12 Tesla Hybrid Block-Coil
Dipole for VLHC
Raymond Blackburn, Nicholai Diaczenko, Tim Elliott, Rudolph
Gaedke, Bill Henchel, Ed Hill, Mark Johnson, Hans Kautzky,
Peter McIntyre, Al McInturff, and Akhdior Sattarov
Texas A&M University
aluminum
compression shell
coil housing
flux return steel
inner expansion
bladders
expansion
bladders
outer expansion
bladders
Kapton
strain gages
laminar springs
mica paper
shear release
outer coils:
50% copper strands
inner coils:
100% superconducting strands
Motto for high-field dipoles:
keep it simple, stupid!
The problems for Nb3Sn high-field dipoles:
• Conductor is fragile – wind & react,
degradation under Lorentz loading
• Filaments are fat –persistent current
multipoles, snap-back
• Preload is immense – how to assemble?
• Conductor is expensive – 10 x NbTi
Block-coil designs enable us to
address these problems
• Stress management: limit coil stress
• Racetrack pancake coils
(bend ends up/down on center layers)
• Close-coupled steel reduces amp-fac, suppresses
persistent-current multipoles
• Simple assembly, preload using expansion
bladders
• Conductor optimization – least superconductor of
any high-field design!
Stress management
• Each block within the coil controls stress so
that it cannot accumulate from inside blocks
to outside blocks:
Pancake coils are compartmentalized
so that they are easy to build, and
control axial stress internally
Center double pancake
top/bottom single pancakes
We have built a NbTi practice dipole
to test fabrication, assembly issues
7 Tesla short-sample field
6 layers
Single-block pancakes
Ends planar
Ribs, plates, springs, shear
release, S-glass insulation,
strain transducers as will be
used in Nb3Sn models.
Vacuum-impregnated coil!
Coil winding
• Winding uses simple
tooling, fixtures
• Tolerances held to .002”
• Transitions, leads made
with S-bends
Splice joints
• Splices were made as
horseshoe, 4” overlap
• Heaters control temp
• Splice rigid on coil end
• For Nb3Sn we will make
straight splices
Ribs & plates control stress
both transverse and axial
• Ribs are EDM cut, give
dimensional control and
bypass of stress.
• Plates are fabricated as
two half shells, welded
together at the ends to
control axial stress.
Bending ends on a pancake is easy!
• Coil package is flexible,
ends are easily bent by
hand.
• Practice dipole was built
with planar coils, Nb3Sn
model will have ends of
center 2 layers bent 90o.
Measure & control coil placement
• Measure coil thickness as
function of compressive
load
• Measure plate, rib locations
as preload is applied, to
assure closure of the
rib/plate interface
Quench heaters – how best to insulate?
• A triple failure: cable frayed
on tight bend, mica paper
frayed in winding, S-glass
fabric shifted in assembly.
• Dilemma between good
electrical insulation, good
heat transport.
Interconnections, final assembly
• Leads brought out along top
and bottom in support rails
• All electrical connections
routed on flex PC top/bottom
Test construction features and essential
elements of stress management
First encounter a turn/turn short… then fix it by fast-ramp quenches.
The fix worked!
High-current quench is quiet.
High-current testing
TAMU-1 Quench History
8
Iq (KA)
6
4
System
Training
2
Ramp-Rate
0
0
5
10
15
Ramp #
20
25
AC losses, splice resistance
quench current (kA)
10
8
6
4
ramp-rate studies
training
2
0
1
10
100
1000
10000
ramp rate (A/s)
Splice resistance measurement
Voltage (  V)
1
0.5
0.28 n
0
-0.5
-1
-1.5
1
2
3
4
5
Current (kA)
5
6
7
12 Tesla Nb3Sn design
No axial Lorentz stress on center winding pair.
Axial preload on outer windings locked into coil structure.
Pancake coils contain internal
complete internal structure
• Side bars give stiff
support, tie ends
• Skins welded to side
bars – preload
• Pusher shoes on
ends – axial preload
• Straight leads
Preload coil within flux return
using additional bladders
coil housing
expansion
bladders
outer expansion
bladders
Kapton
strain gages
laminar springs
mica paper
shear release
outer coils:
50% copper strands
inner coils:
100% superconducting strands
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 fullfield Lorentz load
• In cooldown, Al shell delivers
additional preload
Bladders provide all preloads,
relax tolerances
Magnetics: planar steel, current program
• Planar steel boundaries
• Suppress persistent current multipoles 10x
• Current program Iin, Iout to yield bn< 10-4 cm-n
Magnetics: contoured steel, single current
• All windings operate at a single current
• Contour flux return to cancel b2 at injection
• bn< 10-4 cm-n over 20:1 field range (no holes!)
Optimize the conductor
• Quench stability – enough Cu to heal microquenches –
much less Cu than…
• Quench protection – distribute the energy during a
quench -- jCu < 2,000 A/mm2
• The expensive way: draw Cu into SC strand for both
stability and protection.
• The optimized way:
draw Cu into SC strand only for stability (~40%)
cable pure Cu strands with SC strands for protection.
Half the outer coils are “free” Cu
strands = half the cost!
NbTi
Nb3Sn/NbTi
50
40
coil area (cm2)
quadratic B dependence
LHC
30
SSC
Tevatron
20
VLHC
block-coil
10
Pipe
RHIC
0
0
5
10
field strength (T)
15
Suppression of Persistent-Current
Magnetization Multipoles
• Persistent-current fields are generated from
current loops within the “filaments”.
The steel boundary in a block-coil dipole
suppresses p.c. multipoles at low field
We have evaluated five scenarios for p.c.
multipoles. Same coil assembly in all cases.
1) Flat-pole flux return
2) Curved-pole flux return
3) Flat-pole flux return and 3 mm steel sheet
4) Curved-pole flux return and 3 mm steel sheet
5) No flux return (~equivalent to cos )
The steel flux plate redistributes
flux to suppress multipoles
The flux sheet suppresses
persistent-current multipoles 3x
0.0
-0.1
-0.2 -0.3 -0.4
Central Field (Tesla)
-0.5 -0.6 -0.7 -0.8 -0.9 -1.0 -1.1
-1.2 -1.3 -1.4
-1.5 -1.6
0
-5
b2(10^(-4) units)
-10
-15
-20
No_yoke
Planar_Iron
Curved_Iron
Planar_Sheet
Curved_sheet
-25
-30
-35
-40
-45
Curved_thick_sheet
Current Design - Injection
Current Design – 12 Tesla
Multipoles with Persistent Currents
Magnets are expensive in a hadron
collider, but so is the tunnel
superconductor
35
tunnel
total
cost ($M/TeV)
30
25
20
15
10
5
0
0
5
10
15
field strength (T)
20
25
The Tripler Dipole
13.3 T @ 4.5 K
Single current coil
bn < 10-4 cm-n
Suppression of persistent
current multipoles