Proposal for Tests before LS2

2. Proposal for tests to be performed before LS2 Mirko Pojer and Matteo Solfaroli Beams Department Operation Group

3. Planning • One week at the end of Run II has been reserved for tests on the superconducting circuits • the objective is to train the main dipoles of one sector to 7 TeV equivalent current • some tests are proposed, to investigate possible limitations on the other circuits, that could be fixed or planned in the years to come • A similar campaign was already carried out before LS1, but additional requirements have been now identified, also related to HL-LHC.

4. Motivation • An extensive campaign of powering tests to 7 TeV was already performed before LS1. Why testing again, if already done? • Because some circuits did not reach the nominal current • Some IPQs are showing an increase in the training length • Several nominal currents have been increased N.B. The nominal current for each circuit which is contained in the Design Report or Layout Database is a “hardware-related” value which has no direct relation with the optics used in the machine. It matches the current needed for 7 or 7.5 TeV for the dipoles (fixed ring geometry) and main quadrupole (fixed cell optics), but it might be much different for the matching quadrupoles (variable insertion optics). --> In the following I will use “layout” for currents contained in the two databases above and “nominal” for the current values needed for the ATS optics configurations.

5. Before LS1 • 540 circuits were tested, including IPQs, IPDs, ITs, 600 A and some 120 A circuits; (almost) all were powered up to the layout current for 7 TeV. • 773 successful tests were executed in 10 days (more than 1300 tests performed, including repetitions due to quenches or other kind of faults). • All ITs were commissioned to 7 TeV, with a very limited number of quenches (no quench up to 6.75 TeV). • All IPQs were successfully tested up to the layout current for 7 TeV. • Many quenches were observed, above all on the MQM@4.5 K (53/68 quenches were in fact experienced in 19 of these magnets). • Q5.R2: did not reach the layout current (4310A), as was stopped at 4202 A. • The maximum nominal current will never exceed 3200 A. • All IPDs reached the layout current for 7 TeV, except RD3.L4 which reached 6.9 TeV (5784 A instead of 5860 A) and was stopped not to overstress the magnet.

6. RD3.L4 quench history A. Verweij@ Chamonix ‘17

7. After LS1 • RB and RQD/F circuits were trained to 6.5 TeV, with the exception of: • RB.A45, trained to 6.82 TeV (11536 A) with 24 quenches • RB.A34, trained to 6.74 TeV (11415 A) with 8 quenches, but suffered a short to ground. • IPQs and IPDs were powered up to 6.5 TeV. • RQ5.L8 was changed in LS1 and never pushed to 7 TeV in the LHC. • ITs were powered up to 7 TeV (for IP1 and 5) and 6.7 TeV for the others. • The 600 A circuits were mostly powered to their layout current, excluding the RQTLs (which were limited to current values between 200 and 400 A) and the IT correctors, limited to 400 A in combined powering. • The 120 A circuits were mostly powered to the layout current, with the exception of some non-conform circuits: • RCBYH4.R8B1 – 50 A • RCBYV5.L4B2 – 50 A • RCBYHS4.L5B1 – 50 A • RCBYHS5.R8B1 – 40 A • RCSSX3.L1 – 50 A. • The 60 A circuits were all powered to the layout current. What has not gone to (layout) 7 TeV in the LHC: RBs, RQD/Fs, RD3.L4 and RQ5.L8

8. IPQs A. Verweij@ Chamonix ‘17

9. 7 TeV current needs Taking all possible sets of ATS optics which are developed for LHC for 7 TeV (S. Fartoukh) and converting them in currents with the magnetic models (P. Hagen), we get the list of 7 TeV currents. With operational margin: We will require 100 A more for OP stability Few A are missing from layout current (plus 50 A margin for OP stability) Few A are missing from layout current, (plus 50 A margin for OP stability) >almost 0 quenches in past powering!< Most critical circuit, with current close to Ultimate (3900 A)! Was already tested in YETS17/18… Current increased already in the past. More than 590 A requires a hardware modification. Well below the layout current values

10. Current needs for Q5.L/R6 * Currents include 50 A as a stability margin • 7 TeV • Taking all possible sets of ATS optics which are developed for LHC for 7 TeV (S. Fartoukh) and converting them in currents with the magnetic models (P. Hagen), we get for the four circuits: • 7.5 TeV • From all possible optics and configurations for HL-LHC@7.5 TeV (R. De Maria), not necessarily corresponding to the the previous one, and converting them in currents, we get for the four circuits: HLLHCV1.4 https://indico.cern.ch/event/750135/contributions/3104578/attachments/1702836/2742909/HL14-WP2-2.pptx

11. What to test before LS2 - 1 • Mandatory to train the RB circuit; the best option (from training length point of view) is S12. • To confirm the length of training campaign to 7 TeV. • Check loss of memory of Firm 1 and 2 at higher currents. • To study the (EM induced) multiple quenches, and investigate the possibility of mitigating them (QPS modification or filtering). • For S12, investigation were done at the end of YETS to exclude weak QHs. • S12 went through additional thermal cycles wrt to the others. • Two discharges, medium (500 V) and high voltage (900 V). E. Todesco @ LMC on 24/05/17

12. What to test before LS2 - 2 • Power all RQD/F circuits to 7 TeV • To check training behavior, since no quench was ever observed yet on these circuits at currents higher than 6.5 TeV equivalent. • …or… Quench all RQD/F at 2 current levels… (see later) • https://twiki.cern.ch/twiki/pub/MP3/Meetings/2018-08-29_MP3_meeting.pptx • Push all ITs, IPQs and IPDs to 7 TeV, as done before LS1, the minimum being: • RD3.L4, to check no real limitation exists. • RQ5.L8, changed in LS1 and never sent to 7 TeV in the tunnel. • RQ5.R1 and RQ5.R5, which have shown a long (de)training. • RD2.R8, had cooling problems in the link in 2013. • Few additional IPDs and IPQs, where the nominal current is changed (add 50 A margin!) • Power Q5.R6 to 3760 A (possibly to 4 kA). • To check whether the circuit can reach the ‘new’ nominal current (also in the HL-LHC configuration). • To investigate the (already foreseen) need to cool it down to 1.9 K in the HL-LHC era. • Criteria and number of quenches as already used for RQ5.L6. • Low priority: ISRM(=Internal Splice Resistance Measurement) cycles on some dipoles (principally in S12). G. D’Angelo will confirm, but these are done!

13. What to test before LS2 - 3 • Power all 600 A and 120 A circuits to their nominal 7 TeV current. • Special tests on circuits with limitations (RQTLs and some 80-120 A circuit), to understand whether a degradation has occurred along the years. • For the 120 A circuits with suspected inter-turn short (quench-back effect), the possibility has to be investigated to use a different ramp rate on the way down. • A power cycle with all circuits powered to 7 TeV and the RBs at 6.5 TeV and a heat run of few hours, to study the thermal behavior of current leads and the dynamics and possible limitations of cryogenics, plus investigating possible limitations due to forces between circuits. Cycle to be defined. • Are there additional diagnostics on non-operable circuits which has not been done yet? • diagnostics of the shorts appeared in the RCOs and RCSs (S78)? • Additional diagnostics on ROD.A56B1 (which presents an abnormal resistance)?

14. Additional requests • Power RSD/F circuits to their nominal current with a larger dI/dt (ideally 5 A/s) • Some tests already done during TS1 in S23 -> dI/dt larger than 3 A/s is not reachable without HW or SW modification on QDS • Should be completed on all sectors • Meeting discussion: we will try 3A/s on all and try to increase by 50% the acceleration (0.22 A/s2 vs 0.15 A/s2) (needed in s23/34/67/78 and, for a test, by priority the RSF of b1) • Double di/dt of octupoles (from 5A/s to 10 A/s): it seems there is some margin, otherwise the thresholds will have to be modified. • di/dtsensor measurements in UA27: • 600 A corrector magnet circuit (RQTD) – high inductance corrector • RQ5 magnet circuit (the same of the last campaign 03/2018) – changing the installation and the sensor configuration. • Projointmeasurments in UA27: • RB, RQD  and RQF magnet circuit (plateau and 2kA steps pyramid tests), as we have done during the last test campaign 03/2018. • During meeting pre-request: quench at low current on some dipoles could be proposed, to investigate irradiated diodes (to be further developed by MP3)

15. Global view of training quenches A. Verweij@ Chamonix ‘17 Important: some commissioning currents and test sequences have been changed during the years, so the number of quenches cannot always be directly compared among HWC campaigns.

16. Planning and strategy

17. Some additional considerations • The current requested at 7.5 TeV for all RQTLs is presently within the operational parameters defined at 6.5 TeV. • A couple of cases have to be kept under observation. • To be studied (when time allows) whether a real degradation in performance is visible on these circuits or they simply show an erratic quench behavior. • The same applies to 80-120 A circuits which have been limited in current in the past powering campaigns.

18. Concluding: when and who • One week at the end of Run II has been reserved for tests on the superconducting circuits, from Dec. 3 to 10. • Will be testing around the clock • People availability: • OP will work on 3 shifts • We would like to have: • MP3 working on extended time, from 9 to 21-22 • EPC on call on 2 shifts, from 7 to 23 • PIC on call on extended time, from 9 to 20 • Software experts on call during daytime.

19. 7.5 TeV current needs From all possible optics and configurations for HL-LHC@7.5 TeV (R. De Maria), not necessarily corresponding to the the previous one, and converting them in currents, we get the list of 7.5 TeV currents. With operational margin: As for 7 TeV, this is very critical (1.9 K upgrade foreseen in HL) R6 popping up as well; 1.9 K upgrade foreseen, but might be useful to check the 4.5 K limit, with few training quenches at the end of the year Few A are missing from layout ultimate current (consider 50 A margin for stable operation) Few A are missing from layout ultimate current (consider 50 A margin for stable operation)

20. Quench of all RQD/Fs: Summarising… • The high resistances measured in 32L3 are very likely located at the contact between diode and heat sink. • Measured total value was 153 mW; measured average value about 100 mW. • Assuming that the resistance is on one side of the diode only and assuming that the other contacts are perfect, this should be compared to simulated values, namely 140 mW → 300 K, 240 mW→ 400 K. So there is still some margin… Arjan Verweij, MP3, 29/8/2018

21. Should we investigate further? Arjan Verweij, MP3, 29/8/2018

22. CSCM versus Quench campaign Arjan Verweij, MP3, 29/8/2018

23. Draft proposal • When: • At the end of Run2, just before start of warm-up • Strategy: • RQD and RQF in parallel • 2 current levels, • 3 tests per current level, each quenching 16 or 17 quads (there are circuits with 47 and 51 quads). 4 tests per current level could also be considered. • 8 sectors in parallel • 784 magnet quenches in total Arjan Verweij, MP3, 29/8/2018

24. Type test • A type test at low current is needed in order to check that: • the QPS can trigger the quench heaters of multiple magnets in parallel (with time error << t) – already done in the past (2009?) • PM data transfer is fast enough • PM data are correct (incl. QH’s) • the automatic analysis tool works properly • the measured signals agree with simulations • there are no spurious QPS trips on other magnets • cryogenic recovery is as expected • The type test can be done at 6 kA in one sector during a TS, or whenever there is a free time slot of about 4 hrs. • Since the type test will be done before the end of Run2, we should not include any magnet with whatever kind of non-conformity. Arjan Verweij, MP3, 29/8/2018

25. Thank you for the attention!

26. IPQs quench behavior cern.ch/mp3

27. 60 A circuits A. Verweij@ Chamonix ‘17 • 5 quenches in 5 different magnets • Possibly not all quenches registered • 23 quenches in 21 different magnets • 2 quenches on flat-top (60 A) Other issues: 2 circuits condemned Small number of quenches considering large number of magnets, and all quenches well above the operational currents.

28. 80-120 A circuits A. Verweij@ Chamonix ‘17 Some retraining after thermal cycling. 3 out of 4 triplet quenches due to additional 5 A margin. • Other issues: • nominal current of 4 circuits reduced to 20-50 A • 4 triplet circuits condemned Small number of quenches considering large number of magnets.

29. 600 A circuits A. Verweij@ Chamonix ‘17 Significant retraining after LS1, especially for RQTD/F and RQTL circuits. Other issues: RSS.A34B1, RCS.A78B2 condemned. 9 circuits with less magnets. 35 circuits with reduced nominal current.

30. 600 A circuits - triplets A. Verweij@ Chamonix ‘17 *: not including the quenches due to combined powering • Significant retraining after LS1. • Other issues: • Combined MCBX H-V powering imposes some constraints 600 A circuits: Many circuits to be closely watched in upcoming HWC campaigns.

31. IPD circuits A. Verweij@ Chamonix ‘17 • RD4.L4 and RD4.R4 quenched once in 2014/5, within 100 A of nominal. • Other issues: • RD1.R8 operates with one quench heater.

32. IPQ circuits A. Verweij@ Chamonix ‘17 • Other issues: • RQ9 and RQ10 frequently trip due to thunderstorms and other external pick-ups • RQ8, RQ9, and RQ10 trip when the MB in positions A8, B8, A9, B9, A10 or B10 quenches.

33. IT main circuits A. Verweij@ Chamonix ‘17 3 x RQX.L2 (Q1) 1 x RQX.R2 (Q2) 1 x RQX.L8 (Q3) 1 x RQX.L5 (Q3) Other issues: One training quench during operation in RQX.L2 (Q1) at 6760 A

34. RQ circuits A. Verweij@ Chamonix ‘17 10899 A (28R4) 11280 A (14R5) 10550 A (26L5) 10363 A (15L1) No training quenches during operation