Recent Developments in Nb3Sn
Dipole Technology at Texas A&M
Peter McIntyre
Dept. of Physics
Texas A&M University
p-mcintyre@physics.tamu.edu
The goal:
Dipoles for future hadron colliders
TAMU4: 14.1 T, 4 x 3 cm2 aperture
cm2
LHC Tripler: 24 T, 56 mm aperture
28
superconductor
Windings = Bi-2212 inner, Nb3Sn outer
Collider-quality field, suppress p.c. multipoles
Designing dipoles with Nb3Sn
The challenges
• The conductor is fragile – strain < 0.5%
• High field limit would be imposed by Lorentz stress
• Filaments are large – snap-back too large
The solutions
• Block-coil geometry
• Stress management
• Hydraulic preload
• Flux-plate suppression of snap-back
Stress management
Offload stress from windings to structure
stress (PSI) in structure @ 14 T
stress (PSI) in coils only @ 14 T
Field strength decreases smoothly conductor optimization
Mixed-strand cable places Cu strands where
they are needed for quench protection.
Example: 12 T dipole – outer
winding can even be NbTi!
Pancake coils are easy to build,
control axial stress internally
Center double pancake
top/bottom single pancakes
Provide overall preload using
expansion bladders
• Flux return split vertically,
serves as piston
• Bladders filled with low-melt
Wood’s metal
• Bladders located between
flux return and Al shell
• 2,000 psi pressure delivers
full-field Lorentz load
• In cooldown, Al shell delivers
additional preload
Suppression of multipoles from
persistent current magnetization
• Persistent magnetization is generated from current
loops within the filaments,
• Magnetization relaxes via BIC’s, then snap-back
The steel flux plate redistributes
flux to suppress multipoles
0.5 T
12 T
Multipoles with Persistent Currents
5x suppression of p.c. sextupole – compensates for larger filament size
The Texas A&M program
• TAMU1 (6.5 T)
– evaluate block-coil geometry, winding and impregnation
strategies using NbTi model - tested to short sample
• TAMU2 (5.2 T)
– single-pancake mirror magnet with ITER Nb3Sn
conductor - completed, ready for testing
• TAMU3 (13.5 T)
– double-pancake model with 2.4 kA/mm2 conductor beginning fabrication
• TAMU4 (14.1 T )
– complete Nb3Sn dipole with 4x3 cm bore
TAMU1
• Model dipole to study block coil geometry: cable
preparation, winding techniques, impregnation: treat
exactly according to the design for Nb3Sn.
Testing of TAMU1
TAMU-1 Quench History
8
Iq (KA)
6
4
QH tests
Training
2
Ramp-Rate
0
0
5
10
15
Ramp #
20
25
Winding voltages during quench
AC losses
quench current (kA)
10
1 T/s
8
1.5 T/s
6
4
ramp-rate studies
training
2
0
1
10
100
1000
10000
ramp rate (A/s)
TAMU1 is the first fully impregnated NbTi dipole made in modern times.
It operated to short sample without training and exhibits good AC performance.
This result demonstrates that the helium access thought essential for NbTi
stability is not necessary, provided that stress is managed so as to prevent
conductor motion and friction heat.
TAMU2:
our entry into Nb3Sn technology
TAMU2: 1 single-pancake winding
mirror geometry, ITER superconductor
5.6 T short-sample bore field
Coil winding
Ti mandrel to preserve preload through
cooldown.
Inconel ribs, laminar springs transfer
stress between windings.
Reaction bake @ 650 C
Argon atmosphere purge manifolded throughout coil.
Same furnace can bake 875 C in O2 purge for Bi-2212 and
maintain separate purges of Ar in Nb3Sn, O2 in Bi-2212 windings.
We can react a 3 m long dipole in this furnace.
Splice leads Nb3Sn to NbTi
Lead is supported in rigid frame anchored into winding
superstructure, spliced to a pair of NbTi leads.
Preload, weld pancake subassembly
Preload side bars and end shoes.
Weld cover skin to stabilize coil subassembly.
Note: For stress management, we do not apply large preload, only ~3 MPa, just enough
to remove soft modulus from coil. After impreg and dipole assembly, we will apply larger
preload to the structure to provide stiff walls.
Vacuum impregnation
Horizontal orientation (with tilt), multiple flow paths assure full impregnation
We can impregnate a 3 m long dipole in this retort!
Bladder preload
Entire dipole heated to 80 C.
Bladders preloaded to 2,000 psi using hand pumps.
Pressure sustained while magnet is cooled using water jacket.
TAMU3: going to high field and
testing stress management
TAMU3: 2 single-pancake windings
3 kA/mm2 superconductor
13.5 T short-sample bore field
TAMU4: 2 single-pancakes, one double-pancake
3 kA/mm2 superconductor
14.1 T short-sample bore field
Magnets are getting more efficient!
Bi-2212
coil area (cm2)
NbTi
50
45
40
35
30
25
20
15
10
5
0
quadratic B dependence
LHC Tripler
(6x4 cm)
LHC (7 cm)
SSC (5 cm)
TAMU4 (3 cm)
Tevatron (5 cm)
RHIC (7 cm)
microbore
(3x2 cm)
Nb3Sn
Pipe (2 cm)
0
5
10
15
field strength (T)
20
25
Inject to LHC from SuperSPS
• For luminosity upgrade of LHC, one option is to
replace the SPS and PS with a rapid-cycling
superconducting injector chain.
• 1 TeV in SPS tunnel  1.25 T in hybrid dipole:
flux plate is unsaturated, suppression of snapback multipoles at injection.
• SuperSPS needs 6 T field, ~10 s cycle time for
filling Tripler  >1 T/s ramp rate
Again block-coil geometry is
optimum!
In block-coil dipole, cables are oriented vertically:
In cos  dipole, cables are
oriented on an azimuthal arch:
Result: minimum induced current loop,
minimum AC losses
Result: maximum induced current
loop, maximum AC losses

B nˆ

B  nˆ
Preliminary design for Super-SPS dipole
6 T short-sample field (to allow for AC loss degradation)
LHC NbTi strand (wider cable to optimize geometry, minimize inductance)
We are modeling AC losses, expect to be low.
Flux plate suppresses multipoles from persistent currents, AC-induced currents
(flux plate must be laminated)