For Review Only
Vertical Magnetic Measurement System Commis-sioning and First Measurements of the First Full-Length Prototype
Quadrupole Magnet for the LHC Hi-Lumi Upgrade
Journal: Applied Superconductivity Conference 2018
Manuscript ID Draft
Conference Manuscript Type: Invited Manuscript
Date Submitted by the Author: n/a
Complete List of Authors: Song, Honghai; Brookhaven National Laboratory, Superconducting Magnet Department; DiMarco, Joseph; Fermilab, TD/MSDJain, Animesh; Argonne National Laboratory, Advanced Photon SourceSabbi, GianLuca; LBNL, AFRD-SMPWanderer, Peter; Brookhaven National Lab, Superconducting Magnet DivisionWang, Xiaorong; Lawrence Berkeley National Laboratory, Accelerator Technology and Applied Physics
Keywords:Accelerator magnets < Accelerator magnets: dipoles, quadrupoles, correctors < Superconducting Magnets, Nb3Sn Wire < Niobium-based wires and tapes < Conductors, Superconducting Magnets
draft
For Review Only
> 4LOr3A-02 <
1
Vertical Magnetic Measurement System Commis-sioning and First Measurements of the First Full-
Length Prototype Quadrupole Magnet for the LHC Hi-Lumi Upgrade
H. Song, Senior Member, IEEE, J. DiMarcro, A. Jain, G. Sabbi, P. Wanderer and X. Wang
(Invited)
Abstract—This paper will report the commissioning of the ver-
tical magnetic measurement system and the warm (room temper-ature) and cold (cryogenic) magnetic field measurements of the
first full-length (4.2 m) quadrupole built by the US Accelerator Upgrade Project (AUP) project (formerly the LARP collabora-tion) for the high luminosity upgrade of the Large Hadron Col-
lider at CERN. The magnet, designated MQXFAP2, is a proto-type preceding production for the AUP Accelerator Upgrade Project (AUP). AUP will provide ten 8.4 m superconducting in-
sertion region (IR) quadrupoles for the Hi-Lumi Upgrade. The 8.4 m quadrupoles will be built by assembling two 4.2 m magnets in a single cryostat. Agreement between BNL’s and LBNL’s
warm measurement data has indicated that the upgraded vertical magnetic measurement system at BNL is ready for MQXFA magnet production testing. The cold magnetic measurement has
been performed and preliminary data analysis has been conduct-ed. The measured data indicates that the MQXFAP2 magnet has good magnetic field which meets the preliminary MQXFA mag-
net acceptance criteria.
Index Terms— LARP, AUP, Hi-Lumi, LHC, Nb3Sn, super-
conducting magnets
I. INTRODUCTION
n the framework of the High-Luminosity (HL) upgrade of
the Large Hadron Collider (LHC) to achieve peak luminosi-
ty of 5∙1034 cm-2s-1 and to reach 3000 fb-1 integrated lumi-
nosity in the period of 2024-2026 [1-5], the U.S. AUP (previ-
ously LARP) collaboration and CERN have joined efforts to
develop production 150 mm aperture Nb3Sn quadrupoles for
the LHC interaction regions[6-8]. The HL LHC Accelerator
Upgrade Project (AUP) has been established to fulfill a US
contribution to HL-LHC which will deliver 12 cryostatted as-
semblies, each containing two 4.2 m long Nb3Sn high gradient
quadrupole magnets. They will be components of the triplets
Manuscript receipt and acceptance dates will be inserted here. Acknowledg-
ment of support is placed in this paragraph as well. Consult the IEEE Editorial Style Manual for examples. This work was supported by the IEEE Council on Superconductivity under contract. ABCD-123456789. (Corresponding author: Honghai Song.)
H. Song and P. Wanderer are with Brookhaven National Laboratory, Upton, NY 11973 United States (email: [email protected]). J. DiMacro is with Fermi National Accelerator Laboratory, Batavia, Illinois, United States. A. Jain is with Argonne National Laboratory, 9700 S. Cass Avenue Lemont, IL 60439, USA. G. Sabbi and X. Wang are with Lawrence Berkeley National Laboratory, Berke-ley, CA 94720-8203 United States.
for two LHC insertion regions. To deliver these magnets
(called MQXF), a full range of training and tests on the
MQXF magnets is required to reach the normal operating cur-
rent of 16.47 kA and ultimate current of 17.89 kA, as well as
magnetic field measurement to make sure that the magnetic
field meets functional requirements and acceptance criteria [9-
15]. The vertical superconducting magnet testing facility of
the Superconducting Magnet Division (SMD) at Brookhaven
National Laboratory (BNL) has been significantly upgraded to
perform testing in superfluid He at 1.9 K and 1 bar. This paper
reports new development of the vertical magnetic field meas-
urement facility at BNL and recent magnetic measurement ac-
tivities on the MQXFAP2 magnet including room temperature
and the cold magnetic field measurement at 1.9 K, as well as
comparison with the warm measurement at LBNL and de-
tailed field analysis.
II. THE UPGRADED VERTICAL MAGNETIC MEASUREMENT SYS-
TEM
A. PCB Rotating Probe and Calibration
The magnetic measurements for the MQXFAP2 magnet will
employ Printed Circuit Board (PCB) rotating coil probes as
shown in Figure 1 and the technique has been validated
though the HQ magnetic measurement program[12, 16, 17].
The PCB board length, width and thickness are 425 mm, 95
mm and 4.57 mm respectively. The MQXFAP magnet has 150
I
Figure 1: The magnetic measurement probe with Printed Circuit Board (PCB) rotating coil. (A): the fully assembled PCB rotating coil; (B): the internal PCB board, Probe 220 on the left, and Probe 110 on the right. Note that Probe 110 is above Probe 220 in vertical testing condition
Page 1 of 13 draft
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
For Review Only
2
mm aperture diameter and the reference radius for magnetic
field measurement is 50 mm. The PCB has two circuits (long
and short), 220 mm and 110 mm long (see Figure 1) so they
are called Probe 220 and 110. Their center-to-center distance
is 220 mm. Each has two layers and each layer has 5 different
tracks (one is a spare) side to side across the board. Each track
is 18.55 mm wide and has 12 loops. The loops are combined
in an analog bucking configuration via jumpers on the board
to have Un-Bucked (UB), Dipole-Bucked (DB), and Dipole-
Quadrupole-Bucked (DQB) signals. Each circuit board has
three signals, and signals 1-3 are for Probe 220 (Left in Figure
1) and signals 4-6 are for Probe 110 (Right). The wire re-
sistances for all the 6 signals have been measured as 63.2,
120.6, 235.0, 34.7, 65.0 and 125.4 ohms. Note that the PCB
rotating probe will be vertically placed inside the magnet bore
centered by the support bearings, where Probe 110 is above
Probe 220 in the vertical testing condition.
With all the geometrical information, the sensitivity matri-
ces of Probes 220 and 110 are calculated as input for the rotat-
ing coil signals process. The sensitivities for the quad term are
6.36E-2 m2 in long Probe 220, and 3.17E-2 m2 in short Probe
110. The MQXF magnet has 150 mm inner diameter and a
gradient of 143.2 T/m at the nominal current of 16470 A, re-
sulting in flux in UB signal B2*K2 = 7.16 * 6.36E-2 = 4.55
Vs in Probe 220. With the probe rotating at 1/3.5 Hz, the UB
signal is ~1.63 V. Similarly, the UB signal in Probe 110 is es-
timated at 0.817 V. Furthermore, the PCB coil was calibrated
in a calibration quadrupole at the Superconducting Magnet
Division at BNL. The calibration magnet at BNL was powered
at ~1000 A, and the field is ~0.167 T at a radius of 50 mm.
The measured UB signals in long Probe 220 and short Probe
110 are 0.039 and 0.019 V. The calibration measurement con-
firmed that both circuits measure the magnetic field accurate-
ly, and the consistency between UB and DB signal within each
coil is within +/-0.03%.
B. Vertical Transport System Development
As the MXQF AP2 magnet is ~4563 mm long and the
probes are only 220 m and 110 mm long, one of the key ef-
forts is to transport the PCB rotating probe to scan the field
along the magnet axis with a vertical linear motion transport.
Figure 2 shows the magnet vertical length, Probe 220 and 110
locations and their center-center distance, which is about 220
mm. Considering the detailed design of the PCB, the actual
magnetic length is 108.74 mm for Probe 110, and 217.88 mm
for Probe 220. With Probe 220 as a primary rotating coil and
Probe 110 recorded in parallel, the measurement increment
along the magnet axis (vertical z direction) is 108.94 mm
which is nearly the Rutherford cable twist pitch length. As the
Probe 220 is lower than Probe 110, its first measured point is
220 mm closer to the magnet return end than probe 110.
A LabVIEW based vertical transport motion system has
been developed to transport the rotating PCB probe accurately
and safely. A survey has been made to check the axial motion
accuracy. As shown in Figure 3, the motion has accuracy of
0.6 mm which satisfies the MQXFAP magnet field measure-
ment requirement. To implement the ZSCAN, a script file
which includes all the Z positions will be loaded to the pro-
gram and start the motion. Meanwhile, a HTBasic system can
take a single magnetic field measurement. A hand-shake func-
tion has been established to connect the LabVIEW and
HTBasic systems. The system will take the measurement at
the first Z, and then move to the next Z while the HTBasic
program is waiting. Once the next Z is reached and confirmed,
the next measurement will take place, and so on until the last
Z position, indicated by EOF, when the ZSCAN stops. In ad-
dition, the LabVIEW system online monitoring of three phase
AC status and a real-time watch-dog checks the system to
make sure that the electrical supplies and computer operating
system are working functionally. If any of them malfunctions,
the LabVIEW program will stop running and exit automatical-
ly. Most importantly, the E-Stop hardware is incorporated in
the motion system, so any error message or system malfunc-
tion will stop the motor instantly
4753.61 mm
[187.142 in]
Top of Magnet Shell
Bottom of Warm Bore Tube
Bottom of Magnet Shell
4563 mm
[179.646 in]
Top of Warm Bore Tube
6256.57 mm
[246.322 in]
903.86
[35.585]
Top hat to Lambda
plateRotating coil in
MAX UP Post
656
25.827
Rotating coil
~220 mm
200.8 mm
Home
4
7
3
0
m
m
190.61mm
164.99mm
331.99mm
Switch_UL
Figure 2: A schematic for the MQXFAP magnet and the rotating probe. The total magnet length is 4,563 mm and the probes are only 220 and 110 mm long. The total vertical linear stroke between HOME and the upper limit switch is 4,730 mm at the maximum. When the rotating probe is at HOME, the distance between the probe bottom and the warm bore tube (upper surface) is 355.6 mm, and the distance between the probe bottom and the bottom of magnet shell is 164.99 mm. Limit switches have been placed at the very bottom and top for motion safety.
Figure 3: The LabVIEW-based vertical transport motion for the MQXFAP Magnet. A survey has been conducted to check the motion accuracy. One set of measurements was made from HOME to the further end (1a), and another set forwards (2a) and backwards (2b).
Page 2 of 13draft
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
For Review Only
3
III. MAGNETIC MEASUREMENTS
A. Warm Measurements
The magnetic field in the aperture of the straight section of
a quadrupole magnet can be expressed in terms of field coeffi-
cients in a series expansion in the following complex function
formalism [18]:
(1)
where Bx and By are the horizontal and vertical field compo-
nents in T in Cartesian coordinates, Bn and An are the normal
and skew multipole fields at the reference radius, which is Rref
50 mm. The warm measurement was performed at LBNL in
June 2018 prior to delivery to BNL for vertical magnet testing.
In late August, the warm magnetic measurement was per-
formed at BNL before the magnet cool-down. The warm
ZSCAN measurement has 42 points starting from HOME
(lowest), with an increment of 108.94 mm. Harmonics up to n
of 15 have been measured. To eliminate the remnant field due
to the background field and iron contribution, separate meas-
urements at +/- 15 A currents have been made allowing sepa-
ration of the remnant field from the field due to magnet’s
coils. The transfer function has been measured and plotted
along the Z direction (magnet axis). The BNL and LBNL
transfer function data are plotted in Figure 4 for direct com-
parison. Both are very close to the design transfer function of
8.86 T/m/kA at room temperature. The two curves agree very
well along the magnet axis. Note that there is no BNL data at
the return end due to the limited length of the underground
Dewar.
B. Cold Measurement at 1.9 K and 4.5 K
The quench training test of the second MQXFA prototype
(MQXFAP2) is in progress at BNL Vertical Test Facility. The
magnet reached 15 kA in 9 quenches and showed detraining
Figure 5: Harmonics b3/a3, b6/a6 of the first +/-15 A warm magnetic field measurement before the magnet cooldown. Good agreement between LBL and BNL data has been confirmed.
Figure 4: The averaged +/-15 A warm magnetic field measurement prior to the magnet cooling-down. There is good agreement between LBNL and BNL warm measurements.
Figure 7: Transfer functions and b6 of currents in the stair-step loop measurements.
Figure 6: Transfer functions at 15 A (averaged, warm), 960 A, 6kA and 10 kA in cold MM in the ZSCAN.
Figure 8: Center offset found in the ZSCAN. The X direction has less offset than the Y direction. The higher the current is, the more the Y offset becomes..
Page 3 of 13 draft
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
For Review Only
4
after quench 13. The quench training was interrupted to dis-
cuss magnet performance and test continuation which offered
an opportunity for the cold magnetic field measurements. The
cold measurement has ZSCAN like that in the warm meas-
urement, and ISCAN which is also called stair-step (or DC
loop) measurement. The ZSCAN was performed at currents of
Iinj (960 A) at 1.9 K, Ilim (6kA) at ~2 K, 10 kA at 4.5 K. The
temperature change was due to a temporary problem with the
cryo facility. The short MQXF magnet data has indicated that
the temperature dependence of magnetic fields is not signifi-
cant[12]. The precyle for 960 A measurement was a ramp up
to 10 kA (300 second dwell) and down to Ires (100A), and up
to Iinj (960 A, 1000 second dwell).
The transfer functions (T.F.) of all the three currents are
plotted along with comparison to that of averaged 15 A as
shown in Figure 6. The magnet has highest T.F. at Iinj (960A),
and slightly decreases as current increases due to iron satura-
tion. It agrees with the findings in the short MQXF quadru-
poles. In the ISCAN as plotted in Figure 7, both the T.F. and
the allowed coefficient b6 present considerable hysteresis
which is caused by persistent current. In analyzing the cold
measurement data, it is found that offset occurred between the
PCB probe center and the magnet center as shown in Figure 8.
The Y direction has more offset than the X direction. It has the
largest offset at the lower return end and smallest at the lead
end. Moreover, it increases as the current increases. The warm
bore tube (WBT) which contains the PCB rotating probe may
have tilted during the cool down. This may trace back to the
fact both the WBT and rotating probe are suspended from the
top. It seems more centering is necessary though detailed in-
vestigation is underway. While all the harmonics analysis in
Figure 7 are corrected for the effects of the center offset be-
tween the rotating PCB probe and the magnet, a smaller offset
is still preferred to allow accurate corrections particularly for
high order multipole coefficients.
Furthermore, the cold multipole coefficients of 960 A, 6kA
and 10 kA are plotted along the Z axis, directly compared to
the warm coefficients of 15 A in the ZSCAN as shown in Fig-
ure 9-11. Most of them agree well with each other. The slight
difference in b4 and b5 may be related to the geometric
Figure 9: Nonallowed b3 and a3 at currents of 960 A at 1.9 K, 6 kA at ~2 K, 10 kA at 4.5 K, with comparison to that of averaged 15 A at room temperature.
Figure 11: Allowed b6 and nonallowed a6 at currents of 960 A at 1.9 K, 6 kA at ~2 K, 10 kA at 4.5 K, with comparison to that of averaged 15 A at room temperature.
Figure 10: Nonallowed b5 and a5 at currents of 960 A at 1.9 K, 6 kA at ~2 K, 10 kA at 4.5 K, with comparison to that of averaged 15 A at room temperature.
Page 4 of 13draft
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
For Review Only
5
shrinkage during cooling down. The change in b6 is due to a
combination of iron saturation and persistent currents. The
measured field quality confirms the necessity of having a fine
tuning of the cross-section to increase b6 by about 4 units[19,
20]. A decision has been recently made and shall be imple-
mented for the future magnets through a change of 0.125 mm
shims in the pole and in the midplane. On the other hand, the
non-allowed harmonic that showed large values in the short
model program appear to be close to the targets.
TABLE I
Harmonics bn (n≤6) of MQXFAP2 at 10 kA and 4.5 K.
b3 b4 b5 b6
Current mean rms mean rms mean rms Mean Rms
15A -1.23 1.96 -1.12 1.25 -0.46 0.95 -5.90 1.12
960A -0.92 2.02 -0.91 1.30 -1.38 1.02 -17.03 1.28
6kA -0.67 2.07 -1.11 1.28 -1.12 1.06 -7.37 1.28
10kA -0.87 2.09 -1.25 1.33 -0.99 1.06 -4.18 1.26
TABLE II
Harmonics an (n≤6)) of MQXFAP2 at 10 kA and 4.5 K.
a3 a4 a5 a6
Current mean rms mean rms mean rms mean rms
15A 4.10 1.77 2.61 1.69 1.94 0.78 0.45 0.49
960A 4.19 1.77 3.07 1.83 1.75 0.82 0.47 0.49
6kA 2.79 1.82 2.24 1.96 1.35 0.83 0.40 0.52
10kA 3.17 1.85 2.54 1.95 1.48 0.85 0.40 0.52
The cold and warm harmonics (bn and an, n≤6) of 10 kA
and 15 A in the straight section have been averaged and plot-
ted against each other in Y and X axis as shown in Figure 12.
Most the data points are next to the 45 deg angle line, which
indicates a correlation between the cold and warm field data.
The b6 has the largest value of -5.9 units for 15 A and -4.18
units for 10 kA. The a3 has the largest value of 4.10 units for
15 A and 3.17 units for 10 kA as summarized in TABLE I and
II.
As Probe 220 is twice as long as Probe 110, the Probe 110
is expected to have better resolution with ~110 mm step. Data
of the two probes have been compared in Figure 13. Their
starting points are off by ~220 mm. For the allowed harmonic
Probe 220 Probe 110
Probe 220
Probe 110
Figure 13: Field comparison between Probe 220 and 110.
Figure 12: Cold and warm harmonics correlation for the field in the straight section (Probe 220) between -1.915 m and 1.642 m, 10 kA vs 15 A.
Page 5 of 13 draft
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
For Review Only
6
b6 with 10 – 20 units, their difference is not significant. For
the nonallowed harmonics a6, b4 and a4, the data profiles of
Probe 110 at 15 A and 960 are sharper than those of Probe
220.
IV. CONCLUSIONS
The vertical magnetic measurement system which includes the
newly developed vertical motion transport system has been
commissioned for the magnetic measurement of the
MQXFAP2 magnet. The good agreement between the BNL
and LBNL data in the warm magnetic measurements of aver-
age 15 A indicates that the PCB rotating coil functions well
for the full-length MQXFA production magnets. Furthermore,
the system has been used for the cold measurement and col-
lected ZSCAN and ISCAN (stair-step) data. A preliminary
analysis of the Probe 220 data, corrected for the center offset
between the probe and the magnet, shows good agreement be-
tween the warm and cold measurements. The Probe 110 does
have better resolution than Probe 220 which will be very use-
ful for measurement of the lead end. The measurement indi-
cates that the MQXFAP2 magnet has magnetic field which
meets the preliminary MQXFA magnet acceptance criteria.
The large offset of the measuring coil at the magnet return end
is likely because both the WBT and the probe are suspended
from the top flange. Improvement to better center the rotating
probe in the vertical magnet testing facility seem necessary.
Further investigation of the offset is underway.
ACKNOWLEDGEMENT
The authors thank the magnet testing team at BNL for the
measurement setup and preparations, and general support from
FermiLab and LBNL and thank G. Ambrosio, G. Apollinari,
K. Amm, M. Anerella, S. I. Bermudez, S. Feher, P. Joshi, J.
Muratore, E. Todesco for helpful discussions and suggestions.
This work was supported in part by the U.S. Department of
Energy, Office of Science, Office of High Energy Physics,
through the U.S. LHC Accelerator Research Program, and in
part by the High Luminosity LHC project at CERN.
REFERENCE
1. Ferracin, P., et al., The HL-LHC low-b quadrupole
magnet MQXF: from short models to long
prototypes, Presentation Number: 4LOr3A-01, in The
Applied Superconductivity Conference. 2018, IEEE
Trans. Applied Superconductivity Seattle.
2. Todesco, E., et al., Progress on HL-LHC Nb3Sn
Magnets. Ieee Transactions on Applied
Superconductivity, 2018. 28(4).
3. Rossi, L., The High Luminosity LHC Project. Future
of Our Physics Including New Frontiers, 2017. 53.
4. Todesco, E., et al., A First Baseline for the Magnets
in the High Luminosity LHC Insertion Regions. Ieee
Transactions on Applied Superconductivity, 2014.
24(3).
5. Bottura, L., et al., Advanced Accelerator Magnets for
Upgrading the LHC. Ieee Transactions on Applied
Superconductivity, 2012. 22(3).
6. Ambrosio, G., Nb3Sn High Field Magnets for the
High Luminosity LHC Upgrade Project. Ieee
Transactions on Applied Superconductivity, 2015.
25(3).
7. Apollinari, G., S. Prestemon, and A.V. Zlobin,
Progress with High-Field Superconducting Magnets
for High-Energy Colliders. Annual Review of
Nuclear and Particle Science, Vol 65, 2015. 65: p.
355-377.
8. Wanderer, P., Overview of LARP Magnet R&D. Ieee
Transactions on Applied Superconductivity, 2009.
19(3): p. 1208-1211.
9. Wang, X.R., et al., Analysis of Field Errors for LARP
Nb3Sn HQ03 Quadrupole Magnet. Ieee Transactions
on Applied Superconductivity, 2017. 27(4).
10. Fiscarelli, L., et al., Magnetic Measurements on the
First CERN-Built Models of the Insertion
Quadrupole MQXF for HL-LHC. Ieee Transactions
on Applied Superconductivity, 2018. 28(3).
11. Holik, E.F., et al., Fabrication of First 4-m Coils for
the LARP MQXFA Quadrupole and Assembly in
Mirror Structure. Ieee Transactions on Applied
Superconductivity, 2017. 27(4).
12. DiMarco, J., et al., Magnetic Measurements of the
First Nb3Sn Model Quadrupole (MQXFS) for the
High-Luminosity LHC. Ieee Transactions on Applied
Superconductivity, 2017. 27(4).
13. Sorbi, M., et al., Measurements and Analysis of
Dynamic Effects in the LARP Model Quadrupole
HQ02b During Rapid Discharge. Ieee Transactions
on Applied Superconductivity, 2016. 26(4).
14. Sabbi, G., Nb3Sn IR Quadrupoles for the High
Luminosity LHC. Ieee Transactions on Applied
Superconductivity, 2013. 23(3).
15. Apollinari, G., et al. High-Luminosity Large Hadron
Collider (HL-LHC) Technical Design Report V.0.1 in
CERN Yellow Reports: Monographs. 2017. Geneva:
CERN, https://doi.org/10.23731/CYRM-2017-004
16. DiMarco, J., et al., Application of PCB and FDM
Technologies to Magnetic Measurement Probe
System Development. Ieee Transactions on Applied
Superconductivity, 2013. 23(3).
17. DiMarco, J., et al., Field Quality Measurements of
LARP Nb3Sn Magnet HQ02. Ieee Transactions on
Applied Superconductivity, 2014. 24(3).
18. Jain, A. Basic theory of magnets. in CERN
Accelerator School Meas. Alignment Accel. Detect.
Magnets, no. CERN-98-05. 1998.
19. Todesco, E., HL-LHC Decision management: WP3 -
fine tuning of b6,
https://edms.cern.ch/document/2019517/4. 2018.
20. Bermudez, S.I., et al., Magnetic Analysis of the
MQXF Quadrupole for the High Luminosity LHC:
4LOr3A-04, in The Applied Superconductivity
Page 6 of 13draft
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
For Review Only
7
Conference. 2018, IEEE Trans. Applied
Superconductivity Seattle.
Page 7 of 13 draft
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
For Review Only
> 4LOr3A-02 <
1
Vertical Magnetic Measurement System Commissioning and First Measurements of the First Full-Length Prototype Quadrupole Magnet for the
LHC Hi-Lumi UpgradeH. Song, Senior Member, IEEE, J. DiMarcro, A. Jain, G. Sabbi, P. Wanderer and X. Wang
(Invited)
Abstract—This paper will report the commissioning of the vertical magnetic measurement system and the warm (room temperature) and cold (cryogenic) magnetic field measurements of the first full-length (4.2 m) quadrupole built by the US Accelerator Upgrade Project (AUP) project (formerly the LARP collaboration) for the high luminosity upgrade of the Large Hadron Collider at CERN. The magnet, designated MQXFAP2, is a prototype preceding production for the AUP Accelerator Upgrade Project (AUP). AUP will provide ten 8.4 m superconducting insertion region (IR) quadrupoles for the Hi-Lumi Upgrade. The 8.4 m quadrupoles will be built by assembling two 4.2 m magnets in a single cryostat. Agreement between BNL’s and LBNL’s warm measurement data has indicated that the upgraded vertical magnetic measurement system at BNL is ready for MQXFA magnet production testing. The cold magnetic measurement has been performed and preliminary data analysis has been conducted. The measured data indicates that the MQXFAP2 magnet has good magnetic field which meets the preliminary MQXFA magnet acceptance criteria.
Index Terms— LARP, AUP, Hi-Lumi, LHC, Nb3Sn, superconducting magnets
I. INTRODUCTION
n the framework of the High-Luminosity (HL) upgrade of the Large Hadron Collider (LHC) to achieve peak luminosity of 5∙1034 cm-2s-1 and to reach 3000 fb-1
integrated luminosity in the period of 2024-2026 [1-5], the U.S. AUP (previously LARP) collaboration and CERN have joined efforts to develop production 150 mm aperture Nb3Sn quadrupoles for the LHC interaction regions[6-8]. The HL LHC Accelerator Upgrade Project (AUP) has been established to fulfill a US contribution to HL-LHC which will deliver 12 cryostatted assemblies, each containing two 4.2 m long Nb3Sn
Manuscript receipt and acceptance dates will be inserted here. Acknowledgment of support is placed in this paragraph as well. Consult the IEEE Editorial Style Manual for examples. This work was supported by the IEEE Council on Superconductivity under contract. ABCD-123456789. (Corresponding author: Honghai Song.)
H. Song and P. Wanderer are with Brookhaven National Laboratory, Upton, NY 11973 United States (email: [email protected]). J. DiMacro is with Fermi National Accelerator Laboratory, Batavia, Illinois, United States. A. Jain is with Argonne National Laboratory, 9700 S. Cass Avenue Lemont, IL 60439, USA. G. Sabbi and X. Wang are with Lawrence Berkeley National Laboratory, Berkeley, CA 94720-8203 United States.
high gradient quadrupole magnets. They will be components of the triplets for two LHC insertion regions. To deliver these magnets (called MQXF), a full range of training and tests on the MQXF magnets is required to reach the normal operating current of 16.47 kA and ultimate current of 17.89 kA, as well as magnetic field measurement to make sure that the magnetic field meets functional requirements and acceptance criteria [9-15]. The vertical superconducting magnet testing facility of the Superconducting Magnet Division (SMD) at Brookhaven National Laboratory (BNL) has been significantly upgraded to perform testing in superfluid He at 1.9 K and 1 bar. This paper reports new development of the vertical magnetic field measurement facility at BNL and recent magnetic measurement activities on the MQXFAP2 magnet including room temperature and the cold magnetic field measurement at 1.9 K, as well as comparison with the warm measurement at LBNL and detailed field analysis.
II. THE UPGRADED VERTICAL MAGNETIC MEASUREMENT SYSTEM
A. PCB Rotating Probe and Calibration
The magnetic measurements for the MQXFAP2 magnet will employ Printed Circuit Board (PCB) rotating coil probes as shown in Figure 1 and the technique has been validated though the HQ magnetic measurement program[12, 16, 17]. The PCB board length, width and thickness are 425 mm, 95
I
Figure 1: The magnetic measurement probe with Printed Circuit Board (PCB) rotating coil. (A): the fully assembled PCB rotating coil; (B): the internal PCB board, Probe 220 on the left, and Probe 110 on the right. Note that Probe 110 is above Probe 220 in vertical testing condition
Page 8 of 13draft
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
For Review Only
2
mm and 4.57 mm respectively. The MQXFAP magnet has 150 mm aperture diameter and the reference radius for magnetic field measurement is 50 mm. The PCB has two circuits (long and short), 220 mm and 110 mm long (see Figure 1) so they are called Probe 220 and 110. Their center-to-center distance is 220 mm. Each has two layers and each layer has 5 different tracks (one is a spare) side to side across the board. Each track is 18.55 mm wide and has 12 loops. The loops are combined in an analog bucking configuration via jumpers on the board to have Un-Bucked (UB), Dipole-Bucked (DB), and Dipole-Quadrupole-Bucked (DQB) signals. Each circuit board has three signals, and signals 1-3 are for Probe 220 (Left in Figure 1) and signals 4-6 are for Probe 110 (Right). The wire resistances for all the 6 signals have been measured as 63.2, 120.6, 235.0, 34.7, 65.0 and 125.4 ohms. Note that the PCB rotating probe will be vertically placed inside the magnet bore centered by the support bearings, where Probe 110 is above Probe 220 in the vertical testing condition.
With all the geometrical information, the sensitivity matrices of Probes 220 and 110 are calculated as input for the rotating coil signals process. The sensitivities for the quad term are 6.36E-2 m2 in long Probe 220, and 3.17E-2 m2 in short Probe 110. The MQXF magnet has 150 mm inner diameter and a gradient of 143.2 T/m at the nominal current of 16470 A, resulting in flux in UB signal B2*K2 = 7.16 * 6.36E-2 = 4.55 Vs in Probe 220. With the probe rotating at 1/3.5 Hz, the UB signal is ~1.63 V. Similarly, the UB signal in Probe 110 is estimated at 0.817 V. Furthermore, the PCB coil was calibrated in a calibration quadrupole at the Superconducting Magnet Division at BNL. The calibration magnet at BNL was powered at ~1000 A, and the field is ~0.167 T at a radius of 50 mm. The measured UB signals in long Probe 220 and short Probe 110 are 0.039 and 0.019 V. The calibration measurement confirmed that both circuits measure the magnetic field accurately, and the consistency between UB and DB signal within each coil is within +/-0.03%.
B. Vertical Transport System DevelopmentAs the MXQF AP2 magnet is ~4563 mm long and the
probes are only 220 m and 110 mm long, one of the key efforts is to transport the PCB rotating probe to scan the field along the magnet axis with a vertical linear motion transport. Figure 2 shows the magnet vertical length, Probe 220 and 110 locations and their center-center distance, which is about 220 mm. Considering the detailed design of the PCB, the actual magnetic length is 108.74 mm for Probe 110, and 217.88 mm for Probe 220. With Probe 220 as a primary rotating coil and Probe 110 recorded in parallel, the measurement increment along the magnet axis (vertical z direction) is 108.94 mm which is nearly the Rutherford cable twist pitch length. As the Probe 220 is lower than Probe 110, its first measured point is 220 mm closer to the magnet return end than probe 110.
A LabVIEW based vertical transport motion system has been developed to transport the rotating PCB probe accurately and safely. A survey has been made to check the axial motion accuracy. As shown in Figure 3, the motion has accuracy of 0.6 mm which satisfies the MQXFAP magnet field measurement requirement. To implement the ZSCAN, a script file which includes all the Z positions will be loaded to the program and start the motion. Meanwhile, a HTBasic system can take a single magnetic field measurement. A hand-shake function has been established to connect the LabVIEW and HTBasic systems. The system will take the measurement at the first Z, and then move to the next Z while the HTBasic program is waiting. Once the next Z is reached and confirmed, the next measurement will take place, and so on until the last Z position, indicated by EOF, when the ZSCAN stops. In addition, the LabVIEW system online monitoring of three phase AC status and a real-time watch-dog checks the system to make sure that the electrical supplies and computer operating system are working functionally. If any of them malfunctions, the LabVIEW program will stop running and exit automatically. Most importantly, the E-Stop hardware is
4753.61 mm [187.142 in]
Top of Magnet Shell
Bottom of Warm Bore TubeBottom of Magnet Shell
4563 mm [179.646 in]
Top of Warm Bore Tube
6256.57 mm [246.322 in]
903.86[35.585]
Top hat to Lambda plate
Rotating coil in MAX UP Post
65625.827
Rotating coil
~220 mm
200.8 mm
Home
4730mm
190.61mm164.99mm
331.99mm
Switch_UL
Figure 2: A schematic for the MQXFAP magnet and the rotating probe. The total magnet length is 4,563 mm and the probes are only 220 and 110 mm long. The total vertical linear stroke between HOME and the upper limit switch is 4,730 mm at the maximum. When the rotating probe is at HOME, the distance between the probe bottom and the warm bore tube (upper surface) is 355.6 mm, and the distance between the probe bottom and the bottom of magnet shell is 164.99 mm. Limit switches have been placed at the very bottom and top for motion safety.
Figure 3: The LabVIEW-based vertical transport motion for the MQXFAP Magnet. A survey has been conducted to check the motion accuracy. One set of measurements was made from HOME to the further end (1a), and another set forwards (2a) and backwards (2b).
Page 9 of 13 draft
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
For Review Only
3
incorporated in the motion system, so any error message or system malfunction will stop the motor instantly
III.MAGNETIC MEASUREMENTS
A. Warm MeasurementsThe magnetic field in the aperture of the straight section of
a quadrupole magnet can be expressed in terms of field coefficients in a series expansion in the following complex function formalism [18]:
(1)
where Bx and By are the horizontal and vertical field components in T in Cartesian coordinates, Bn and An are the normal and skew multipole fields at the reference radius, which is Rref 50 mm. The warm measurement was performed at LBNL in June 2018 prior to delivery to BNL for vertical magnet testing. In late August, the warm magnetic measurement was performed at BNL before the magnet cool-down. The warm ZSCAN measurement has 42 points starting from HOME (lowest), with an increment of 108.94 mm. Harmonics up to n of 15 have been measured. To eliminate the remnant field due to the background field and iron contribution, separate measurements at +/- 15 A currents have been made allowing separation of the remnant field from the
field due to magnet’s coils. The transfer function has been measured and plotted along the Z direction (magnet axis). The BNL and LBNL transfer function data are plotted in Figure 4 for direct comparison. Both are very close to the design transfer function of 8.86 T/m/kA at room temperature. The two curves agree very well along the magnet axis. Note that there is no BNL data at the return end due to the limited length of the underground Dewar.
B. Cold Measurement at 1.9 K and 4.5 K
Figure 5: Harmonics b3/a3, b6/a6 of the first +/-15 A warm magnetic field measurement before the magnet cooldown. Good agreement between LBL and BNL data has been confirmed.
Figure 4: The averaged +/-15 A warm magnetic field measurement prior to the magnet cooling-down. There is good agreement between LBNL and BNL warm measurements.
Figure 7: Transfer functions and b6 of currents in the stair-step loop measurements.
Figure 8: Center offset found in the ZSCAN. The X direction has less offset than the Y direction. The higher the current is, the more the Y offset becomes..
Page 10 of 13draft
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
For Review Only
4
The quench training test of the second MQXFA prototype (MQXFAP2) is in progress at BNL Vertical Test Facility. The magnet reached 15 kA in 9 quenches and showed detraining after quench 13. The quench training was interrupted to discuss magnet performance and test continuation which offered an opportunity for the cold magnetic field measurements. The cold measurement has ZSCAN like that in the warm measurement, and ISCAN which is also called stair-step (or DC loop) measurement. The ZSCAN was performed at currents of Iinj (960 A) at 1.9 K, Ilim (6kA) at ~2 K, 10 kA at
4.5 K. The temperature change was due to a temporary problem with the cryo facility. The short MQXF magnet data has indicated that the temperature dependence of magnetic fields is not significant[12]. The precyle for 960 A
measurement was a ramp up to 10 kA (300 second dwell) and down to Ires (100A), and up to Iinj (960 A, 1000 second dwell).
The transfer functions (T.F.) of all the three currents are plotted along with comparison to that of averaged 15 A as shown in Figure 6. The magnet has highest T.F. at Iinj (960A), and slightly decreases as current increases due to iron saturation. It agrees with the findings in the short MQXF quadrupoles. In the ISCAN as plotted in Figure 7, both the T.F. and the allowed coefficient b6 present considerable hysteresis which is caused by persistent current. In analyzing the cold measurement data, it is found that offset occurred between the PCB probe center and the magnet center as shown in Figure 8. The Y direction has more offset than the X direction. It has the largest offset at the lower return end and smallest at the lead end. Moreover, it increases as the current increases. The warm bore tube (WBT) which contains the PCB rotating probe may have tilted during the cool down. This may trace back to the fact both the WBT and rotating probe are suspended from the top. It seems more centering is necessary though detailed investigation is underway. While all the harmonics analysis in Figure 7 are corrected for the effects of the center offset between the rotating PCB probe and the magnet, a smaller offset is still preferred to allow accurate corrections particularly for high order multipole coefficients.
Figure 6: Transfer functions at 15 A (averaged, warm), 960 A, 6kA and 10 kA in cold MM in the ZSCAN.
Figure 9: Nonallowed b3 and a3 at currents of 960 A at 1.9 K, 6 kA at ~2 K, 10 kA at 4.5 K, with comparison to that of averaged 15 A at room temperature.
Figure 10: Nonallowed b5 and a5 at currents of 960 A at 1.9 K, 6 kA at ~2 K, 10 kA at 4.5 K, with comparison to that of averaged 15 A at room temperature.
Page 11 of 13 draft
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
For Review Only
5
Furthermore, the cold multipole coefficients of 960 A, 6kA and 10 kA are plotted along the Z axis, directly compared to the warm coefficients of 15 A in the ZSCAN as shown in Figure 9-11. Most of them agree well with each other. The slight difference in b4 and b5 may be related to the geometric shrinkage during cooling down. The change in b6 is due to a combination of iron saturation and persistent currents. The measured field quality confirms the necessity of having a fine tuning of the cross-section to increase b6 by about 4 units[19, 20]. A decision has been recently made and shall be implemented for the future magnets through a change of 0.125 mm shims in the pole and in the midplane. On the other hand, the non-allowed harmonic that showed large values in the short model program appear to be close to the targets.
TABLE I
Harmonics bn (n≤6) of MQXFAP2 at 10 kA and 4.5 K.
b3 b4 b5 b6
Current mean rms mean rms mean rms Mean Rms
15A -1.23 1.96 -1.12 1.25 -0.46 0.95 -5.90 1.12
960A -0.92 2.02 -0.91 1.30 -1.38 1.02 -17.03 1.28
6kA -0.67 2.07 -1.11 1.28 -1.12 1.06 -7.37 1.28
10kA -0.87 2.09 -1.25 1.33 -0.99 1.06 -4.18 1.26
TABLE II
Harmonics an (n≤6)) of MQXFAP2 at 10 kA and 4.5 K.
a3 a4 a5 a6
Current mean rms mean rms mean rms mean rms
15A 4.10 1.77 2.61 1.69 1.94 0.78 0.45 0.49
960A 4.19 1.77 3.07 1.83 1.75 0.82 0.47 0.49
6kA 2.79 1.82 2.24 1.96 1.35 0.83 0.40 0.52
10kA 3.17 1.85 2.54 1.95 1.48 0.85 0.40 0.52
The cold and warm harmonics (bn and an, n≤6) of 10 kA and 15 A in the straight section have been averaged and plotted against each other in Y and X axis as shown in Figure 12. Most the data points are next to the 45 deg angle line, which indicates a correlation between the cold and warm field data. The b6 has the largest value of -5.9 units for 15 A and -4.18 units for 10 kA. The a3 has the largest value of 4.10 units for 15 A and 3.17 units for 10 kA as summarized in TABLE I and II.
As Probe 220 is twice as long as Probe 110, the Probe 110 is expected to have better resolution with ~110 mm step. Data of the two probes have been compared in Figure 13. Their starting points are off by ~220 mm. For the allowed harmonic b6 with 10 – 20 units, their difference is not significant. For the nonallowed harmonics a6, b4 and a4, the data profiles of Probe 110 at 15 A and 960 are sharper than those of Probe 220.
Figure 11: Allowed b6 and nonallowed a6 at currents of 960 A at 1.9 K, 6 kA at ~2 K, 10 kA at 4.5 K, with comparison to that of averaged 15 A at room temperature.
Probe 220 Probe 110
Probe 220
Probe 110
Figure 13: Field comparison between Probe 220 and 110.
Figure 12: Cold and warm harmonics correlation for the field in the straight section (Probe 220) between -1.915 m and 1.642 m, 10 kA vs 15 A.
Page 12 of 13draft
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
For Review Only
6
IV. CONCLUSIONS
The vertical magnetic measurement system which includes the newly developed vertical motion transport system has been commissioned for the magnetic measurement of the MQXFAP2 magnet. The good agreement between the BNL and LBNL data in the warm magnetic measurements of average 15 A indicates that the PCB rotating coil functions well for the full-length MQXFA production magnets. Furthermore, the system has been used for the cold measurement and collected ZSCAN and ISCAN (stair-step) data. A preliminary analysis of the Probe 220 data, corrected for the center offset between the probe and the magnet, shows good agreement between the warm and cold measurements. The Probe 110 does have better resolution than Probe 220 which will be very useful for measurement of the lead end. The measurement indicates that the MQXFAP2 magnet has magnetic field which meets the preliminary MQXFA magnet acceptance criteria. The large offset of the measuring coil at the magnet return end is likely because both the WBT and the probe are suspended from the top flange. Improvement to better center the rotating probe in the vertical magnet testing facility seem necessary. Further investigation of the offset is underway.
ACKNOWLEDGEMENT
The authors thank the magnet testing team at BNL for the measurement setup and preparations, and general support from FermiLab and LBNL and thank G. Ambrosio, G. Apollinari, K. Amm, M. Anerella, S. I. Bermudez, S. Feher, P. Joshi, J. Muratore, E. Todesco for helpful discussions and suggestions. This work was supported in part by the U.S. Department of Energy, Office of Science, Office of High Energy Physics, through the U.S. LHC Accelerator Research Program, and in part by the High Luminosity LHC project at CERN.
REFERENCE
1. Ferracin, P., et al., The HL-LHC low-b quadrupole magnet MQXF: from short models to long prototypes, Presentation Number: 4LOr3A-01, in The Applied Superconductivity Conference. 2018, IEEE Trans. Applied Superconductivity Seattle.
2. Todesco, E., et al., Progress on HL-LHC Nb3Sn Magnets. Ieee Transactions on Applied Superconductivity, 2018. 28(4).
3. Rossi, L., The High Luminosity LHC Project. Future of Our Physics Including New Frontiers, 2017. 53.
4. Todesco, E., et al., A First Baseline for the Magnets in the High Luminosity LHC Insertion Regions. Ieee Transactions on Applied Superconductivity, 2014. 24(3).
5. Bottura, L., et al., Advanced Accelerator Magnets for Upgrading the LHC. Ieee Transactions on Applied Superconductivity, 2012. 22(3).
6. Ambrosio, G., Nb3Sn High Field Magnets for the High Luminosity LHC Upgrade Project. Ieee Transactions on Applied Superconductivity, 2015. 25(3).
7. Apollinari, G., S. Prestemon, and A.V. Zlobin, Progress with High-Field Superconducting Magnets for High-Energy Colliders. Annual Review of Nuclear and Particle Science, Vol 65, 2015. 65: p. 355-377.
8. Wanderer, P., Overview of LARP Magnet R&D. Ieee Transactions on Applied Superconductivity, 2009. 19(3): p. 1208-1211.
9. Wang, X.R., et al., Analysis of Field Errors for LARP Nb3Sn HQ03 Quadrupole Magnet. Ieee Transactions on Applied Superconductivity, 2017. 27(4).
10. Fiscarelli, L., et al., Magnetic Measurements on the First CERN-Built Models of the Insertion Quadrupole MQXF for HL-LHC. Ieee Transactions on Applied Superconductivity, 2018. 28(3).
11. Holik, E.F., et al., Fabrication of First 4-m Coils for the LARP MQXFA Quadrupole and Assembly in Mirror Structure. Ieee Transactions on Applied Superconductivity, 2017. 27(4).
12. DiMarco, J., et al., Magnetic Measurements of the First Nb3Sn Model Quadrupole (MQXFS) for the High-Luminosity LHC. Ieee Transactions on Applied Superconductivity, 2017. 27(4).
13. Sorbi, M., et al., Measurements and Analysis of Dynamic Effects in the LARP Model Quadrupole HQ02b During Rapid Discharge. Ieee Transactions on Applied Superconductivity, 2016. 26(4).
14. Sabbi, G., Nb3Sn IR Quadrupoles for the High Luminosity LHC. Ieee Transactions on Applied Superconductivity, 2013. 23(3).
15. Apollinari, G., et al. High-Luminosity Large Hadron Collider (HL-LHC) Technical Design Report V.0.1 in CERN Yellow Reports: Monographs. 2017. Geneva: CERN, https://doi.org/10.23731/CYRM-2017-004
16. DiMarco, J., et al., Application of PCB and FDM Technologies to Magnetic Measurement Probe System Development. Ieee Transactions on Applied Superconductivity, 2013. 23(3).
17. DiMarco, J., et al., Field Quality Measurements of LARP Nb3Sn Magnet HQ02. Ieee Transactions on Applied Superconductivity, 2014. 24(3).
18. Jain, A. Basic theory of magnets. in CERN Accelerator School Meas. Alignment Accel. Detect. Magnets, no. CERN-98-05. 1998.
19. Todesco, E., HL-LHC Decision management: WP3 - fine tuning of b6, https://edms.cern.ch/document/2019517/4. 2018.
20. Bermudez, S.I., et al., Magnetic Analysis of the MQXF Quadrupole for the High Luminosity LHC: 4LOr3A-04, in The Applied Superconductivity Conference. 2018, IEEE Trans. Applied Superconductivity Seattle.
Page 13 of 13 draft
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Top Related