Superferric Magnets for JLEIC Peter McIntyre, Dior Sattarov, Jeffrey - - PowerPoint PPT Presentation

Peter McIntyre

October 8, 2015

Superferric Magnets for JLEIC

Peter McIntyre, Dior Sattarov, Jeffrey Breitschopf, Daniel Chavez, James Gerity, Joshua Kellams, Katie O’Quinn, Tim Elliott, Ray Garrison Texas A&M University

Superferric Magnets for the Ion Ring and Booster

Arc, 261.7°

IPs ions

Ion Ring– 128 arc half-cells Booster – 32 arc half-cells Superferric magnets have been designed for the requirements of arc half-cells: Ø Ion Ring: 8-100 GeV protons: 0.25 -- 3 T dipoles, 52 T/m quads Ø Booster: 0.2-8 GeV protons: 0.24 – 3 T dipoles, 6 T/m quads

Half-cell cryostat geometry for Ion Ring arcs

= leads = fixed anchor to support post = sliding anchor to support post = sliding shroud section of vacuum vessel

F S F F F S S

11.4 m overall half-cell length Dipole aperture requirement: betatron amplitude (15 s) @ injection: ±3 cm dispersion of ±0.5% momentum spread: ±1 cm sagitta (with 4 m dipole length): ±1.8 cm ±5 cm Quad aperture radius requirement: 4 cm

Each half-cell contains two 4 m dipoles, one 0.8 m quadrupole, 1 sextupole to correct body sextupole in dipoles (Neuffer):

Superferric Magnets – Cost Minimum up to ~4 T

3 T SSC dipole 2 T pipe dipole for VLHC 1 T strong-focusing cyclotron 3 T proton gantry for particle beam therapy 4.5 T for 100 TeV hadron collider 2 T FAIR dipoles

MEIC Arc dipole

The biggest challenge is to create a 10 cm x 6 cm aperture with the field quality needed for high-luminosity collisions with long luminosity lifetime – dynamic aperture

Initial 3 T design Initial 3.5 T @15kA design, conventional winding 3.5 T design, barrel winding 3.0 T design, barrel winding, b2<2 units

Design stages of CIC-based >3T JLEIC dipole

all allowed multipoles <1 unit all allowed multipoles <1.5 unit Green is good!

13.7 kA @ 3.0 T, bn < 1 unit all fields

I cable 13.7 kA Estored 64 kJ/m Inductance 0.68 mH/m Bcoil 3.5 T Tquench 5.94 K ISSL 19.7 kA Bcoil 4.6 T Bbore 3.85 T Nturns/bore 2x12 Nstrands 15 dstrands 1.20 mm Cu/Non Cu 1.5 Area 8.23 cm2 Expect ~1% SSL reduction due to solid conductor approximation. Magnet will operate at 79% of SSL

(Fx kN/m, Fy kN/m) (Fx kN/m, Fy kN/m) (Fx kN/m, Fy kN/m) (Fx kN/m, Fy kN/m) (31.95,-3.06) (38.25,-3.47) (30.24,-7.37) (36.13,-8.63) (18.44,-7.65) (20.77,-8.23) (4.19,-7.84) (4.97,-7.92) (30.82,-3.26) (36.54,-3.74) (18.29,-4.13) (20.80,-4.25) (3.34,-4.40) (4.21,-4.38) (31.61,0.98) (37.05,1.28) (19.28,-0.86) (21.94,-0.71) (3.54,-2.09) (4.42,-1.91) (21.78,0.42) (24.56,0.47) (4.33,-0.50) (5.03,-0.44)

Lorenz body forces on each turn of the MEIC dipole @ 3 T (top set from 1.2 m 3d model, bottom set from 2d model):

Cable-in-Conduit: Dubna to GSI to MEIC

The SIS-100 ring uses superferric dipoles operating at 1.8 T. Its conductor is a semi-rigid cable-in-conduit, in which the helium cryogen flows internally so that the magnet is not immersed in liquid helium. Cable-in-conduit makes a much simpler end geometry for a large-bore dipole. The windings can be supported in a reinforced polymer structure, with tight precision.

We follow the Dubna/GSI CIC strategy with a few improvements for higher-field operation

15 NbTi/Cu wires are cabled onto a perforated spring tube. The cable is inserted in a sheath tube, and the sheath is drawn

• nto the cable to just compress the wires against the spring tube.

1.2m long straight section Symmetric 3D model of MEIC dipole at 2.92T central field Max field in the conductor 3.33T (3.42T if scaled to 3T central field ) located at (0.0242,0.042,0) Location and magnitude (3.49T) of max field point is consistent with 2d simulations done in COMSOL

3-D FEA to simulate deflections from axial forces on end windings

The end windings are supported in a nest of G-11 forms, bolted to the SS bars and beam tube and vacuum-impregnated with filled epoxy. Only the end region is impregnated. Deflections due to Lorentz forces on ends are shown. G-11 support matrix is suppressed in this image for clarity. Maximum displacement ~ 4 µm

Superferric Dipoles: How we build them

3. Insert the SS beam tube, seal the ends, and epoxy impregnate the gap between segments and beam tube.

• 1. Fabricate inner form segments from 4”-

thick G-11 fiber-reinforced epoxy slabs.

• 2. Assemble stack of segments for dipole

body, using the CIC channels for alignment.

• 3. Improvement: Channels for cables are split

equally on the facing segments. Strategy:

• All cables are positioned

sandwiched between layers of precision-machined structure.

• Ends are formed to the side of

the dipole, then popped into place in the structure layer.

• Overall coil assembly is

preloaded within steel flux return, all windings immobilized.

Assembly/alignment / epoxy impregnation

• f beam tube & winding structure

All ends are formed with 2” radius bends. 1. Bend a U with the correct horizontal spacing.

• 2. Bend the U to form a 90° ear, with offset

for layer-layer transitions.

• 3. ‘Odd-man’ turns require

forming a ‘dog-bone’ end.

Forming the flared ends requires production tooling

We have validated that bends preserve internal structure, do not damage NbTi wires.

See videos of the real tools making these bends at https://goo.gl/VoSDOS

Motorized Benders in operation

Impregnated assembly ready for layer 1

Mockup Winding – Layer 1 wound

Measurement of cable positions for Layer 1

1 2 5 6 4 3

3 4 5 6 7 8 9 10 11 12 13 16 Right left 1 2 15 14

Error matrix from magnetic model: bn, an from .001” error on each cable x,y

Cable placement measurements for Layer 1:

Measure cable positions for Layer 1

Extract random multipoles from cable placements in Layer 1:

Caps/forms in place – ready to wind layer 2

Layer 2 complete

Caps/forms in place – ready to wind layer 3

Layer 3 complete

Final caps in place – preload and QC

Apply flat plates to top/bottom surfaces, clamp sides and top/bottom, dilatometry of x/y cable positions through open slots in G-11 structure.

5/20/2016: Mockup winding complete

Measure all cable positions – Calculate contributions to multipoles

Shimming strategy to trim 2 dominant multipoles from warm measurements

Developing this workable strategy to trim a1, b2 multipoles after warm measurement is an important milestone in maturing the CIC superferric dipole for use in a collider. Doing it in practice will be a goal for model dipole construction and testing.

You goof something once in a while…

We reached the last turns of layer 2, and realized that the 3-D CAD design of the end topology had a half- twist that would require passing the 1m dia. feed spool of cable through the last finished turn! We managed to do it (!), and corrected the design…

Next: the complete fabrication

Body segments assembled on beam tube, jig-located, epoxy- impregnated SS bars, end frames installed, ready to wind first layer First layer wound Second layer wound Third layer wound Cable frame complete, End covers installed Ti channels, SS top/bottom skins installed Flux return halves installed and closed, SS shells welded

Quench Protection

Quench heater foils are bonded in a 10 cm end segment of the G-11 structure on both ends of the dipole. Every cable turn is driven normal in ~10 ms by a current pulse to the heater foils.

50 100 150 200 250 0.05 0.1 0.15 0.2 T (K) t (s) Peak Temperature quench both ends quench one end 5 10 15 20 25 50 100 150 200 MIITs, kA2 s T, K MIITs

Maximum pressure during quench in the 250 m length CIC of a 4 m dipole = 41 bar

Quench is driven at both ends of the dipole

Redundancy: Even if one quench heater ckt fails, the remaining

• ne protects the magnet.

Supercritical He is single-phase – it cannot boil. Enthalpy/volume is a function of pressure only, not temp.

Peak temperature in quench= 150 K Peak pressure in CIC cable = 41 bar

Provision for differential contraction, preload against Lorentz forces in body

Horizontal contraction

• 3
• 1

1 3 0.5 1 1.5 2 2.5 3 3.5 units Bcentral (T) Central Multipoles Cold+Forces @2cm b2 b4 b6 b8

• Orient G-11 with fabric perpendicular to

beam tube axis – small contraction.

• Locate Ti U-channels to compensate

differential contraction – warm-cold ~+.002”

• Locate SS/mica paper slip plane between G-

11 and steel flux return.

Ti G-11 316 SS Steel

CIC winding limits temperature rise in event of heat from beam losses

Simulated temperature distribution in the presence of 1 W heat deposition in a MEIC dipole winding.

4.5 K 4.4 4.3 4.2

CIC structure controls Lorentz stress, prevents coil motion, error fields

Y displacement (RT – 3 T) .001” X displacement (RT – 3 T) .001” Multipole effects <0.2 units

Half-cell cryostat

Support load from 5 reentrant feet. Supports integrate provisions for precise positioning & internal alignment of all elements. 50 K shield, MLI, and top-half shell go on after all

• alignment. Ports for checking alignments.

Static heat loads ~0.5 W to 4.5 K, 50 W to 50 K.

injection extraction

RF cavity

Crossing angle: 75 deg.

Booster (8 GeV, gt = 10)

g =

56 t

S M

M56 = D ρ ds

∫

Injection: multi-turn 6D painting

0.22-0.25 ms long pulses ~180 turns Proton single pulse charge stripping at 285 MeV Ion 28-pulse drag-and-cool stacking at ~100 MeV/u Ion energies scaled by mas-to-charge ratio to preserve magnetic rigidity Ekin = 285 MeV – 7.062 GeV Ring circumference: 273 m (≈ 2200/8) 38

272.306 70 7

• 7

BETA_X&Y[m] DISP_X&Y[m] BETA_X BETA_Y DISP_X DISP_Y

Straight

• Inj. Arc (2550)

Straight (RF + extraction) Arc (2550) =

56

273 M cm

Bogacz

Booster arc magnets

For planning purposes we could provide the required fields and apertures for the Booster magnets by building Ion Ring arc dipoles and quads with appropriate lengths: Dipole 1.2 m Quad 0.4 m It may likely prove to be the case that making the dipoles of a common design is less expensive than making dedicated designs with smaller quad gradient.

8.76306 3 BETA_X&Y[m] DISP_X&Y[m] BETA_X BETA_Y DISP_X DISP_Y

Bend: Lb = 120 cm (magnetic length) Lead ends: 2×22 cm B = 2.73 Tesla bend ang. = 7.08 deg. Sagitta =1.8 cm Bend

Sextupole

Bend Sextupole: Ls = 10 cm S = 750 Tesla/m2

Correctors BPM

Quad

Quadrupole: Lq = 40 cm G = 1.2-5.8 Tesla/m Correctors (H/V): 20 cm BPM can: 20 cm Dual-dipole

Quad

Half-cell cryomodule

Current Status of Development:

üWe have built a 1.2m mockup winding. üWe have measured cable positions to determine structure multipoles. We met the tolerances to provide the specified multipoles. üWe developed a shim procedure to further reduce important multipoles if desired.

Hypertech is endeavoring to adapt their Continuous Tube-Forming Process to form the sheath tube by rolling and welding a perforated strip directly onto the cable

15 NbTi/Cu wires are cabled onto a perforated spring tube.

First efforts show He leaks along seam before/after bending the tube. For model dipole, we will lay out 125 m straight tube on floor, pull cable into it, draw down in the corridor of our building.

Work in progress

• We are validating single-strand performance of

the NbTi wire chosen for the CIC cable to verify that there is no degradation from cabling and bending.

• We are ready to order long-length wire, sheath

tube, fabricate cable, pull into sheath, draw to final size. Target piece length = 125 m

• Long enough for full winding of 1.2 m model

dipole.

• Long enough for top/bottom half-winding of 4 m

dipole.

• Waiting for $ to fabricate improved G11

structural beam, then wind the model dipole.

• Fabricate CIC cable
• Build laminated flux return
• QC cable positions of all turns in dipole

Texas A&M group will propose the balance

• f NP$ needed to build model dipole
• Assemble winding in flux return, preload
• Measure cable positions under preload
• Warm measure, Shim to kill multipoles
• Cold test the model dipole

It is worth a fresh look at the projected cost to build a first 4 m JLEIC dipole.

The total costs for each category are ~1.5 x higher than the TAMU estimates in our first cost exercise in December 2014. But this is our current-best projection for the cost to build the first dipole. I am optimistic that we will meet our

• riginal cost and performance objectives

for the manufactured run of 256 dipoles for JLEIC.

Bottom-up cost detail – dipole cold mass

Dipole cold mass cost estimate presented at MEIC review 12/2014.

CIC Magnets for IR

Ø Quadrupoles must operate in the fringe field of the ~3 T detector solenoid. Ø The FF magnets must operate over a large range of beam energies: no PM. Ø FF quads must focus ions after IP collision must match to the collider lattice, but must have large aperture to pass scattered. QF1 requires 12 T in windings. Ø E, ion quads are close to one another, must not produce field on the other beam. Ø All FF magnets must operate with high rad damage & heat load from losses.

QFFB4e QFFB3e QFFB2e QFFB1e QFFB1e QFFB2e QFFB3e QFFB4e

We have prepared conceptual designs for the four most difficult IR magnets

• All designs utilize CIC conductor.

– adaptable for challenging coil geometries – compact end windings

• Utilize MgB2 or REBCO superconductor for temperature

margin in high radiation loss in QFFB1e, QFFB2e.

• Utilize Nb3Sn superconductor for high gradient in QFFB1.
• Utilize sheath solenoid winding to cancel external flux

from spectrometer solenoid.

Ion Beam:

QFFB1: 90 T/m, 9 cm half-aperture, 36 cm from e-beam SB1: 2 T, 340 mm aperture, 25 cm from the electron beam

Electron beam:

QFFB2e: 58 T/m gradient, 3 cm half-aperture, 10.5 cm from the ion beam QFFB1e, QFFB2e are immersed in fringe field of spectrometer solenoid

Interaction Region: Ion Beam Magnets

IP

QFFB3_US QFFB2_US QFFB1_US SB1 QFFB1 QFFB2 QFFB3 SB1

Assuming 100 GeV/c Parameters are determined primarily by detection requirements rather than beam dynamics Bottom-up study of multipole requirements in progress Note: parameters are still being fine-tune but no major changes

IR Ion Magnet Parameters

Name Type Length [m] Good-field radius [cm] Inner radius [cm] Outer radius [cm]

• Min. beam

separation [cm] Strength [T or T/m] Pole-tip field [T]

QFFB3_US Quad [T/m] 1 3 4 12 36.0

• 116
• 4.6

QFFB2_US Quad [T/m] 1.5 3 4 12 26.5 149 6 QFFB1_US Quad [T/m] 1.2 2 3 10 18.0

• 141
• 4.2

SB1 Dipole [T] 1 4 17 24 25.0

• 2
• 2

QFFB1 Quad [T/m] 1.2 4 9 17.1 35.9

• 88
• 8

QFFB2 Quad [T/m] 2.4 4 15.7 24.7 48.2 51 8 QFFB3 Quad [T/m] 1.2 4 17 26.7 67.2

• 35
• 6

SB2 Dipole [T] 4 4 40 90 102 4.7 4.7

Interaction Region – e-Beam Magnets

IP

QFFB4e QFFB3e QFFB1e QFFB1e_US QFFB2e_US QFFB3e_US QFFB2e

Assuming 10 GeV Parameters are determined primarily by beam size and available space Multipole tolerance study has not been done yet One has to consider effect of the solenoid fringe field Note: parameters are still being fine-tune but no major changes

IR Electron Magnet Parameters

Name Type Length [m] Good-field radius [cm] Inner radius [cm] Outer radius [cm]

• Min. beam

separation [cm] Strength [T/m] Pole-tip field [T]

QFFB4e Quad 0.5 4 5 11 21

• 3.1
• 0.16

QFFB3e Quad 0.58 4 5 11 15 47.9 2.39 QFFB2e Quad 0.7 2 3 8 10.5

• 57.7
• 1.73

QFFB1e Quad 0.4 1.2 2 6 8 24.4 0.49 QFFB1e_US Quad 0.7 2 3 7 12

• 43.9
• 1.32

QFFB2e_US Quad 0.7 4 5 10 16 45.5 2.28 QFFB3e_US Quad 0.5 4 5 10 22

• 16.4
• 0.82

FF1 FF2 e- P+ 1 2 3 4 5

• 1

30 mm 90 mm FF3 60 mm

a b c d FF2 e- P+ 1 2 3 4 5

• 1

30 mm 90 mm FF3 FF1 PM 60 mm

a b c d

Option 1: Option 2:

Synchrotron Light is a Major Challenge

Ion beam quad QFFB1: 90 T/m, 9 cm half- aperture, 36 cm from e-beam

T

Reverse-current winding kills fringe field at the location of the electron beam.

electron beam

Nb3Sn windings, 4.2 K 9 kA cable current

Dipole SB1: 2 T, 340 mm aperture, 25 cm from the electron beam

m

Window-frame C-geometry dipole configured as a Lambertson septum to suppress fringe field at electron beam.

electron beam T

MgB2 windings, 10 K 4 kA cable current

Quadrupole QFFB2e: 58 T/m gradient, 3 cm half-aperture, 10.5 cm from the ion beam

MgB2 windings @ 10 K REBCO windings @ 50 K?

ion beam

CORC: Cable-in-Conduit using REBCO

CORC cable: a) forming flat REBCO tapes onto a center tube; b) cross-section of completed CORC cable; coiled turn of CORC cable illustrating its flexibility and stability. Marrying CORC and CIC: a) form the REBCO tapes onto perforated core tube, and pull CORC cable into sheath; b) draw sheath onto CORC cable.

QFFB1e - Quad with Iron Flux Return: Cancel flux from 3 T spectrometer solenoid

QFFB1e QFFB2e 3 T

The problem:

• Need superconducting quads (not PM quads) in e-beam FF to be able to operate

with range of electron energies, tune the FF optics for optimum dynamic aperture.

• Superconducting quad needs iron flux return, but iron pulls in fringe flux from

spectrometer solenoid and saturates.

Stealth magnetics: exclude fringe field

Problem: Iron sucks in solenoid fringe field, saturates. Impossible to shape and control gradient. Solution: wrap superconducting solenoid winding on flux return, Adjust K(z) to exclude flux from spectrometer.

QFFB1e, QFFB2e with and without suppression of solenoid fringe

By adjusting K(z) we can exclude the fringe field of the spectrometer solenoid, so that the iron-clad quads operate in the normal fashion.

Conclusions

• We have made preliminary magnetic designs that can

achieve the requirements for all of the special magnets for the IR region.

• CIC windings provide flexibility to meet the

requirements with compact structure, supress fields at close-lying neighbor beam.

• CIC windings make it possible to use Nb3Sn, MgB2

superconductor to provide thermal headroom.

• Stealth magnetics can be used to operate

superconducting quads in fringe field of spectrometer.

Texas A&M group will propose NP funding for development of serious conceptual designs for IR magnets

• Optimize magnetics for required field quality
• Build/test short segments of CIC with Nb3Sn, MgB2
• Develop concepts for fabrication strategy
• Prepare details for building model of QFFB1e,

testing in fringe field of existing solenoid in FY2018