Power converter requirements D Gamba G Arduini J
Power converter requirements D. Gamba, G. Arduini, J. M. Coello de Portugal, R. De Maria, S. Fartoukh, M. Fitterer, M. Giovannozzi, R. Tomás Acknowledgements: A. Ballarino, R. Bruce, J. -P. Burnet, B. Holzer, Q. King, E. Mc. Intosh, E. Métral, Y. Sammer, F. Savary, F. Schmidt, H. Thiesen, E. Todesco, M. Martino and the LHC@Home volunteer logo area 6 th HL-LHC Collaboration Meeting – Paris (FR) – 15/11/2016
Outline § Latest circuits layout based on current baseline (from June 2016) § Operational requirements § § § Inner triplet Dipole correctors Non-linear correctors D 1/D 2 11 T magnets § Beam Dynamics requirements § Power Converters ripple logo area D. Gamba et al. - 6 th HL-LHC Collaboration Meeting – Paris (FR) – 15/11/2016 2
General requirements [6][8] § Requirements from operation § Maximise physics time § Minimise time of squeeze § Minimise time to put beams in collision § while keeping acceptable loss spikes for orbit feedback § Minimise time for ramp-down of magnets § Minimise time for pre-cycles. § Requirements from beam dynamics § Give sufficient current range for optics gymnastics § Minimise time-dependent effects, such as ripple § Impact on emittance § Impact on luminosity § Minimise duration of transients (e. g. putting beams in collisions) § Avoid instabilities § Allow optics measurements, e. g. k-modulation logo area D. Gamba et al. - 6 th HL-LHC Collaboration Meeting – Paris (FR) – 15/11/2016 3
Operational requirements logo area D. Gamba et al. - 6 th HL-LHC Collaboration Meeting – Paris (FR) – 15/11/2016 4
Circuits Layout [3] logo area D. Gamba et al. - 6 th HL-LHC Collaboration Meeting – Paris (FR) – 15/11/2016 5
Latest circuits table (01/11/2016) [3] Inner Triplet Circuits for Hi. Lumi D 2 D 1 Q 4 Q 5 Q 6 L per circuit [m. H] R per circuit [mΩ] PC Quad Ramp rate Acceleration Numbe [A/s] rate [A/s 2] r 2 tbd tbd 4 15 5 4 tbd Triplet Q 1, Q 2 a, Q 2 b, Q 3 Trim Q 1 Trim Q 3 Trim Q 2 a Orbit correctors Q 2 a/b - vertical Orbit correctors Q 2 a/b - horizontal Orbit correctors CP - vertical Orbit correctors CP - horizontal Superferric, order 2 MQXFA / MQFXB MCBXFB MCBXFA MQSXF 1 1 2 2 1 1 1 4 (IR 1/5) 8 (IR 1/5) 4 (IR 1/5) 16. 5 2 2 0. 12 1. 6 1. 47 0. 182 17. 82 2 2 0. 12 1. 73 1. 59 0. 2 255 69 69 58. 5 59 135 109 247 1247 0. 264 1. 44 13. 372 1. 512 1. 656 1. 728 9. 853 Superferric, order 3, normal and skew MCSXF / MCSSXF 2 8 (IR 1/5) 0. 105 0. 12 118 13. 372 4 tbd Superferric, order 4, normal and skew MCOXF / MCOSXF 2 8 (IR 1/5) 0. 105 0. 12 152 13. 372 4 Superferric, order 5, normal and skew MCDXF / MCDSXF 2 8 (IR 1/5) 0. 105 0. 12 107 13. 372 MCTXF MCTSXF 1 1 4 (IR 1/5) 0. 105 0. 12 229 52 Separation dipole D 1 MBXF 1 4 (IR 1/5) 12 12. 96 Recombination dipole D 2 MBRD 1 4 (IR 1/5) 12 Orbit correctors D 2 MCBRD 4 16 (IR 1/5) Individually powered quad Q 4 (1. 9 K) MQY 2 Orbit correctors Q 4 (1. 9 K) MCBY Individually powered quad Q 5 (1. 9 K) Ramp Up Time [s] Ramp Down References Time [s] tbd tbd 115. 34 106 tbd tbd tbd 4 tbd tbd 13. 372 4 4 tbd tbd 27 0. 27 1 20 2 648 500 12. 96 25 0. 234 1 20 2 648 534. 19 0. 54 600 1. 08 4 2 1 270 8 (IR 1/5) 4. 5 4. 86 74 0. 65 1 10 2 486 569. 24 EDMS 1375861 8 32 (IR 1/5) 0. 088 0. 1 5270 tdb 4 0. 667 0. 25 149. 93 EDMS 1375861 MQY 2 8 (IR 1/5) 4. 51 4. 88 74 0. 6 1 10 2 488 616. 67 EDMS 1375861 Orbit correctors Q 5 (1. 9 K) MCBY 6 24 (IR 1/5) 0. 072 0. 08 5270 tdb 4 0. 667 0. 25 119. 95 EDMS 1375861 Individually powered quad Q 6 (4. 5 K) MQML 2 8 (IR 1/5) 4. 31 4. 66 21 0. 47 1 12 2 388. 34 223. 41 EDMS 1375861 Orbit correctors Q 6 (4. 5 K) MCBC 2 8 (IR 1/5) 0. 08 0. 09 2840 tdb 4 0. 667 0. 25 134. 94 EDMS 1375861 11 T dipole, MBH - - 2 (IR 7) 11. 85 0. 25 12. 798 0. 25 15734 127. 1 1 15 2 4 10 0. 5 1 1300 Superferric, order 6, skew 11 T Magnet Type Number of Total I_nominal I_ultimate circuits per number of (7 Te. V) [k. A] IP side circuits [k. A] 11 T dipole, MBH Trim circuit logo area D. Gamba et al. - 6 th HL-LHC Collaboration Meeting – Paris (FR) – 15/11/2016 6
Ramping down and precycle [6][8] § Requirement: the ramp down and precycle time of the new magnets needs to be in the shadow of that of the main magnets. MQs ≈ 16 min MBs ≈ 21 min MQXs (IR 2/8) ≈ 32 -37 min D 1 s ≈ 19 min Q 4 s (IR 2/8) ≈ 17 -24 min § The main magnets feature a ramp down time between 16 -21 min. § We keep this as a reference time in the following. § The slowest magnets are the ITs of IR 2/8 (about 32 -37 min). § Reduction of the ramp-down time would benefit the turn-around time § Superferric circuits: how many cycles of degaussing required? logo area From “Beam Dynamics requirements” - M. Giovannozzi – (link) [6] D. Gamba et al. - 6 th HL-LHC Collaboration Meeting – Paris (FR) – 15/11/2016 7
IT quadrupoles Inner Triplet Circuits for Hi. Lumi Triplet Q 1, Q 2 a, Q 2 b, Q 3 Trim Q 1 Trim Q 3 Trim Q 2 a Magnet Type MQXFA / MQFXB - Number of Total I_nomina I_ultimate circuits per number of l (7 Te. V) [k. A] IP side circuits [k. A] 1 1 4 (IR 1/5) 16. 5 2 2 0. 12 17. 82 2 2 0. 12 L per circuit [m. H] 255 69 69 58. 5 u Proposed values R per Ramp PC Quad. Ramp rate Acceleration Ramp Up circuit Down References 2 Number [A/s] rate [A/s ] Time [s] [mΩ] Time [s] 0. 264 1. 44 13. 372 2 4 4 4 15 ± 2 ± 1 ± 1 ± 1 1200 ≈1300 New base line Q 1 -Q 2 -Q 3: § smaller tune ripple than Q 1 -Q 3 Q 2 a-Q 2 b for current control regime § best compensation voltage control regime § range sufficient for baseline optics. § About 10% range in current imbalance between quadrupoles [8] -> Trim = 2 k. A § Ramp rates dictated by: § need to follow main magnet ramp and allow ramp down within ~21 minutes § K-modulation § Acceleration rates determined by: § Deceleration/acceleration at the end/beginning of ramp-up/ramp-down § Deceleration/acceleration for k-modulation § Ramp/Acc. rates comparable with specification of current LHC circuits powering tests [1] logo area D. Gamba et al. - 6 th HL-LHC Collaboration Meeting – Paris (FR) – 15/11/2016 8
IT correctors Inner Triplet Circuits for Hi. Lumi Orbit correctors Q 2 a/b - vertical Orbit correctors Q 2 a/b - horizontal Orbit correctors CP - vertical Orbit correctors CP - horizontal Superferric, order 2 Superferric, order 3, normal and skew Superferric, order 4, normal and skew Superferric, order 5, normal and skew Superferric, order 6, skew Magnet Type MCBXFB MCBXFA MQSXF MCSXF / MCSSXF MCOXF / MCOSXF MCDXF / MCDSXF MCTSXF Number of I_nomina L per Total number I_ultimate circuits per l (7 Te. V) circuit of circuits [k. A] IP side [k. A] [m. H] 2 2 1 1 1 2 2 2 1 1 8 (IR 1/5) 4 (IR 1/5) 8 (IR 1/5) 4 (IR 1/5) 1. 6 1. 47 0. 182 0. 105 1. 73 1. 59 0. 2 0. 12 59 135 109 247 118 152 107 229 52 u u R per circuit [mΩ] 1. 512 1. 656 1. 728 9. 853 13. 372 Hypothetical values Confirmed values PC Quad Ramp rate Acceleration Ramp Up Number [A/s] rate [A/s 2] Time [s] 4 4 4 4 4 ± 15 ± 3 ± 3 ± 3 ± 5 ± 5 ± 1 ± 1 ± 0. 1 115. 34 106 *** *** *** Ramp Down References Time [s] 115. 34 106 *** *** *** Triplet orbit correctors MCBXF: § Ramp rates and acceleration rates dictated by: § Possibility to include the MCBXF in the orbit feedback (presently limited by maximum achievable acceleration and ramp rates) § Need to bring the beam in collision on a time scale comparable to that of the LHC § Cross the separation region below 2 sigma in less than 3 seconds § Values still to be confirmed by latest information on orbit knob. Superferric non-linear correctors: • Proposed ramp and acceleration rates dictated by: • K-modulation to determine the magnetic center by beam-based measurements (up to b 4) • critical for non-linear corrector strategy according to 2016 experience • degauss cycles and procedures: needs input from WP 3 • Proposed numbers based on the assumption that 4 cycles should be completed in 20 mins *** Ramp up/down is not the main parameter: important is the length of the degaussing cycles! logo area D. Gamba et al. - 6 th HL-LHC Collaboration Meeting – Paris (FR) – 15/11/2016 9
D 1, D 2 dipoles and D 2 correctors u D 2 D 1 Circuits for Hi. Lumi Number of Total I_nomina I_ultimate Magnet Type circuits per number of l (7 Te. V) [k. A] IP side circuits [k. A] L per circuit [m. H] R per circuit [mΩ] Confirmed values PC Ramp Quad Ramp rate Acceleration Ramp Up Down References 2 Numbe [A/s] rate [A/s ] Time [s] r Separation dipole D 1 MBXF 1 4 (IR 1/5) 12 12. 96 27 0. 27 1 20 2 648 500 Recombination dipole D 2 MBRD 1 4 (IR 1/5) 12 12. 96 25 0. 234 1 20 2 648 534. 19 Orbit correctors D 2 MCBRD 4 16 (IR 1/5) 0. 54 600 1. 08 4 ± 2 ± 1 270 § D 1/D 2: Proposing to keep similar specs as for LHC [R. De Maria]. § Ramp rate 20 A/s; Acc. Rate 2 A/s 2 § Energy ramp in the shadow of these values. § Values compatible with current specifications for LHC circuits test [1]. § MCBRD: same consideration as for the correctors in the triplet: § We need to include it into the orbit feedback. § Similar kick/s as for current LHC correctors [R. De Maria]. § Ramp rate 2 A/s; Acc. rate 1 A/s 2 logo area D. Gamba et al. - 6 th HL-LHC Collaboration Meeting – Paris (FR) – 15/11/2016 10
Matching quadrupoles Q 6 Q 5 Q 4 Circuits for Hi. Lumi Number of Total I_nominal I_ultimate L per circuit Magnet Type circuits per number of (7 Te. V) [k. A] [m. H] IP side circuits R per circuit [mΩ] Individually powered quad Q 4 (1. 9 K) MQY 2 8 (IR 1/5) 4. 5 4. 86 74 0. 65 Orbit correctors Q 4 (1. 9 K) MCBY 8 32 (IR 1/5) 0. 088 0. 1 5270 tdb Individually powered quad Q 5 (1. 9 K) MQY 2 8 (IR 1/5) 4. 51 4. 88 74 0. 6 Orbit correctors Q 5 (1. 9 K) MCBY 6 24 (IR 1/5) 0. 072 -> 0. 088 0. 08 -> 0. 1 5270 tdb Individually powered quad Q 6 (4. 5 K) MQML 2 8 (IR 1/5) 4. 31 4. 66 21 0. 47 Orbit correctors Q 6 (4. 5 K) MCBC 2 8 (IR 1/5) 0. 08 0. 09 2840 tdb u Confirmed values PC Ramp Quad Ramp rate Acceleration Ramp Up Down References 2 Numbe [A/s] rate [A/s ] Time [s] r EDMS 1 10 2 486 569. 24 1375861 EDMS 4 ± 0. 667 ± 0. 25 149. 93 1375861 EDMS 1 10 2 488 616. 67 1375861 EDMS 4 ± 0. 667 ± 0. 25 119. 95 1375861 EDMS 1 12 2 388. 34 223. 41 1375861 EDMS 4 ± 0. 667 ± 0. 25 134. 94 1375861 Requirements [8]: § 50% ratio between Q 4 aperture due to three current leads. § sufficient for non-high β* optics § Fast discharge during ramp-down and precycle for quadrupoles. § Ramp-down values in the table = 5*L/R § Fast discharge during squeeze for Q 4/Q 5/Q 6 needed. § With 2 diodes in series for Q 5/Q 6 will balance IR 5/1 and IR 2 squeeze times at ≈200 s [16]. § Note: without diodes and 1 -quadrant PC squeeze would require 826 s (based on worse case scenario) [16]. § It could be tested (and beneficial) for LHC [minutes of 16] § The Q 4/5/6 ramp/acc. rates in the tables are almost below what specified in [1] for LHC. § Speed of correctors as the LHC ones (equal to circuits powering tests specs [1]). logo area D. Gamba et al. - 6 th HL-LHC Collaboration Meeting – Paris (FR) – 15/11/2016 11
11 T magnets expected trim requirements § They will be powered in series with main dipoles. § The magnetic response of these magnets is different than main dipoles. § Need of correctors or trim PC to compensate [2]. § Baseline is to use trim [4] Positive and Negative Currents [-250 A … 100 A] 6. 5 Te. V IRB = 11 k. A ΔI 11 T = -67 A 7 Te. V (Nominal) IRB = 11. 85 k. A ΔI 11 T = 0 A Courtesy of M. Giovannozzi 7. 5 Te. V (Ultimate) IRB = 12. 84 k. A ΔI 11 T = 91 A E ≈ 3. 5 Te. V (*) Preliminary curves – should be confirmed by operation – plan is to install the two units in LS 2 logo area From “Magnet powering and operation requirements” – S. Yammine (link) [4] D. Gamba et al. - 6 th HL-LHC Collaboration Meeting – Paris (FR) – 15/11/2016 12
11 T magnets 11 T Circuits for Hi. Lumi 11 T dipole, MBH Trim circuit Number of I_nomina Total number I_ultimate circuits per l (7 Te. V) of circuits [k. A] IP side [k. A] Magnet Type L per circuit [m. H] u Proposed changes R per Ramp PC Quad Ramp rate Acceleration Ramp Up circuit Down References Number [A/s] rate [A/s 2] Time [s] [mΩ] Time [s] 11 T dipole, MBH - 2 (IR 7) 11. 85 12. 798 15734 1 2 10 1 1300 - - 2 (IR 7) 0. 25 127. 1 15 4 ± 1 ± 0. 04 600 1300 § From typical trim function one would expect: § Ramp up time (0 -250 A) ≈ 600 s § Ramp rate: 250 A / 500 s ≈ 0. 5 A/s § Acc. rates: § 2*0. 5 A/s / 200 s = 0. 005 A/s 2 for I inversion. § 0. 5 A/s / 25 s = 0. 02 A/s 2 for end of ramp. § Assuming x 2 margin: § Ramp rate = 1 A/s; Acc. rate = 0. 04 A/s 2 Note that in case of trim failure at -250 A one should expect up to 5 σ aperture lost in the arc, and one would need 40 -80% of correctors strength to partially mitigate the effect [8]. logo area D. Gamba et al. - 6 th HL-LHC Collaboration Meeting – Paris (FR) – 15/11/2016 13
Updated table (08/11/2016) Number of Total I_nominal I_ultimate Magnet Type circuits per number of (7 Te. V) [k. A] IP side circuits Inner Triplet Circuits for Hi. Lumi D 2 D 1 Q 4 Ramp Up Time [s] Ramp Down References Time [s] 1 1 2 2 1 1 1 4 (IR 1/5) 8 (IR 1/5) 4 (IR 1/5) 16. 5 2 2 0. 12 1. 6 1. 47 0. 182 17. 82 2 2 0. 12 1. 73 1. 59 0. 2 255 69 69 58. 5 59 135 109 247 1247 0. 264 1. 44 13. 372 1. 512 1. 656 1. 728 9. 853 2 4 4 4 4 15 ± 2 ± 15 ± 15 ± 3 ± 1 ± 1 ± 5 ± 5 ± 1 1200 115. 34 106 *** ≈1300 115. 34 106 *** Superferric, order 3, normal and skew MCSXF / MCSSXF 2 8 (IR 1/5) 0. 105 0. 12 118 13. 372 4 ± 3 ± 1 *** 2 8 (IR 1/5) 0. 105 0. 12 152 13. 372 4 ± 3 ± 1 *** 2 8 (IR 1/5) 0. 105 0. 12 107 13. 372 4 ± 3 ± 0. 1 *** 1 1 4 (IR 1/5) 0. 105 0. 12 229 52 13. 372 4 4 ± 3 ± 0. 1 *** *** Superferric, order 6, skew Q 5 R per PC Quad Ramp rate Acceleration circuit Number [A/s] rate [A/s 2] [mΩ] MQXFA / MQFXB MCBXFB MCBXFA MQSXF Superferric, order 5, normal and skew Q 6 L per circuit [m. H] Proposed changes Hypothetical values Confirmed values To be defined Triplet Q 1, Q 2 a, Q 2 b, Q 3 Trim Q 1 Trim Q 3 Trim Q 2 a Orbit correctors Q 2 a/b - vertical Orbit correctors Q 2 a/b - horizontal Orbit correctors CP - vertical Orbit correctors CP - horizontal Superferric, order 2 Superferric, order 4, normal and skew 11 T u u MCOXF / MCOSXF MCDXF / MCDSXF MCTSXF Separation dipole D 1 MBXF 1 4 (IR 1/5) 12 12. 96 27 0. 27 1 20 2 648 500 Recombination dipole D 2 MBRD 1 4 (IR 1/5) 12 12. 96 25 0. 234 1 20 2 648 534. 19 Orbit correctors D 2 MCBRD 4 16 (IR 1/5) 0. 54 600 1. 08 4 ± 2 ± 1 270 Individually powered quad Q 4 (1. 9 K) MQY 2 8 (IR 1/5) 4. 5 4. 86 74 0. 65 1 10 2 486 569. 24 EDMS 1375861 Orbit correctors Q 4 (1. 9 K) MCBY 8 32 (IR 1/5) 0. 088 0. 1 5270 tdb 4 ± 0. 667 ± 0. 25 149. 93 EDMS 1375861 Individually powered quad Q 5 (1. 9 K) MQY 2 8 (IR 1/5) 4. 51 4. 88 74 0. 6 1 10 2 488 616. 67 EDMS 1375861 Orbit correctors Q 5 (1. 9 K) MCBY 6 24 (IR 1/5) 0. 088 0. 1 5270 tdb 4 ± 0. 667 ± 0. 25 119. 95 EDMS 1375861 Individually powered quad Q 6 (4. 5 K) MQML 2 8 (IR 1/5) 4. 31 4. 66 21 0. 47 1 12 2 388. 34 223. 41 EDMS 1375861 Orbit correctors Q 6 (4. 5 K) MCBC 2 8 (IR 1/5) 0. 08 0. 09 2840 tdb 4 ± 0. 667 ± 0. 25 134. 94 EDMS 1375861 11 T dipole, MBH - - 2 (IR 7) 11. 85 0. 25 12. 798 0. 25 15734 127. 1 1 15 2 4 10 ± 1 1 ± 0. 04 1300 600 1300 11 T dipole, MBH Trim circuit logo area *** Ramp up/down is not the main parameter: important is the length of the degaussing cycles! D. Gamba et al. - 6 th HL-LHC Collaboration Meeting – Paris (FR) – 15/11/2016 14
Power Converters ripple and instabilities logo area D. Gamba et al. - 6 th HL-LHC Collaboration Meeting – Paris (FR) – 15/11/2016 15
From PC output to magnetic field [6][8][12] PCs work in two regimes: § current control (f < 0. 1 Hz (or f 0)): § voltage control (f > 0. 1 Hz (or f 0)): with: current ripple (Power converter specifications) voltage ripple (Power converter specifications) transfer function of the load (circuit) seen by the power converter: transfer function from magnet input current to magnetic field § Can be assumed to be constant only in current control regime [12] § At high frequency it is affected by eddy currents and other “losses” [12] transfer function of the cold bore, absorber, beam screen etc. , TVacuum ≤ 1 (not (yet) taken into account) § input form WP 3/6 expected [12] logo area D. Gamba et al. - 6 th HL-LHC Collaboration Meeting – Paris (FR) – 15/11/2016 16
Expected voltage spectrum for HL-LHC [6][8] current control: Class 1 PC: 1 ppm** Class 2 PC: 10 ppm 300 Hz voltage control (Class 1/2 PC): § 50 Hz harmonics (main grid): 50 Hz: 3. 2 m. V R. M. S. 100 Hz: 0. 8 m. V R. M. S. § 300 Hz harmonics (diode rectifier): 300 Hz: 10. 0 m. V R. M. S. 600 Hz: 2. 5 m. V R. M. S. § 20 k. Hz harmonics (ITPT converters): 20 k. Hz: 10. 0 m. V R. M. S. 40 k. Hz: 2. 5 m. V R. M. S. § 10 MHz harmonics: 10 MHz: 1. 0 m. V R. M. S. (0. 5 m. V) § all other frequencies: 0. 5 m. V R. M. S 50 Hz 20 k. Hz 600 Hz 40 k. Hz 100 Hz 10 MHz Courtesy EPC group § 50 Hz harmonics could controlled to the desired value (J-P. Burnet) § More refined spectrum is being evaluated by EPC group. ** From more recent estimation, Class 1 PC should have 0. 1 ppm-σ Gaussian jitter From “Update oflogo the requirements on circuits parameters from beam physics” – R. De Maria– (link) [8] area D. Gamba et al. - 6 th HL-LHC Collaboration Meeting – Paris (FR) – 15/11/2016 17
Power supply ripple: why important? (1/2) § It is particularly relevant for magnets at locations with high beta function (e. g. IT) also due to their large number (e. g. main quadrupoles, dipoles…) § LHC (4. 5 km bmax): long efforts for the determination of maximum tolerable tune ripple from power converters. LHC -> d. Qtotal < 10 -4 § Tolerance based on: § Long-term tracking studies § Experimental studies (SPS) and HERA experience § HL-LHC (21. 7 km - round, 43 km - flat - bmax): new regime to be explored. § Tolerance based on analysis of the performance of K-modulation method for b* control HL-LHC -> d. Qtotal < 10 -5 Courtesy F. Carlier logo area From “Beam Dynamics requirements” - M. Giovannozzi – (link) [6] D. Gamba et al. - 6 th HL-LHC Collaboration Meeting – Paris (FR) – 15/11/2016 18
Power supply ripple: current control regime [6] Assuming Class 1 PC: ± 1 ppm** Tune shift induced by uniformly distributed random error on current: § With class 2 PC one should expect x 10 amplification § Class 1 PC for D 1/D 2 needed also for keeping jitter at IP << 1 σ [8] § § Same for correctors MCBXFA/FB/RD/YY since just factor 3 to 10 smaller than D 1/D 2 [8] trims not taken into account but expected to be negligible Round optics HL-LHC (IT nominal LHC) IT Q 1 -Q 2 -Q 3 D 1 L D 2 L D 1 R D 2 R Q 4* rms((Qz-Qz 0)) [10 -4] rms((z-z 0)) at IP [μm] 0. 25 ? 0. 67 (0. 17**) ? 0. 027 (0. 007**) 0. 040 (0. 01**) ? ? Note that from previous slide we would need d. Qtotal < 10 -5 *Q 4 is now an existing MQY: Previous simulations were performed assuming MQYY (edms) ** New estimation: Class 1 PC should only have 0. 1 ppm-σ Gaussian jitter [J. Coello de Portugal] logo area D. Gamba et al. - 6 th HL-LHC Collaboration Meeting – Paris (FR) – 15/11/2016 19
Power supply ripple: why important? (2/2) Tune modulation induces resonance sidebands [22 -23]: slow modulation (e. g. 50 Hz): distances between the sidebands are small but amplitudes decrease only slowly with increasing order fast modulation (e. g. 600 Hz): distances between the sideband are large and amplitudes decrease rapidly with increasing order slow+fast modulation: the sidebands of the fast modulation form the seeds for the sidebands of the slow modulation (“seeding resonances”) => additional resonances from PC ripple can lead to higher diffusion rates, losses and ultimately emittance growth. Limits are usually around tune shift of �� Q=10 -6 up to �� Q=10 -4 (simulations + experiments SPS and HERA). logo area From “Impact of low frequency noise on emittance blow-up” – M. Fitterer (link) [9] [18] D. Gamba et al. - 6 th HL-LHC Collaboration Meeting – Paris (FR) – 15/11/2016 20
Power supply ripple § Tolerances defined by that no degradation of DA should be observed from PC ripple § Conclusion in terms of emittance growth limits from PC ripple in HL-LHC: 300 Hz, 600 Hz colored lines: tune ripple for expected frequency spectrum and different powering schemes orange bars: limits from DA simulations with single frequency • already tune modulation of 10 -5 to 10 -6 could be relevant for HL-LHC (higher diffusion in tails and also emittance growth) • The critical frequencies (300/600 Hz) might be reduced according to PC experts. logo area From “Impact of low frequency noise on emittance blow-up” – M. Fitterer (link) [9] D. Gamba et al. - 6 th HL-LHC Collaboration Meeting – Paris (FR) – 15/11/2016 21
Conclusions (1/2) – Operational Requirements § Key aspects are: § the squeeze and rump-down time for maximising the integrated luminosity. § We took as a reference 20’ for ramp-down, even though some magnets (e. g. at IR 2/8) will still require about 30’. § Degaussing cycle for superferric correctors needs to be in the shadow of ramp-down/precycle. § Speed of orbit collapse for avoiding beam-beam driven instabilities. § Need of fast MCBX/MCBRD orbit corrector to be compatible with orbit feedback. § Limiting factor also for the LHC. To be discussed. logo area D. Gamba et al. - 6 th HL-LHC Collaboration Meeting – Paris (FR) – 15/11/2016 22
Conclusions (2/2) – Beam Dynamics § Clear requirement of class 1 PC for the main elements (quadrupoles, D 1/D 2, correctors) § Even with latest (promising!) estimation of low frequency noise (<0. 1 ppm-σ), k-modulation measurements could be still affected. § New simulations required to quantify the contribution from matching section quadrupoles (e. g. Q 4 now a MQY) § Potential orbit jitter at the IP to be studied. § Earlier DA studies showed potential issues, with emphasis on 300 and 600 Hz frequencies. § The situation needs to be re-assessed using the new layout and circuits specifications. § New model/measurements of ΔB/B 0(ΔI/I 0) at f>f 0 are expected from EPC and MSC for better estimate the actual ripple spectra [12]. § We are working on finalising the requirements in an official note. logo area D. Gamba et al. - 6 th HL-LHC Collaboration Meeting – Paris (FR) – 15/11/2016 23
Thank you for your attention. logo area D. Gamba et al. - 6 th HL-LHC Collaboration Meeting – Paris (FR) – 15/11/2016 24
Used references 1) 2) 3) 4) 5) 6) 7) 8) 9) 10) 11) 12) 13) 14) 15) 16) 17) 18) 19) 20) 21) 22) 23) Parameters for LHC Superconducting Circuits Powering Tests (08/01/2016) (edms) Conceptual Design Review of the HL-LHC Magnet Circuits (21 -23/03/2016) (report) Magnet Circuit Forum (MCF) (espace) Review of the 11 T dipoles at Collimator Section for the HL-LHC (06 -08/04/ 2016) (report) HL_TDR_V 07. 2016. 10. 04. Version 17. 05. pdf (edms) “Operation and beam dynamics requirements„ - M. Giovannozzi 22/03/2016 (link) Beam Dynamics requirements - M. Giovannozzi 19/05/2016 (link) Update of the requirements on circuits parameters from beam physics - R. De Maria 19/05/2016 (link) Impact of low frequency noise on emittance blow-up – M. Fitterer 20/05/2016 (link) Magnets for Insertion Regions (espace WP 3) HL-LHC Circuits and PC nomenclature – S. Yammine 01/11/2016 (link) Modeling of the PC Output to the Magnetic Field – M. Martino 01/11/2016 ( link) Optics transition between injection and collision optics for the HL-LHC upgrade project – M. Korostelev et al. IPAC 2013 (link) Transition between injection optics and collision optics (including ATS squeeze) – M. Korostelev 14/11/2013 (link) IR 6 squeeze for layout HLLHCV 1. 2 – M. Fitterer 09/09/2015 (link) Analysis of time required for pre-squeeze and squeeze – Q. King 19/02/2015 (link) Beam-beam simulations for IT PC tolerances – M. Fitterer 09/03/2015 (link) Powering schemes for triplet and matching sections – M. Fitterer 20/11/2014 (link) Separation collapsing speed for HL-LHC – C. Tambasco 19/04/2016 (link) HL-LHC Optics and layout update – R. De Maria 02/09/2016 (link) International Review of the Inner Triplet Quadrupoles – 07 -10/06/2016 ( link) O. S. Brüning, F. Willeke, Phys. Rev. Lett. 76, No. 20 (1995) O. S. Brüning, Part. Acc. 41, pp. 133 -151 (1993) logo area D. Gamba et al. - 6 th HL-LHC Collaboration Meeting – Paris (FR) – 15/11/2016 25
Additional slides logo area D. Gamba et al. - 6 th HL-LHC Collaboration Meeting – Paris (FR) – 15/11/2016 26
Faster MCBX/MCBRD correctors requirements § Trying to get similar ramp rates and accelerations rates in terms of angle for all correctors. [R. De Maria]. § Similar specifications also for D 1/D 2. [R. De Maria]. HLLHC MCBXFA MCBXFB MCBRD LHC MCBY MBXF MBRD MCBX MCBY MBX MBRC Integrated field [T m] 4. 50 2. 50 5. 00 2. 79 35. 00 1. 51 2. 25 39. 69 Nom. Current [A] 1600 430 88 12000 550 72 5750 6050 Ramp rate [A/s] 15. 00 2. 00 0. 67 20. 00 2. 50 0. 67 18. 00 13. 00 Field Rate [m. Tm/sec] 42. 19 23. 44 23. 26 21. 15 58. 33 6. 85 20. 84 124. 25 85. 28 Angle Rate [murad/sec@7 Te. V] 1. 81 1. 00 0. 91 2. 50 0. 29 0. 89 5. 32 3. 65 Ramp Acc. [A/s 2] 5. 00 1. 00 0. 25 2. 00 0. 20 0. 25 2. 00 Field Acc. [m. Tm/sec 2] 14. 06 7. 81 11. 63 7. 93 5. 83 0. 55 7. 81 13. 12 Angle Acc. [murad/sec 2@7 Te. V] 0. 60 0. 33 0. 50 0. 34 0. 25 0. 02 0. 33 0. 59 0. 56 Time to full rate [sec] 3. 00 2. 67 10. 00 12. 50 2. 67 9. 00 6. 50 logo area D. Gamba et al. - 6 th HL-LHC Collaboration Meeting – Paris (FR) – 15/11/2016 27
From PC output to magnetic field [12] § A more accurate model: logo area From “Modeling of the PC Output to the Magnetic Field” – M. Martino (link) [12] D. Gamba et al. - 6 th HL-LHC Collaboration Meeting – Paris (FR) – 15/11/2016 28
D. Gamba et al. - 6 th HL-LHC Collaboration Meeting – Paris (FR) – 15/11/2016 “Modeling of the PC Output to the Magnetic Field” – M. Martino (link) [12] 29 Courtesy TE-MSC C. Giloux - 2009 Dramatic effect due to the Beam Screen? for SC: “circuit” only: C output to magnetic field [12] logo area
Power supply ripple (2) [6] [17 -18] § In this case the tolerances on: § Ripple amplitude of the overall spectrum § Determination of most dangerous frequencies § Have been determined with long-term simulations to assess the impact on the ripple on the dynamic aperture. § Several configurations have been probed: § Without beam-beam § With beam-beam and with/without crab crossing, εN=3. 75 μm, Nb=2. 2 x 1011, 4 D beam-beam lens § With beam-beam and crab crossing, εN=2. 5 μm, Nb=2. 2 x 1011, 6 D beam-beam lens § optics: s. LHCV 3. 1 b, β*=15 cm in IR 1/5, β*=10 m in IR 2/8 § max number of turns: 106 § Criterion for effect of ripple: DA changes in respect to reference cases without ripple logo area D. Gamba et al. - 6 th HL-LHC Collaboration Meeting – Paris (FR) – 15/11/2016 30
Power supply ripple (2) [6][17 -18] § 300 and 600 Hz have a sizeable impact on dynamic aperture if the amplitude of tune modulation is 10 -4. § With beam-beam, the realistic spectrum starts showing an impact on dynamic aperture if its amplitude is multiplied by a factor of 10. Reminder: sixty realisations of magnetic field errors are used. Continuous lines: average (over realisations) DA Dashed lines: minimum or maximum (over realisations) DA Physical mechanism and extrapolation to lifetime drop of the 300/600 Hz effect is under study [8] logo area From “Beam Dynamics requirements” - M. Giovannozzi – (link) [6] D. Gamba et al. - 6 th HL-LHC Collaboration Meeting – Paris (FR) – 15/11/2016 31
Notes: § From “Report from the Review Panel for the Conceptual Design Review of the HL-LHC Magnet Circuits (21 -23 March 2016) (report)”: § Suggested to better investigate the need of a single PC+3 trims for the triplet instead of 2 PC+2 trims. § Decision was taken at “HL-TCC of 28/04/2016” (link). § See also presentation at MCF from Arjan on 9/8/2016 (link) § Suggested to investigate the option of not having independent trim for the 11 T magnets and to use correctors for compensating the different strength. § Decision was taken at 06/04/2016 (link) to use the trim. § Going into collision from 2σ to 1σ < 1 s (E. Me tral, 19/05/2016 link) § In the minutes it is stated that the actual requirement is to go from 2 to 0σ in less than 3 s logo area D. Gamba et al. - 6 th HL-LHC Collaboration Meeting – Paris (FR) – 15/11/2016 32
- Slides: 32