1. MQXFPM1 and MQXFS1b Test Results
GianLuca Sabbi
HL LHC Collaboration Meeting 2016

2. Mirror Test Objectives and Features
Magnetic mirror structure allows standalone testing of coils
Developed by FNAL GARD and used extensively by LARP
Provides early feedback on cable and coil design/fabrication
Main driver is single coil requirement, not faster assembly/test
Useful at the start of a new series, or to check design/processing variants
Strength: demonstrated capability to achieve conductor limit
Weakness: generally slow training (but: faster quench recovery due to lower stored energy)
Note: some results not representative of quadrupole layout (e.g. ramp rate dependence)
A key step in the development of both short and long coils
R. Bossert
G. Sabbi – MQXFPM1and MQXFS1b Test Results 11/15/2016

3. MQXFPM1 Conductor Parameters and SSL
Short Sample Limit at 1.9K
Parameter
Unit
MQXFSM1
MQXFPM1
Alloy
(Nb-7.5w%Ta)3Sn
(Nb-1.5a%Ti)3Sn
Sub-element layout
108/127
Non-Cu fraction %
43.6
45.5
Cable Width mm
18.094
18.156
Cable Mid-Thickness
1.529
1.525
Keystone angle
Deg.
0.52
0.56
Magnet
Current
Field
Unit kA T
MQXFPM1
22.09
15.05
MQXFSM1
21.07
14.5
MQXFS1
21.50
14.6 XS
RRR
Average
295
Maximum
341
Minimum
206
MQXFPM1 HT
Temp [C]
Time [h]
210 74
401 49
641 48
A. Ghosh, I. Pong
G. Sabbi – MQXFPM1and MQXFS1b Test Results 11/15/2016

4. MQXFPM1 Assembly Steps and Pre-Load
Welded s.s. skin
SM1
Bolted s.s. skin
Mirror magnet azimuthal pre-load (MPa)
Assembly Step
Design
Actual Short Mirror
Actual Long Mirror
In press, before Al clamp insertion 80 79 74
Side clamp inserted, press released 50 66
After shell welding (bolting) 70 77 98
R. Bossert, I. Novitski
G. Sabbi – MQXFPM1and MQXFS1b Test Results 11/15/2016

5. MQXFPM1 Fabrication and End Load
Coil in lower yoke
End Plates welded, and installing End Preload Bolts
End pre-load in mirror is much lower than in quadrupole short model
R. Bossert, I. Novitski
G. Sabbi – MQXFPM1and MQXFS1b Test Results 11/15/2016

6. MQXFSM1 and MQXFPM1 Quench Training
Long Mirror PM1
A3-A4 and A4-A5
Short Mirror SM1
G. Chlachidze, S. Stoynev,
J. Muratore
G. Sabbi – MQXFPM1and MQXFS1b Test Results 11/15/2016

7. MQXFPM1 Protection, Status and Next Steps
Test Plan Assumptions:
33 mW dump resistor
<15 ms detection time
<21 MIITs during training
<30 MIITs for quench studies
Initial quenches:
16.5 mW dump for lower voltage
22 ms detection time (16 kA)
Measured quench load: 30.8 MIITs
Status and next steps:
Performed data analysis and comparison with models
Replaced one IGBT that was damaged in the second training quench
Increased dump resistor back to 33 mW
Performed hi-pot ok up to 800 V (80K); system checks at 4.5K in progress
Planning to decrease 10 ms validation time and (possibly) detection threshold
E. Ravaioli, J. Muratore
G. Sabbi – MQXFPM1and MQXFS1b Test Results 11/15/2016

8. Reference: MQXFS1 Design Report,
MQXFS1 Quadrupole
First-generation cable and coil design (larger keystone angle)
Two coils from CERN and two from LARP
CERN (#103 and #104) – RRP 132/169
LARP (#3 and #5) – RRP 108/127
Magnetic length 1.2 m, Coil length 1.5m
Reference: MQXFS1 Design Report,
MQXFS1 Coils: Outer Layer
MQXFS1 Coils: Inner Layer
CERN 103 & 104
CERN 103 & 104
LARP 3 & 5
LARP 3 & 5
D. Cheng, J. Schmalzle, J.C. Perez, M. Yu et al.
G. Sabbi – MQXFPM1and MQXFS1b Test Results 11/15/2016

9. Short Sample and Reference Current Levels
Parameter
Current
Field
Gradient
Inom/Iss
Temperature
1.9 K
Unit kA T
T/m %
Coil 103
21.50
14.6
169.08 77
Coil 104
21.78
14.77
171.03 76
Coil 3
22.28
15.08
174.57 74
Coil 5
21.85
14.81
171.52 75
Current [kA]
Symbol
Gradient [T/m]
Remarks
0.01
I.warm
0.09
Warm measurements
0.1
I.res
0.9
Reset level for pre-cycle
0.96
I.inj
8.5
Injection level
6.0
I.lim
48.8
Current limit (pre-training)
16.48
I.nom
132.6
Nominal level
17.76
I.ult
143.2
Ultimate level
21.50
I.ssl
171.0
1.9K Short Sample Limit
Parameter
Current
Field
Gradient
Unit kA T
T/m
Coil 103
19.55
13.38
155.16
4.3K SSL:
B. Bordini, A. Ghosh, P. Ferracin
G. Sabbi – MQXFPM1and MQXFS1b Test Results 11/15/2016

10. MQXFS1 Quench Analysis Quench locations:
Quenches observed in all coils and distributed over many locations in the high field areas
About 50% in coil 103 (both lower margin and earlier unloading)
Some quenches in end regions and layer ramp
Strain gauges:
Preload increase during cool- down lower than expected
During excitation, deviation from linear behavior starts at about 65% of SSL
H. Pan, T. Strauss,
G. Vallone
G. Chlachidze, S. Stoynev
Strain gauges, distribution of quench locations, and quench level independent of temperature consistently indicate low pre-load as the main limiting factor
No indication of coil performance degradation up to the quench level achieved
G. Sabbi – MQXFPM1and MQXFS1b Test Results 11/15/2016

11. HQ02 Training Rate vs. Pre-load
HQ02a/a2 slow training above 80% SSL, attributed to low pre-load
HQ02b: Much faster training up to 95% SSL after pre-load increase
H. Bajas, M. Bajko, G. Chlachidze, M. Martchevsky, F.Borgnolutti, D. Cheng, H. Felice, et al.
G. Sabbi – MQXFPM1and MQXFS1b Test Results 11/15/2016

12. From MQXFS1 to MQXFS1b
Average azimuthal pre-load increased from 80 to 100 MPa
No change in axial pre-load: 600 kN, half of Lorentz force
D. Cheng, H. Pan, G. Vallone
G. Sabbi – MQXFPM1and MQXFS1b Test Results 11/15/2016

13. MQXFS1b vs MQXFS1 Quench Training
MQXFS1b fully retained the maximum quench level achieved in MQXFS1
Quench level further improved, but with slower rate and some fall backs
Highest quench so far: 18.8kA, 87.4%
Next steps: protection studies and further localization of quench origin
G. Chlachidze, S. Stoynev
G. Sabbi – MQXFPM1and MQXFS1b Test Results 11/15/2016

14. MQXFS1b Quench Instrumentation
Voltage Taps
Axial field quench antenna
Antenna is installed in warm bore – can be moved and reconfigured
M. Martchevsky
G. Sabbi – MQXFPM1and MQXFS1b Test Results 11/15/2016

15. MQXFS1b Quench Locations
Most quenches developed in the end regions:
Voltage taps: IL Pole block A3-A4 segment
QA: LE, segment #1 (11 quenches) and RE, segment #7 (4)
Segment #8 not connected due to limited DAQ channels
1 seg - Coil #3, quench in A3-A4
2 seg – Coil #3, quench in A3-A4 and A4-A5
Additional training quenches are planned in addition to protection studies
Quench antenna is being reconfigured for more precise quench localization
G. Chlachidze, S. Stoynev, M. Martchevsky, T. Strauss
G. Sabbi – MQXFPM1and MQXFS1b Test Results 11/15/2016

16. QA reconfiguration LE quench localization
QA coils are now closely spaced around the LE quench area
Accurate localization of QA relative to magnet coil is also required
Possibility to cross-correlate with voltage taps or MM probe
Increased delays are being considered (sufficient QI margin)
Installation on warm bore provides much increased flexibility
Probe can also be rotated to provide left-right positioning
T. Strauss, M. Martchevsky
G. Sabbi – MQXFPM1and MQXFS1b Test Results 11/15/2016

17. Comparison between HQ02 and MQXFS1 Ends
MQXF coil-spacer interface has tension not observed in HQ02
Tension appears to increase with higher pre-load (needs to be confirmed/quantified)
Also high compression, but different areas for MQXF vs. HQ
One additional block in MQXF than in HQ
Flexible spacers also implemented in MQXF
HQ02a (L1)
HQ02b (L1)
MQXFS1a (L1)
MQXFS1b (L1)
H. Pan, G. Vallone
G. Sabbi – MQXFPM1and MQXFS1b Test Results 11/15/2016

18. Magnetic Measurements and Protection Studies
Effect of magnetic shim
Magnetic Measurements: [S. Izquierdo Bermudez presentation]
Reduced plan in MQXFS1b relative to MQXFS1
Focus on harmonics correction with magnetic shim
Two harmonics were targeted: effect is consistent with calculation, but in one case sign is opposite
Need to ensure correct referencing of magnetic measurements, calculations and shim installation
J. DiMarco
Protection studies: [E. Ravaioli presentation]
Extensive protection studies are planned in both MQXFS1b and MQXFPM1
MQXF1b: first opportunity to test CLIQ on the MQXF design
Mirror: first opportunity to verify heater performance and modelling predictions on a long MQXF coil
Long Mirror Quench #1 Analysis
G. Sabbi – MQXFPM1and MQXFS1b Test Results 11/15/2016

19. Summary Long Mirror MQXFPM1: MQXFS1b Quadrupole:
First feedback on MQXF Long Coil Performance
Start of training is comparable to short mirror
New test facility commissioned at BNL
Need to proceed with caution to ensure safe operation during training
MQXFS1b Quadrupole:
Training started at the maximum quench level achieved in MQXFS1
Quench level further improved, but with slower rate and some fall backs
Quenches are primarily in the ends, different from HQ/LARP experience
Need to determine cause for end quenches, and corrective actions
Diagnostic effort centered on axial field antenna to precisely locate quenches
Highest quench so far: 18.8kA, 87.4%
Magnetic shim correction: wo harmonics were targeted, effect is consistent with calculation, but in one case sign is opposite
Extensive protection studies underway: first opportunity to test CLIQ in MQXF
G. Sabbi – MQXFPM1and MQXFS1b Test Results 11/15/2016

20. Backup Slides
G. Sabbi – MQXFPM1and MQXFS1b Test Results 11/15/2016

21. Previous QA Configurations
G. Sabbi – MQXFPM1and MQXFS1b Test Results 11/15/2016

22. D20 Training History
G. Sabbi – MQXFPM1and MQXFS1b Test Results 11/15/2016