MQXF quench heaters tests and investigation CERN Paolo

MQXF quench heaters: tests and investigation (CERN)¶ Paolo Ferracin, Arnaud Foussat, Juan Carlos Perez, Jan Petrik on behalf of the MQXF collaboration logo area MQXF Video-meeting 03 April 2019 CERN

Outline • Overall design • Electrical Test after coil fabrication • Electrical Test on MQXFS 3 and MQXFS 5 coils (after cold test) • Test up to 5 k. V of MQXFS 3 coils (105, 106, 107, L 8) • Test up to 3. 7 k. V of MQXFS 5 coils (203, 204, 205, 206) • Partial discharge test on MQXFS 3 coils, MQXF 5 coils and 207 (virgin) • Breakdown test on MQXFS 3 coils and 206 (limiting coil in MQXFS 5) Paolo Ferracin 2

MQXF coil and cavity size • Radius insulated cable after reaction • • R 1: 75. 000 mm R 2: 93. 653 mm R 3: 94. 313 mm R 4: 112. 965 mm Cable before reaction Bare width mm 18. 150 Bare thickness in mm 1. 462 Bare thickness out mm 1. 588 Bare mid-thickness mm 1. 525 Keystone angle Cable after reaction Bare width mm 18. 363 1. 17 % growth in width Bare thickness in mm 1. 530 4. 5% growth in thickness Bare thickness out mm 1. 658 Bare mid-thickness mm 1. 594 0. 400 Insulation thickness mm 0. 145 Ins. width mm 18. 440 Ins. thickness in mm 1. 752 Ins. thickness out mm 1. 878 Ins. mid-thickness mm 1. 815 Keystone angle Insulation thickness Ins. width Ins. thickness in Ins. thickness out Ins. mid-thickness 0. 400 0. 145 18. 653 1. 820 1. 948 1. 884 Paolo Ferracin 3

Impregnation – New coils (second generation) Impreg Design: 310 µm 3 layers S 2 Glass Hexcel 4522 CERN meas @ 5 MPa: 290 µm plain CERN meas @ 1 MPa: 329 µm plain 250 µm G 11 Silicone Mold Putty (removed after impregnation) Impreg Design: 150 µm 100 µm Polyimide trace (real thk. 75 µm) Radial filler: 0. 5 mm Radial filler 0. 6 mm Coating on 4 sides of the fillers for 0. 120 mm total 100 µm Polyimide trace (real thk. 75 µm) 1 layer S 2 Glass Hexcel 4522 Midplan LARP meas @ 5 MPa: 121 µm plain LARP meas @ 1. 6 MPa: 133 µm plain e CERN meas @ 0 MPa (micrometer): 160 µm plain 250 µm G 11 CERN mould has 40 µm additional inner radius due to Teflon coating 4

Ground insulation Paolo Ferracin 5

Voltage withstand level table • EDMS 1963398 v 4 (2018 -04 -23) Paolo Ferracin 6

Outline • Overall design • Electrical test after coil fabrication • Electrical Test on MQXFS 3 and MQXFS 5 coils (after cold test) Paolo Ferracin 7

Test after coil fabrication Short model • Coil to quench heater: 3680 V (since ~04/18) • Out of 29 coils electrically tested, no issues, but… • • 1 tested only up to 2000 V 23 tested only up to 2500 V 4 tested up to 3000 V 1 tested up to 5000 V Paolo Ferracin 8

Test after coil fabrication MQXFB prototype (to be updated) • Coil to quench heater: 3680 V (since ~04/18) • Out of 8 coils electrically tested, no issues, but • 4 tested to 3000 V • 4 (MQXFBP 1) tested to 3700 V Paolo Ferracin 9

Test after coil fabrication MQXFA prototype (to be updated) • Coil to quench heater: 3680 V (since ~04/18) • Out of 15 coils electrically tested, no issues, but • 2 tested to 2500 V • 11 tested to 3000 V • 2 tested to 3680 V Paolo Ferracin 10

Test after coil fabrication All (to be updated) • Coil to quench heater: 3680 V (since ~04/18) • Out of 52 coils electrically tested, no issues, but • the test voltage ranged from 2000 V to 5000 V Paolo Ferracin 11

MQXFS 3 • Coil 105, 106, 107 • 5 thermal cycles • 1 full disassembly • ~100 quenches • Coil 8 • 2 thermal cycles • ~50 quenches Paolo Ferracin 12

MQXFS 3 coil tested up to 5 k. V Coil QH Test to…. Bdv k. V 105 OLL 5 3. 9 External wire 105 OLH 5 2. 3 External wire 105 ORH 5 3. 7 External wire 105 ORL 5 2. 3 External wire 106 OLL 5 NO 106 OLH 5 3. 6 106 ORH 5 4. 5 External wire 106 ORL 5 3. 8 External wire 107 OLL 5 NO 107 OLH 5 NO 107 ORH 5 0. 1 107 ORL 5 NO LARP 8 OLH 5 NO LARP 8 ORH 5 0. 6 LARP 8 ORL 5 NO Paolo Ferracin 13

MQXFS 3 coil tested up to 5 k. V Coil QH Test to…. Bdv k. V 105 OLL 5 3. 9 External wire 105 OLH 5 2. 3 External wire 105 ORH 5 3. 7 External wire 105 ORL 5 2. 3 External wire 106 OLL 5 NO 106 OLH 5 3. 6 106 ORH 5 4. 5 External wire 106 ORL 5 3. 8 External wire 107 OLL 5 NO 107 OLH 5 NO 107 ORH 5 0. 1 107 ORL 5 NO LARP 8 OLH 5 NO LARP 8 ORH 5 0. 6 LARP 8 ORL 5 NO Paolo Ferracin 14

MQXFS 5 • Coil 203, 204, 205, 206 • 3 thermal cycles • ~50 quenches Paolo Ferracin 15

MQXFS 5 coil tested up to 3. 7 k. V Coil QH Test to…. Bdv k. V 203 OLL 3. 7 NO 203 OLH 3. 7 NO 203 ORL 3. 7 NO 204 OLH 3. 7 NO 204 ORL 3. 7 NO 205 OLH 3. 7 NO 205 ORL 3. 7 NO 206 OLH 3. 7 2. 0 206 ORH 3. 7 NO 206 ORL 3. 7 NO Paolo Ferracin 16

Partial discharge Occurs when high enough excitation voltage(for given insulation system) is present Represents miniature failures in insulation (bubbles, voids, discharges. . ) Can accumulate over time and result in failure Requirement for PD free insulation in power transformers and other expensive equipment which are subjected continuous AC excitation – not our case Very sensitive Good for relative comparison 17 Jan Petrik, 20/03/2019

Overall Results Assembly MQXFS 3 MQXFS 5 MQXFS 3 Virgin MQXFS 3 MQXFS 5 MQXFS 5 MQXFS 3 MQXFS 5 Coil LARP 8 206 204 106 207 107 205 203 206 204 203 204 105 206 Side Right Right Left Right Left Right Field High High High Low Low High Low PD Inception Voltage - [ACV] 559 619 821 837 850 853 902 929 960 961 976 980 986 992 1000 Assembly Virgin MQXFS 5 MQXFS 3 MQXFS 3 MQXFS 5 MQXFS 3 Coil 207 205 107 105 203 106 105 107 203 LARP 8 106 205 LARP 8 105 LARP 8 Side Right Left Left Right Right Left Field Low Low High Low Low High Low PD Inception Voltage - [ACV] 1000 1011 1022 1053 1055 1069 1072 1086 1096 1139 1150 1218 1234 1238 1342 1349 107 Right-High – QH already shorted logo are Jan Petrik, 20/03/2019

PD Results The Worst 10 – all are High field QH 7 out of 10 – Right side, High field QH Right side, high field is the worst QH for 8 out of 9 coils…#107 – RH QH was already busted… Virgin coil (#207) is comparable with others This would indicate (= not conclusive!!) that there is no damage during assembly and cold testing To be sure we would need to measure #207 after cold testing. 19 Jan Petrik, 20/03/2019

Breakdown voltage test • Tests • Break down test – ramp 1 k. V/min • Burn mode – Step, 500 V increase • Capacitive discharge to do more damage • 5 coils • 105, 106, 107, LARP 8, 206 Paolo Ferracin 20

Paolo Ferracin 21

Breakdown voltage test Coil QH Test to…. Bdv k. V 105 OLL 10 NO 105 OLH 10 NO 105 ORH 10 6. 5 105 ORL 10 NO 106 OLH 10 3. 6, then lower Located 106 ORH 10 8. 5 Located 106 ORL 10 NO 107 OLH 10 NO 107 ORH 10 0. 1 107 ORL 10 8 LARP 8 OLL 10 7 LARP 8 OLH 10 NO LARP 8 ORH 10 0. 6 LARP 8 ORL 10 NO 206 OLL 10 2 ? ? Located 206 OLH 10 2. 5 Located 206 ORH 10 5 206 ORL 10 NO Paolo Ferracin Located 22

Locations Paolo Ferracin 23

Breakdown voltage test (above 3. 7 k. V) Coil QH Test to…. Bdv k. V 105 OLL 10 NO 105 OLH 10 NO 105 ORH 10 6. 5 105 ORL 10 NO 106 OLH 10 3. 6, then lower 106 ORH 10 8. 5 106 ORL 10 NO 107 OLH 10 NO 107 ORH 10 0. 1 107 ORL 10 8 LARP 8 OLL 10 7 LARP 8 OLH 10 NO LARP 8 ORH 10 0. 6 LARP 8 ORL 10 NO 206 OLL 10 2 ? ? 206 OLH 10 2. 5 206 ORH 10 5 206 ORL 10 NO Paolo Ferracin 24

Appendix Paolo Ferracin 25

Coil design Paolo Ferracin 26

Cable insulation thickness Paolo Ferracin 27

Change in coil size with heaters outside • Removal of • 0. 310 mm impregnated fiberglass (3 layers) • 0. 100 mm trace • Adding of • ~0. 100 mm fiberglass (11 T scenario) • Total • 0. 310 mm smaller impregnated coil • 0. 210 mm smaller coil including external quench heaters Paolo Ferracin 28

Voltage withstand level table • How to obtain the test voltage levels Paolo Ferracin 29

Test after coil fabrication Short model • Coil to quench heater: 3680 V (since ~04/18) Paolo Ferracin 30

Test after coil fabrication MQXFB prototype • Coil to quench heater: 3680 V (since ~04/18) Paolo Ferracin 31

Test after coil fabrication MQXFA prototype • Coil to quench heater: 3680 V (since ~04/18) Paolo Ferracin 32

MQXFS coil 105 Paolo Ferracin 33

MQXFS coil 106 Paolo Ferracin 34

MQXFS coil 107 Paolo Ferracin 35

MQXFS coil 8 Paolo Ferracin 36

MQXFS 5 • Coil 203, 204, 205, 206 • 3 thermal cycles • ~50 quenches Paolo Ferracin 37

MQXFS coil 203 Paolo Ferracin 38

MQXFS coil 204 Paolo Ferracin 39

MQXFS coil 205 Paolo Ferracin 40

MQXFS coil 206 Paolo Ferracin 41

Partial Discharge Filter AC PD Detector DUT Jan Petrik, 11/01/2019 42

What we evaluated PD inception voltage PD vs voltage PD pattern Tricky Comparison based 43 Jan Petrik, 20/03/2019

Limits We do not have bare coil QH contains Voltage taps traces Cables We measure everything together Where are the limits of used cables? Quick test with separate cable - 1689 V Single test, let’s assume that above 1300 V we are not sure… Distinctive PD pattern If PD activity was low, we’ve kept increasing test voltage, in 3 cases the coil DP pattern was swamped by cables logo are Jan Petrik, 20/03/2019

Results Per Coil PD Inception Voltage - [ACV] Coil Virgin Coil logo are QH Count Avg Min Max STD Larp 8 4 1071. 25 559 1349 304. 9347266 206 3 859. 66667 619 1000 170. 9587345 204 4 937 821 986 67. 60547315 106 3 1019. 6667 837 1150 133. 0321599 207 4 928. 5 850 1011 77. 10544728 107 3 1017 902 1096 83. 19054433 205 4 1099 922 1234 131. 9507484 203 4 1027. 5 929 1136 80. 38812101 105 4 1118. 75 992 1342 133. 2692294 Due to RH QH Jan Petrik, 20/03/2019

Results per QH QH All QH High Field only Low Field Only Left Side only Right side only Left & High only Left & Low only Right & High only Right & Low only QH Count 33 16 17 8 8 8 9 Avg 1012. 6364 931. 3125 1089. 1765 1080. 8125 948. 47059 1046. 5 1115. 125 816. 125 1066. 1111 PD Inception Voltage - [ACV] Min Max 559 1349 559 1342 961 1349 853 1349 559 1218 853 1342 980 1349 559 992 961 1218 STD 169. 9536155 188. 4667473 102. 7431036 141. 0528353 169. 8588056 155. 9503126 114. 5277668 142. 0655109 84. 55562856 And per Assembly MQXFS 3 MQXFS 5 logo are QH Count Avg 14 1062. 1429 15 988. 86667 PD Inception Voltage - [ACV] Min Max STD 559 1349 196. 6711749 619 1234 145. 3230272 Jan Petrik, 20/03/2019

CERN short models Magnet tests summary – updated 8. 10. 2018 Reception (warm, air) In LHe Test station Coil-Gnd Coil-QH 3680 V 368 V 460 V 1840 V 2300 V 3 a 3700 3000 1500 2300 HFM 3 b - - 1500 2300 HFM 5_1 3700 3000 1500 1000 Cluster D 5_2 - - 3000 1000 Cluster D 3 c_1 3700 3000 1000 Cluster D? 3 c_2 - - 1000 Cluster D? 3 c_3 - - 3700 500/1000 800 1100 Cluster D 4 a 3700 3000 1000 950 1800 Cluster D 4 b - - 1000 500 1880 HFM EDMS 1963398 Chronological order After LHe (warm, air/He) Target >> EDMS 3000 1000 Target = EDMS Target < EDMS Target not reached

Details on MQXFS 5 tests Reception (warm, air) EDMS 1963398 After LHe (warm, air/He) In LHe Test station Coil-Gnd Coil-QH 3680 V 368 V 460 V 1840 V 2300 V 5, 1 st run: 3 a tests Coil-Gnd at cold 1000 V OK 3700 3000 OK at 500, 1000 V. Then failed at 1500 V, but tested 2300 again at HFM 3 b - - 3000 5_1 3700 3000 5_2 - - 3000 3 c_1 3700 3000 3 c_2 - 3 c_3 1000 1500 2300 HFM 1500 1000 Cluster D 1000 Cluster D 3700 3000 1000 HFM? - 1000 HFM? - - 3700 500/1000 800 1100 Cluster D 4 a 3700 3000 1000 950 1800 Cluster D 4 b - - 1000 500 1880 HFM 5, 2 nd run: tests Coil-Gnd at cold OK at 500, 1000 V. Did not try to go beyond that Target >> EDMS Target = EDMS Target < EDMS Target not reached

Details on MQXFS 3 c, run 3 tests Reception (warm, air) After LHe (warm, air/He) In LHe Test station Coil-Gnd Coil-QH 3680 V 368 V 460 V 1840 V 2300 V 3 a 3700 3000 1500 2300 HFM 3 b - - 1500 2300 HFM 5_1 3700 3000 1500 1000 Cluster D 5_2 - - 1000 Cluster D EDMS 1963398 3000 1000 3 c, 3 rd run: up to the EDMS value did not show the defective 3 c_1 testing 3700 3000 recommended 3700 3000 1000 quench HFM? heater found at our own target value (1 k. V). 3 c_2 - - 1000 HFM? 3 c_3 - - 3700 500/1000 800 1100 Cluster D 4 a 3700 3000 1000 950 1800 Cluster D 4 b - - 1000 500 1880 HFM For this run we also tested the coil-QH insulation in gaseous helium at 80, 150 and 280 K. Target >> EDMS Target = EDMS Target < EDMS Target not reached

Details on MQXFS 4 tests Reception (warm, air) After LHe (warm, air/He) In LHe Test station Coil-Gnd Coil-QH 3680 V 368 V 460 V 1840 V 2300 V 3 a 3700 3000 1500 2300 HFM 3 b - - 1500 2300 HFM EDMS 1963398 3000 1000 4 a: tests 5_1 at cold failed the EDMS target value. Maximum reached was 900 3700 to reach 3000 1500 1000 V coil- Cluster D ground and 1700 V- coil-QH. 5_2 3000 1000 Cluster D 3700 the auxiliary 3000 leads for 1000 4 b: after 3 c_1 changing test facility 3000 and upgrading CLIQ, the maximum voltage 3 c_2 reached was and - 1140 V coil-ground 10001400 V coil-QH 1000 Breakdown in coil-QH at 1880 V happened after ~25 s of maintaining the voltage 3 c_3 HFM? - - 3700 500/1000 800 1100 Cluster D 4 a 3700 3000 1000 950 1800 Cluster D 4 b - - 1000 500 1880 HFM Target >> EDMS Target = EDMS Target < EDMS Target not reached

Possible test plan • Additional tests to 3. 7 k. V • MQXFAP 2 coils • For all the coils of MQXFS 3 and for coil 206 of MQXFS 5 • Identify failure mode destructive discharge test • After destructive discharge test, micrograph on a cut coil • We have pieces of LARP coil 7 (MQXFS 3 ab) already available • Partial discharge tests • Multiple cool-downs? Paolo Ferracin 51