Performance of Superconducting Magnet Prototypes

for LCLS-II Linear Accelerator

Vladimir Kashikhin, Nikolai Andreev, Joseph DiMarco, Alexander Makarov, Michael Tartaglia, George Velev

Abstract— The new LCLS-II Linear Superconducting

Accelerator at SLAC needs superconducting magnet packages

installed inside SCRF Cryomodules to focus and steer an electron

beam. Two magnet prototypes were built and successfully tested

at Fermilab. Magnets have an iron dominated configuration,

quadrupole and dipole NbTi superconducting coils, and splittable

in the vertical plane configuration. Magnets inside the

Cryomodule are conductively cooled through pure Al heat sinks.

Both magnets performance was verified by magnetic

measurements at room temperature, and during cold tests in

liquid helium. Test results including magnetic measurements are

discussed. Special attention was given to the magnet performance

at low currents where the iron yoke and the superconductor

hysteresis effects have large influence. Both magnet prototypes

were accepted for the installation in FNAL and JLAB prototype

Cryomodules.

Index Terms—Accelerator, Cryomodule, Linac, Magnet,

Superconducting, Conduction cooling.

I. INTRODUCTION

HE new Linear Superconducting Accelerator LCLS-II [1]

needs superconducting magnet packages installed inside

Cryomodules which are based on the superconducting radio

frequency technology (SCRF). Many different magnet

packages were built and successfully tested for Linear

Accelerators [2] – [4]. The first large scale Superconducting

Linear Accelerator XFEL [3] used superconducting magnets

cooled by a liquid helium bath [5].

In recent years a more advanced approach was developed

based on magnets with conduction cooling [6] – [12]. In order

to avoid combined installation of magnets and SCRF cavities

in a very clean room, these magnets are made splittable in the

vertical plane. This allows the magnet installation after the

cavity string is assembled, and the inner volume is sealed to

avoid contaminations cavity surfaces.

The magnet (see Fig. 1) is cooled through pure aluminum

heat sinks thermally attached to the 2 K, 5 K, and 50 K

cryomodule cooling pipes. In these magnets the main

quadrupole field is formed by four iron poles, and four

racetrack type superconducting coils. A vertical and horizontal

dipole are wound on top of each quadrupole coil to steer an

electron beam to the quadrupole field center. The magnet

design and fabrication described in [12].

Manuscript received September 1, 2016. This work was supported in part

by Fermi Research Alliance, LLC, under contract No. DE-AC02-07CH11359

with the U.S. Department of Energy. V. Kashikhin#, N. Andreev, J. DiMarco, A. Makarov, M. Tartaglia, G.

Velev are with the Fermi National Accelerator Laboratory, Batavia, IL 60510,

USA (corresponding author# phone: 630-840-2899; fax: 630-840-6766; email: kash@fnal.gov).

Fig. 1. Magnet inside the FNAL SCRF Prototype Cryomodule.

II. MAGNET PACKAGE MAIN PARAMETERS

Two magnets were fabricated to be installed in prototype

cryomodules at FNAL and JLAB. The magnet package cross-

section is shown in Fig. 2. This is a rather short 322 mm long

iron dominated magnet with relatively large 90 mm aperture.

Fig. 2. The magnet package cross-section.

The magnet parameters for the LCLS-II Cryomodules are

shown in Table 1. At the accelerator front end the quadrupole

integrated gradient is very low, only 0.05 T. Therefore the iron

core remnant and hysteresis effects could spoil the field

quality and reproducibility required by the accelerator. Both

prototype magnets were tested at FNAL Test Stand 3 in a

helium bath cooling mode, and underwent a series of quench

performance and magnetic characterization measurements

described next.

T

FERMILAB-CONF-616-TD ACCEPTED

2LPo2C-04

2

TABLE I

LCLS-II MAGNET PACKAGE PARAMETERS

Parameter Unit

s

Value

Integrated peak gradient at 10 GeV T 2.0

Integrated peak gradient at 0.4 GeV T 0.05 Clear bore aperture mm ≥78

Ferromagnetic pole tip bore diameter mm 90

Effective length mm 230 Peak quadrupole gradient T/m 8.67

Quadrupole field non-linearity at 10 mm diameter

Quadrupole field reproducibility

%

%

≤1.0

≤1.0 Quadrupole magnet DC inductance H 0.66

Number of superconducting coil packages 4

Number of superconducting sections in the coil 3 Number of turns in the quadrupole section 426

Number of turns in dipole sections 39

Peak operating current A ≤20 NbTi superconductor diameter mm 0.5

Superconductor filament size µm 3.7 Dipole correctors integrated strength T-m 0.005

Max magnetic center offset in Cryomodule mm ≤0.5

Magnet physical length mm 340 Magnet width/height mm 322/220

Quantity required 35

III. MAGNET PACKAGE ELECTRICAL AND QUENCH

PERFORMANCE TESTS

The first magnet prototype SPQA01 was cold tested in

October 2015 in Test Stand 3. Fig. 3 shows an overview and

close-up of the magnet and top plate assembly ready to install

in the helium dewar, with a 30 mm warm bore tube mounted

through and centered in the magnet aperture for magnetic

measurements. Warm electrical checks of the assembly and

instrumentation were performed prior to cool down, and

repeated when cold. Instrumentation on this magnet consists

of one cernox RTD and three silicon diodes, all mounted on

the inner coil or outer iron surfaces. The RTD resistance was

verified to be consistent with calibration values at room

temperature and in the 4.3 K helium bath. The silicon diode

voltages were measured cold with 10 µA excitation and were

consistent with the standard voltage response at that

temperature. After low current magnetic measurements were

completed (up to 10 A), the quench performance was tested:

the quadrupole, vertical dipole, then horizontal dipole were

individually ramped at 0.5 A/s to 30 A, with no quenches. All

three circuits were then powered simultaneously at 30 A for

several minutes with no quench, before ramping down to

0 A. High current magnetic measurements were then

completed, again with no quenches.

The second magnet prototype SPQA02 test began in

November, 2015 and the second thermal cycle was

completed in December. Warm and cold electrical tests were

performed as with SPQA01; an additional cold hipot test of

high voltage insulation integrity was made between the outer

(horizontal) dipole winding and the heater, which was missed

on SPQA01 (but passed warm, following the cold test).

A quench performance test was made in the first thermal

cycle, again following the low current magnetic

measurements, and all three coils were ramped to 30 A

individually and collectively held there without any

quenches. No quenches occurred during any of the magnetic

measurement ramps in either thermal cycle.

Fig. 3. SPQA01 magnet assembly ready for installation in the Stand 3

dewar for cold testing in 4.5 K liquid helium bath.

IV. MAGNET PACKAGE MAGNETIC MEASUREMENTS

The magnet package magnetic measurements were

performed by rotational coils. The rotational coil system

utilizes a PC Board design [13] and provides a measurement

accuracy of ~1 unit (10-4). The probe rotates in an anti-cryostat

(warm bore tube) placed within the magnet aperture as the

assembly is suspended in the LHe vessel. The probe radius is

limited by the ~30 mm inner diameter of the warm bore. The

PCB is 1 m long and extends beyond both ends of the magnet.

However, owing to the magnet and warm bore position in the

cryostat, the probe was not centered in the magnet, and only

extended out the far (lead) end by about 100 mm; ~200 mm

short of capturing the full end field. The board was a spare

from a previous project [8], in a temporary fixture, and as

such, the probe (and the acquisition system) were not

optimized for these measurements. All harmonics are reported

here at a reference radius of 10 mm.

The field strength of the quadrupole was measured over

three cycles at different currents. For the low field bi-polar

measurements, a bipolar 10 A Kepco power supply was used.

For current from 10 A to 30 A, a unipolar Lambda power

supply was used. The measurements match the 0.125 T/A

design value well. The hysteresis width at 1 A shows that the

change in the transfer function (TF) at lower current is about

± 5 % for both magnets.

Fig. 4. Quadrupole SPQA02 integrated gradient transfer function.

2LPo2C-04

3

The field strength of the horizontal and vertical dipole

correctors measured over 3 cycles at different currents are

shown in Fig. 5 for SOQA01 - again for both Kepco and

Lambda power supplies. The measurements are close to their

design values of 0.28 mT-m/A.

Fig. 5. Strength TF for the horizontal dipole corrector of SPQA01.

The hysteresis width at 1 A shows that the change in TF at

lower current is about ±7 % for both magnets. It should be

noted that when current goes closer to zero we are

approaching a singularity point where the relative field

distortions could be discontinuously large. But at the same

time the absolute field value is very low and comparable with

the Earth and fringe fields around the magnet.

All dominant quadrupole integrated field harmonics were

measured. The largest harmonics are below 0.1 %, except at

the lowest current measured of 0.4 A, where they are still less

than 0.5 % for SPQA01 and 0.25 % for SPQA02, including

any persistent current or magnetization contributions.

Nevertheless, tests showed rather high hysteresis effects at

low magnet currents caused by the iron yoke made from

AISI 1006 low carbon steel. A special cold test program was

developed to investigate the field reproducibility effects.

V. SPQA03 MAGNET COLD TEST

The first production magnet SPQA03 was fabricated at

FNAL with the intent to make more comprehensive cold

studies to establish the magnet operational scenarios for the

accelerator and verify reproducibility of magnetic conditions.

For reducing the remnant and hysteresis field effects,

degaussing and standardization procedures were developed.

For degaussing the following current drive formula was used:

I

where k, τ, m are coefficients that define the peak current, the

current amplitude decay, and the cycle time period. On the

base of this formula at k=64, τ=20, m=400 was programmed

the degaussing cycle for the regulated power supply shown in

Fig. 6.

Fig. 6. Power supply current variation during degaussing.

The quadrupole unipolar cycling of current after the initial

degaussing is shown in Fig. 7. It confirms the previous result

of 5 % variations in the quadrupole transfer function (TF)

during current ramp up and down. So, these variations are

above the specified value of 1 %.

Fig.7. Quadrupole TF variations for the unipolar current cycling in the range

of 0.4 A – 18 A.

At the same time TF hysteresis loops are very reproducible

(see Fig. 8) when the current went up or down. In this case the

reproducibility is better than 0.5 %, and meets specification.

Fig. 8. Quadrupole TF variations for different current ramps.

Fig. 9 and Fig. 10 show expected linear dependence of the

quadrupole magnetic center displacement for different

combinations of the horizontal and vertical dipole corrector

2LPo2C-04

4

currents. It should be noted that dx and dy displacements are

fully decoupled.

Fig. 9. Quadrupole magnetic center displacement at different Horizontal

Dipole Corrector currents.

Fig. 10. Quadrupole magnetic center displacement at different Vertical Dipole

Corrector currents.

The integrated magnetic field quality was also investigated.

The measured magnetic field quadrupole harmonics are less

than 5 units at the reference radius of 10 mm. When the

quadrupole field is combined with the dipole corrector all

harmonics are also less than 5 units except the sextupole

which has a maximum of 110 units at current 20 A. It should

be noted that the maximum needed dipole corrector strength is

reached at 20 % of the quadruple current, in order to

compensate a possible 0.5 mm quadrupole magnetic center

shift caused by magnet installation accuracy and thermal

effects. In this case the sextupole field component will be two

times lower.

During accelerator operations each magnet will operate at its

fixed nominal operating current. Because some SCRF cavities

might be turned off, a 20 % quadrupole strength adjustment

might be needed. Fig. 11 and Fig. 12 show that 20 % magnet

strength change causes less than 0.4 % TF variations.

One of the main results of this test is that after an initial

degaussing the field increases in a very reproducible 0.5 %

way to the nominal operational value. Subsequent 20 % field

adjustments up or down result in 0.4 % magnetic field

reproducibility.

Fig. 11. Quadrupole TF change for 20 % current variation in the range of 0.4 A – 0.5 A.

Fig. 12. Quadrupole TF change for 20 % current variation in the range of

4 A – 5 A.

VI. CONCLUSION

Two prototypes of the splittable conduction cooled magnet

packages, and the first production magnet were thoroughly

tested and showed a good performance. Prototype magnets are

now installed in FNAL and JLAB cryomodules. The magnet

package combines a quadrupole with orthogonal dipole

correctors. During cold tests the following features were

observed and verified:

- The field quality and reproducibility are acceptable.

- The field geometric harmonics are low and meet the

specification.

- Both magnets were successfully excited to 30 A

without quench (20 A is the peak operating current).

The successfully completed tests validated the magnet

design and fabrication for use in the LCLS-II SCRF

cryomodule.

ACKNOWLEDGMENT

The authors would like to thank Prof. Akira Yamamoto

(KEK), Chris Adolphsen, and Paul Emma (SLAC) for very

useful discussions. We are very grateful to the SLAC team for

providing and commissioning a regulated power supply, and

to all FNAL Technical Division personnel involved in the

design, fabrication and tests of these magnets.

2LPo2C-04

5

REFERENCES

[1] “Linac Coherent Light Source,” LCLS-II, 2015:

https://portal.slac.stanford.edu/sites/lcls_public/Pages/Default.aspx [2] “International Linear Collider Technical Design Report,” June 2013,

http://www.linearcollider.org/ILC/Publications/Technical-Design-

Report [3] “TESLA Design Report,” Hamburg, DESY 2001-11, 2001.

[4] “Fermilab Proton Improvement Plan,” PIP-II, 2013:

http://www-ad.fnal.gov/proton/PIP/PIP_index.html [5] F. Toral et al, “Design and fabrication study on the TESLA500

superconducting magnets,” IEEE Trans. Applied Superconductivity 12,

282 (2002). [6] V.S. Kashikhin, et al., “Test results of a superconducting quadrupole

model designed for linear accelerator applications,” IEEE Transactions

on Applied Superconductivity, vol. 19, Issue 3, Part 2, June 2009, pp. 1176-1182.

[7] V.S. Kashikhin, N. Andreev, Y. Orlov, D. F. Orris, and M. A. Tartaglia,

“Superconducting Magnets for SCRF Cryomodules at Front End of

Linear Accelerators,” Proceedings of IPAC’10, Kyoto, Japan, 2010, pp.

379-381.

[8] V. S. Kashikhin, N. Andreev, J. Kerby, Y. Orlov, N. Solyak, M. Tartaglia,and G. Velev, “Superconducting splittable quadrupole magnet

for linear accelerators,” IEEE Trans. Appl. Supercond., vol. 22, no. 3, p.

4002904, June 2012. [9] N. Andreev, V. S. Kashikhin, J. Kerby, N. Kimura, M. Takahashi,

M. A. Tartaglia, T. Tosaka, and A. Yamamoto, “Conduction cooling test

of a splittable quadrupole for ILC cryomodules,” IEEE Trans. Appl. Supercond., vol. 23, no. 3, p. 3500305, June, 2013.

[10] N. Kimura, N. Andreev, V. S. Kashikhin, J. Kerby, M. A. Tartaglia, and

A. Yamamoto, “Cryogenic performance of a conduction cooling splittable quadrupole magnet for ILC cryomodules,” Adv. Cryogenics

Engineering, 59A, 2014, pp.407-415.

[11] R. Carcagno, et al., “Magnetic and Thermal Performance of a Conduction-Cooled Splittable Quadrupole,” IEEE Trans. Appl.

Supercond., vol. 24, no. 3, p. 4001604, June, 2014.

[12] V.S. Kashikhin, et al., “Performance of Conduction Cooled Splittable Superconducting Magnet Package for Linear accelerators,” IEEE Trans.

Appl. Supercond., vol. 26, Issue: 4, p. 4103405, 2016.

[13] J. DiMarco, et al.,”Application of PCB and FDM Technologies to Magnetic Measurement Probe System Development,” IEEE Trans. Appl.

Supercond., vol. 23, no. 3, 9000505, June 2013.