1. 1 SIS 300 Dipole Low Loss Wire and Cable J. Kaugerts, GSI TAC, Subcommittee on Superconducting Magnets Nov15-16, 2005

2. 2 Collider vs Fast-Ramped Synchrotron Operation For beam colliders, such as RHIC, magnet AC losses were not an important consideration, given low magnet ramp rate (0.042 T/s) and infrequent ramps. For fixed target fast-ramping synchrotrons, such as GSI‘s SIS 200 at 4 T ( and now SIS 300 at 6T) the ramp rate is high (1T/s) and ramps are frequent, so AC loss reduction is an important consideration

3. 3 Conductor Losses Wire losses 1)Filament hysteresis P f = (4d f /3  ) ∫ dB  J c (T, B  ) Coupling (eddy) current loss/volume P w P w = 2  /  0 (dB  /dt) 2  = (  0 /2 ρ et )(p/2  ) 2  ~ coupling current time constant ρet~ transverse resistivity p~ filament twist pitch Cable losses (scale with R c & R a ) 1) Crossover strand resistance R c 2) Adjacent strand resistance R a ·

4. 4 Dipole GSI 001 A 1m long dipole was built and tested at BNL for the earlier ( 4T, 1 T/s) SIS 200 synchrotron design, which was updated to the 6 T, 1 T/s present SIS 300.

5. 5 GSI 001 Dipole Losses/cycle/m assuming RHIC wire & cable (1 T/s ramp) rampingmeanloss/fraction power cycleof total Watts Joules% transv'se cr'sover1610.48 3220.692.0% transv'se adjacent10.69 21.40.6% parallel adjacent0.16 0.30.0% filament coupling111.94 223.96.4% hysteresis15.55 31.10.9% delta hysteresis1.34 2.70.1% total hysteresis16.89 33.81.0% total magnet1750.16 3499.9100.0% R c =8μ Ω no core R a =64μ Ω 13 mm fil. twist pitch

6. 6 GSI 001 Dipole Calculated Conductor Loss (as built) rampingmeanloss/fraction power cycle/mof total Watts Joules% transv'se cr'sover0.21 0.40.5% transv'se adjacent10.69 21.427.7% parallel adjacent0.16 0.30.4% filament coupling10.60 21.227.5% hysteresis15.55 31.140.3% delta hysteresis1.34 2.73.5% total hysteresis16.89 33.843.8% total magnet38.54 77.1100.0% SS core in cable R c =62.5 mΩ R a =64 μΩ Fil twist pitch=4 mm

7. 7 SIS 300 Dipole Loss Reduction Previous slide shows that R a, coupling currents and filament hysteresis constitute major loss sources for cored cable conductor. Loss reduction: 1) increase R a 2) Increase matrix resistivity, to reduce coupling currents 3) decrease filament diameter, to reduce hysteresis loss

8. 8 R a Loss Reduction R a can be increased by heating cable in air R a increase may reduce current sharing capability of wire and decrease conductor stability. No quantitative data are available, to my knowledge.

9. 9 Higher resistance wire matrix Cold working the copper in the wire during it‘s production can provide a higher resistivity wire matrix, but cable heat treatment, due to coil curing, or heat treatment to increase R a will reduce this resistivity again. High resistivity barriers (such as CuNi) around filaments or filament regions increase the effective,or transverse, resistivity of the wire A Cu.0.5-0.6% Mn interfilamentary matrix also increases the transverse resistivity and is unaffected by cable curing or heat treatment.

10. 10 Small filament wire Below about 3.5 micrometer filament size, proximity coupling again increases filament hysteresis loss in an all-copper matrix wire ( keeping s/d constsnt) s~filament spacing, d~fil. dia. Use of a CuMn interfilamentary matrix eliminates proximity coupling effects for filament sizes down to around 1 micrometer

11. 11 SIS 300 Dipole Wire Parameters (with Cu matrix wire) Strand diameter 0.825 mm Filament diameter 3.5 micrometers Filament twist pitch 5 mm Matrix/NbTi ratio 1.4 (1.5) Strand transverse resistivity  et (4 + 0.9 B)10 -10 Ohm m (goal) Strand transverse resistivity  et (1.4 +0.9B)10 -10 Ohm m (calculated with all-copper matrix, 1.5 Cu/SC ratio) Present EAS wire with 4.3 micrometer filaments, 1.75 Cu/SC ratio has measured  et =(0.58+0.9B)10 -10 Ohm-m Wire Strand coating Sn Ag (Stabrite) Critical current density Jc 2700 A/mm2 ( 5 T, 4.2 K) Critical current density Jc 2130 A/mm2 ( 6T, 4.2 K)

12. 12 SSC Cu-0.6%Mn Interfilamentary matrix 2.5 micron filament wire* Global matrix ratio: 1.7 Filament number:22686 Filament diameter: 2.63 μm Wire twist pitch: 12.5 mm Transverse resistivity ρ et = (4.15 + 1.9B)10 -10 Ωm ( For RHIC wire ρ et = (1.24 + 0.9B)10 -10 Ωm) Wire diameter: 0.651 mm Jc =2760 A/mm 2 ( 5 T, 4.2 K) ( best value achieved) * Made for possible use in the SSC High Energy Booster (HEB), using a double stacking production method, and tested for GSI at Twente TU

13. 13 Another Possible CuMn Interfilamentary Matrix Wire for SIS 300 IGC fabricated a 309 mm billet into wire of 2.6 micron filament diameter, with a Cu-0.6%Mn interfilamentary matrix, using a patented single stack approach, also for SSC use. Further parameters are: Filament number: 38663 Matrix to NbTi ratio: 1.5 Wire diameter: 0.808 mm J c = 2753 A/sqmm at 5T, 4.2 K ( best value achieved) Such a conductor requires scaling up by a factor of 1.02 in diameter, for application in the SIS 300 dipole. Calculated value for transverse resistivity ρ et =3.410 -10 Ωm Coupling current time constant  =1.17 msec for 5 mm fil. twist pitch.

14. 14 SIS 300 dipole Loss/cycle-m with Cu matrix ramp'gmeanloss/fraction power cycleof total Watts Joules% transv'se cros'r0.36 3.24.0% transv'se adj'nt1.21 10.713.2% parallel adjacent0.01 0.10.1% fil'nt coupling2.36 20.825.7% hysteresis5.12 45.055.7% delta hysteresis0.12 1.01.3% total hysteresis5.24 46.157.0% total magnet9.30 80.8100.0% Rc=20 mΩ,SS core in cable Ra=200 μΩ 5 mm fil. Twist pitch 3.5 μm filaments

15. 15 SIS 300 Dipole Loss/cycle-m with CuMn interfilamentary Matrix ramp'gmean loss/ fraction power cycle of total Watts Joules % transv'se cros'r0.36 3.2 5.7% transv'se adj'nt1.21 10.7 19.0% parallel adjacent0.01 0.1 0.2% fil'nt coupling1.05 9.3 16.5% hysteresis3.66 32.2 57.3% delta hysteresis0.08 0.7 1.3% total hysteresis3.74 32.9 58.6% total magnet6.38 56.2 100.0% R c =20 m Ω SS core in cable R a =200 μΩ 5 mm fil. Twist pitch 2.5 μm filaments

16. 16 Loss Reduction with CuMn interfilamentary matrix Higher transverse resistivity and smaller filament size give 32% loss reduction over all-cu matrix.

17. 17 Tested Wires 1.40 0.94 0.92 1.15 1.10 1.23 2A12 3N7 RHIC K2 001T4 G2 001T6 SSC CuMn Ratio J cm /J t double stacked single stacked single stacked double stacked double stacked triple extruded, double stacked Wire ID

18. 18 Single stacked wire

19. 19 Filament Distortion Effects Wires made with a double stacking process show a greater filament distortion than wires made with a single stacking process, as shown by the difference in magnetization & transport current densities for the preceding wires.

20. 20 Other Interfilamentary Matrix Materials Aside from Cu-0.5wt% Mn, Cu-10wt%Ni and Cu-30wt%Ni have been used to reduce eddy current losses in low loss strands

21. 21 Wire Coupling Current   for SIS 300 Wire with Various Interfilamentary Matrices and Barriers Case No. Interfil. Matrix mat. Barrier Mat. Filament Diam. d f (  m)  (msec)  et  -10  m) Notes 1Cunone3.52.781.43 RRR Cu =278 RRR Cuint =25 2 Cu- 0.5wt%Mn none2.51.342.97  CuMn =250 3 Cu- 0.5wt%Mn none2.51.073.72 RRR Cu =220 (this case only) 4 Cu- 0.5wt%Mn Cu- 10wt%Ni 2.5 0.488.37  CuNi =1400 5 Cu- 10wt%Ni 2.50.449.00 6 Cu- 10wt%Ni none2.51.3093.04 7 Cu- 30wt%Ni none2.51.3043.05  cuNi = 3640

22. 22 CuMn versus CuNi Interfilamentary matrix Cu-10wt%Ni is about 6 times more resistive than Cu-0.5wt%Mn. For stability reasons, avoid making matrix more resistive than needed to reduce AC loss. Cu-0.5%Mn is as effective as Cu-10wt%Ni in reducing strand eddy current loss CuNi contains 0.15-1.0 % Mn, so the “ active ingredient“ for proximity effect suppression appears to be Mn is both cases

23. 23 Jc (A/mm 2 vs B (T) for pure CuNi matrix switch wire

24. 24 Switch wire performance conclusions Short samples instabilities: Inception of instabilities at low field, depending on wire diameter d Self field instability Virtually independent of Filament size CuNi composition (between CuNi 30 and CuNi10) Stability limit: J c d ~ 2000 A/mm

25. 25 Low Loss Wire Conclusion A Cu-0.5-0.6%Mn interfilamentary matrix wire with fine ( 2.5 μm or less) filaments, made by a double stacking process ( assembly easier & better stability) appears to give a wire with the lowest loss. Jc above 3200 A/mm 2 has been achieved for commercially available CuMn interfilamentary matrix wires with 5.3 micron filaments and Jc above 2700 A/mm 2 has been achieved for 2.5 micron filament conductor, but R&D is probably required to optimize J c & piece length

26. 26 Present Wire Status EAS has produced 84.6 kg of 0.825 mm dia. wire, with 12318 filaments of 4.3 μm diameter, Cu/sc ratio of 1.75 and J c ( 5 T, 4.2 K) in the range 2711-2752 A/mm 2 The wire has been sent to Alstom for cabling A 1m long model magnet would require at least 55 kg of wire (341 m of cable). A 2.908 m ( eff. length) prototype magnet would require 160 kg of wire. Alstom can provide 300 m of cable ( 0.825mm dia. Wire) with19200 3.5 μm filaments ( single stacked), Cu/sc ratio of 1.9, 3-4 months after order.

27. 27 Problem We need to order more wire, to build SIS 300 model, or prototype, dipoles. Lead time between wire RFQ and wire receipt is 9-12 months Solution Make two 200 kg billets. First one, with 2.5 micron filament wire. During fabrication of this wire, determine possible Jc degradation with further filament decrease and determine minimum filament twist pitch before Jc decreases. Make second billet with optimum parameters.

28. 28 Cable R a & R c Rc> 100 mΩ for cables heat treated at 200C for 4 hours ( IHEP tests on cored LHC outer layer cable) R a ~ 200-300 μΩ for 8 hour bare cable heat treatment at 200 o C & 30 minute cure cycle of polyimide tape insulated samples at 195 o C & 15-70 MPa (BNL tests). Need more statistics.