-H method at constant transport current in the sample. Both methods showed low current quenches at low fields (see Fig.6). These results are consistent with the instability model. The model needs to be improved to achieve better quantitative agreement between calculations and the experimental data obtained using different methods. 2000 P S LIMIT 1500 1000 V-I measurements <=V-H measurement s 500 0 2 4 6 MJR-28-1.0_FNAL PIT-28-1.0_FNAL 2 10 12 6 8 10 12 High efficiency of tests at Fermilab using the SC transformer allowed studying different cable made of different strands, impregnated and non-impregnated with epoxy, with different value of RRR [30]. The effect of RRR on average quench currents for impregnated and non-impregnated Nb3Sn cables made of MJR, RRP and PIT strands measured in self-field at 4.2 K is shown in Fig.11. These data and also measurements performed on strands [27-29] show the importance of strand RRR for the cable stability at low fields. 30000 25000 20000 Non-impregnated MJR 15000 Impregnated MJR Non-impregnated KS PIT 10000 Impregnated RC PIT 5000 Non-impregnated RRP Impregnated RRP 0 20 TRANS ITIO NS 8 4 Fig.7. 28 strand MJR and PIT cable samples tested at BNL, CERN and Fermilab at 4.3 K. 0 Q UE NCHES MJR-20-1.0_BNL Magnetic field, T Fig.5. Strand critical and quench current, calculated using adiabatic flux jump model developed at Fermilab. The dashed lines represent two most important magnet load lines [26]. 0 MJR-28-1.0_CERN 0 12 Average quench current, A Current, A 2500 Current, A 4 14 40 60 80 100 RRR 16 Magnetic field, T Fig.6. Reduction of strand current carrying capability observed in V-I and V-H measurement on 1 mm MJR strands at Fermilab [27]. B. Cable Short Sample Tests The instability studies were performed also on cable short samples. The cable samples were tested at Fermilab using 28 kA SC transformer in self-field at temperatures T=2-4.3K [30], as well as at BNL in external magnetic fields up to 7 T at 4.3K, and at CERN in external fields up to10 T and temperatures 1.8-4.2 K [31]. The results of testing of 28-strand MJR and PIT cable samples at BNL, CERN and Fermilab at 4.3 K are summarized in Fig.7. One can see excellent correlation of experimental data for similar samples tested at three different test facilities. The results are also consistent with Fermilab’s instability calculations and HFDA02-04 and HFDM01-02 quench performance [21]. Fig. 11. Effect of RRR on average quench currents for impregnated and nonimpregnated Nb3Sn cables made of MJR, RRP and PIT strands measured in self-field at 4.2 K [30]. Interstrand contact resistance (ICR) plays an important role in cable stability. We developed technique which allowed ICR measurements on samples extracted from magnets [32]. The first results of these measurements in turns with different position in the coil are shown in Table I. As it can be seen the adjacent resistance is uniform and quite low. It certainly helps for current sharing in the cable. The crossover resistance varies in large range. To make it more uniform and reduce its effect on AC losses and field dynamic effects, a SS core can be used. Table I. ICR measurement on HFDA04 coil samples. Position Midplane Pole Coil Layer Inner Outer Inner Outer Ra (µΩ) 2.3-3 2.5-3 0.8-1.9 4.25 Rc (µΩ) ≥ 500 20-30 4-6 4.38 3LA02 C. Cable testing using Small Racetracks The Nb3Sn cables are being tested at Fermilab using compact coil systems that implement main features of Nb3Sn W&R technology and real cable operation conditions. We use the simple and reliable mechanical structure developed at LBNL [33] and 2-layer coil wound from single piece of cable into common coil configuration. Fermilab’s small racetrack coil and small racetrack assembly are shown in Fig.12. The details of racetrack design and fabrication procedures, and test results are reported in [34]. 5 After a few quenches at 2.2 K magnet current increased to 22 kA. Then magnet quenched at 4.5 K at 21 kA. This result as well as the ramp rate study at 4.5 K confirms that the magnet has reached its short sample limit at 4.5 K. The maximum field reached in the coil was ~10 T. Fig.14. Mirror magnet HFDM03 (non-lead end, before installation of second half-shell of bolted skin). 25000 Two cable types, used in Fermilab’s cos-theta dipole models, were tested in two small racetracks SR01 and SR02. These cables were made of PIT and MJR strands 1.0 mm in diameter. Both strands had high critical current density and effective filament sizes of 0.05 and 0.11 mm respectively. Quench performance of both racetracks is shown in Fig.13. SR01, made of PIT strand, reached its high field short sample limit at 4.5K. Racetrack SR02, made of MJR strand, was limited by quenches in the low field regions near the coil lead splices. Both results were expected based on instability analysis and strand and cable measurements. Quench current, A 30000 Quench current, A 4.5 K Fig. 12. Fermilab small racetrack coil (left) assembled with LBNL-type mechanical structure. 20000 20 A/s 300 A/s 175 A/s 150 A/s 200 A/s 100 A/s 75 A/s 50 A/s Iq_max @4.5 K 15000 10000 5000 2.2 K 4.5 K 0 0 10 20 30 40 50 Quench number Fig. 15. Mirror magnet HFDM03 quench history. IV. 10 T DIPOLE MODEL HFDA05 4.5K 25000 2.2K Recently we fabricated and tested a new cos-theta dipole model (HFDA05) made of Nb3Sn PIT cable (see Fig.16). 4.5K 20000 4.5K 15000 2.2K 10000 SR02 5000 SR01 0 0 10 20 30 40 Quench number 50 60 Fig.13. Quench performance of small racetrack magnets SR01 (a) and SR02 (b). Only quenches at low ramp rate of 20 A/s are plotted. D. Magnet quench performance improvement The stability analysis as well as experimental data obtained on strands, cables and small racetrack confirmed acceptable stability of PIT Nb3Sn strand. This strand was used in our last cos-theta dipole model. To test and optimize PIT coil quench performance, the first PIT coil was assembled and tested in magnetic mirror configuration (HFDM03) shown in Fig.14. The details on magnet fabrication and tests results are reported in [35]. Quench history of this magnet is presented in Fig.15. After 25 quenches at 4.5 K HFDM03 reached a plateau at 21 kA. Fig.16. Dipole model HFDA05 (lead end, before installation of end plate and half-coil splicing). The design of this magnet is similar to other HFDA models [20]. One half-coil of this magnet was previously tested in mirror configuration (HFDM03) described above. This coil was then extracted from the mirror and was assembled with a new PIT half-coil in dipole model HFDA05. The details of magnet design and fabrication procedure as well as the test results are reported in [35]. Quench history of this magnet is shown in Fig.17. After 23 quenches at 4.5 K the magnet reached a stable plateau at 16.8 3LA02 kA. After a few quenches at 2.2 K the magnet current increased to 17.9 kA. When the magnet was excited again at 4.5 K, it quenched at the same current of 16.8 kA. After a thermal cycle (TC) to room temperature the magnet showed short re-training with the first quench only 3% below the short sample limit. The training data show that the magnet has reached its short sample limit at 4.5 K. The maximum field in the bore at 4.5 K was 9.5 T and at 2.2 K (in TCII) was 10.1 T. [6] [7] [8] [9] [10] 20000 TCI Quench current, A [5] 15000 TCII 4.5 K 20 A/s 300 A/s 125 A/s 150 A/s 200 A/s 100 A/s 75 A/s 50 A/s 10 A/s Iq_max @4.5 K 10000 5000 [11] [12] [13] 2.2 K 4.5 K [14] [15] 0 [16] 0 20 40 Quench number 60 [17] Fig. 17. Dipole model HFDM05 quench history. [18] Successful fabrication and tests of the two cos-theta magnets based on PIT Nb3Sn strands have proven an importance of the conductor stability for magnet quench performance predicted by the stability model. The magnetic field of 10 T has been reached in the Fermilab’s dipole design. The mechanical structure and technology developed for these magnets demonstrated reliable performance at fields up to 10 T. [19] [20] [21] [22] V. CONCLUSION Fermilab has sound SC accelerator magnet R&D program. Reaching of 10 T field level in short dipole models is an important milestone towards its goal of 10-12 T accelerator dipole magnet for future hadron collider and of 15 T largeaperture quadrupoles for the LHC luminosity upgrade. The studies of magnetic instabilities in Nb3Sn strands, cable and magnets, performed at Fermilab, provide a key contribution to the field of applied superconductivity and accelerator magnet technologies and will fuel further theoretical and experimental works in these fields. [23] [24] [25] [26] [27] [28] ACKNOWLEDGMENT The authors thank Fermilab’s Technical Division staff for their contribution to this effort and our colleagues at BNL, LBNL and CERN for collaboration in instability studies. REFERENCES [1] [2] [3] [4] O. Brüning, et al., LHC Luminosity and Energy Upgrade: A Feasibility Study, LHC Project Report 626, December 2002. P. Limon et al., “Design study for a staged Very Large Hadron Collider”, Fermilab-TM-2149, June 4, 2001. S. A. Gourlay, “High Field Magnet R&D in the USA”, IEEE Trans. on Applied Superconductivity, vol. 14, No.2, June 2004, p.333 A. Devred et al., “High Field Accelerator Magnet R&D in Europe”, IEEE Trans. on Applied Superconductivity, vol. 14, No.2, June 2004, p.339. View publication stats [29] [30] [31] [32] [33] [34] [35] 6 E. Barzi, Superconductor and Cable R&D for High Field Accelerator Magnets at Fermilab, Fermilab internal note, TD-04-024, June 2004. R. Kephart et al., The U.S. LHC Accelerator Research Program: A Proposal, .http://www-td.fnal.gov/LHC/USLARP.html A.V. Zlobin et al., “Conceptual Design Study of Nb3Sn Low-beta Quadrupoles for 2nd Generation LHC IRs”, IEEE Trans. on Applied Superconductivity, v. 13, No. 2, June 2003, p.1266. A.V. Zlobin et al., “Aperture limitations for 2nd generation Nb3Sn LHC IR quadrupoles”, PAC2003, Portland (OR), 2003. V.V. Kashikhin et al.,“2nd LHC IR Quadrupoles Based on Nb3Sn Racetrack Coils”, EPAC 2004, Lucerne, July 5-8 2004. D.R. Chichili et al., “Fabrication of Nb3Sn Shell-Type Coils with PrePreg Ceramic Insulation”, CEC/ICMC2003, September 22-25 2003. S. Yadav et al.,“Coil Design Issues for the High Field Dipole at Fermilab”, IEEE Trans. on Appl. Superc., v. 11, No. 1, 2001, p. 2284. D. R. Chichili et al., “Investigation of Alternative Materials for Impregnation of Nb3Sn Magnets”, IEEE Trans. on Applied Superconductivity, v. 13, No. 2, June 2003, p.1792. G. Ambrosio et al., “Development of the 11 T Nb3Sn Dipole Model at Fermilab”, IEEE Trans. on Appl. Supercon, v. 10, No. 1, 2000, p.298. F. Asner et al., “First Nb3Sn, 1 m Long Superconducting Dipole Model Magnet for LHC Break the 10 Tesla Field Threshold”, Proc. of the Magnet Technol. Conf. (MT-11): Elsevier Applied Science, 1990, p.36. A. den Ouden et al., “An Experimental 11.5 T Nb3Sn LHC Type of Dipole Magnet”, IEEE Trans. On Magnetics, vol.30, no.4, 1994, p.2320. D. Dell’Orco et al., “ Design of the Nb3Sn Dipole D20”, IEEE Trans. On Applied Superconductivity, vol. 3, No.1, p.82, 1993 D.R. Chichili et al., “Design, Fabrication and Testing of Nb3Sn Shell Type Coils in Mirror Magnet Configuration”, CEC/ICMC 2003, Alaska, September 22-25 2003. E. Barzi et al., “Field Quality of the Fermilab Nb3Sn Cos-theta Dipole Models”, EPAC2002, Paris, June 3-7 2002, p.2403. V.V. Kashikhin et al., “Passive Correction of the Persistent Current Effect in Nb3Sn Accelerator Magnets”, IEEE Trans. on Applied Superconductivity, v. 13, No. 2, June 2003, p.1270. N. Andreev et al., “Development and test of single-bore cos-theta Nb3Sn dipole models with cold iron yoke”, IEEE Trans. on Applied Superconductivity, v. 12, No. 1, March 2002, p. 332. S. Feher et al., “Test Results of Shell-type Nb3Sn Dipole Coils”, IEEE Trans. On Applied Superconductivity, vol. 14, No.2, June 2004, p.349. G. Ambrosio et al., “Development of React & Wind Common Coil Dipoles for VLHC”, IEEE Trans. on Applied Superconductivity, v. 11, No. 1, March 2001, p. 2172. G. Ambrosio et al., “Fabrication and Test of a Racetrack Magnet Using Pre-Reacted Nb3Sn Cable”, IEEE Trans. on Appl. Supercon., v. 13, No. 2, 2003, p.1284. V.S. Kashikhin et al., “Development and Test of Single-Layer Common Coil Dipole Wound With Reacted Nb3Sn Cable”, IEEE Trans. On Applied Superconductivity, vol. 14, No.2, June 2004, p.353. V.S. Kashikhin et al., “Field Quality Measurements of Fermilab Nb3Sn Common Coil Dipole Model”, IEEE Trans. On Applied Superconductivity, vol. 14, No.2, June 2004, p.287. V.V. Kashikhin, A.V. Zlobin, “Magnetic instabilities in Nb3Sn strands and cables”, this conference 5LB02 E. Barzi et al., “Transport Critical Current of Nb3Sn Strands at Low and High Magnetic Fields”, this conference 2MB04 A.K. Ghosh et al., “Dynamic stability threshold in high-performance internal-tin Nb3Sn superconductors for high field magnets” submitted to Superconductor Science and Technology. D.R. Dietderich et al., “Correlation Between Strand Stability Measurements and Magnet Performance”, this conference 1LX01 E. Barzi et al., ”Study of Current Carrying Capability of Nb3Sn Cables in Self-field Using a SC Current Transformer”, this conference 1LX06 G. Ambrosio et al., “Critical Current Measurement of Nb3Sn Rutherfordtype Cables for High Field Accelerator Magnets”, this conference 2LR03 G. Ambrosio et al., “Measurement of Inter-Strand Contact Resistance in Epoxy Impregnated Nb3Sn Rutherford Cables”, CEC/ICMC2003, Alaska, September 22-25 2003. R.R. Hafalia et al., “An Approach for Faster High Field Magnet Technology Development”, IEEE Trans. On Applied Superconductivity, vol. 13, No.2, June 2003, p.1258. S. Feher et al., “Cable Testing for Fermilab`s High Field Magnets Using Small Racetrack Coils”, this conference 2LR05 A.V. Zlobin et al., ”Development and Test of Nb3Sn Cos-theta Dipoles Based on PIT Strands”, this conference 3LG02