1. Space charge studies at the CERN PS A. Huschauer, Raymond Wasef H. Damerau, S. Gilardoni, S.Hancock, D. Schoerling, R. Steerenberg Acknowledgements: All PS/PSB operators MSWG, May 7, 2013

2. 1 A. HUSCHAUER and R. WASEF, MSWG, May 7, 2013 Reminder † : - Control of the working point - Identification of resonances Free space within the tune diagram Compensation of resonances Intermediate conclusion Space charge at injection Measurement settings 4 th order resonance (4q y =1) Results Conclusion Overview † MSWG, August 17, 2012

3. 2 A. HUSCHAUER and R. WASEF, MSWG, May 7, 2013 Reminder: control of the working point Two different ways of controlling the tune Pole Face Windings (PFW) and Figure of 8 Loop (F8L) o in total 5 circuits to control tunes, linear chromaticities and (in theory) one of the second order chromaticities Low Energy Quadrupoles (LEQ) o 2 families: focusing/defocusing o RMS current limited to 10 A ξ x = -0.83 ξ y = -1.12 Bare machine no LEQ, no PFW very linear behaviour LEQ don’t influence linearity PFW significant alteration of linear machine unbalanced narrow and wide circuits Q x =6.10, Q y =6.20 Q x =6.25, Q y =6.28

4. 3 A. HUSCHAUER and R. WASEF, MSWG, May 7, 2013 Reminder: identification of resonances * A benchmarking experiment in SIS18 for dynamic aperture induced beam loss, GSI-Acc- Note-2004-05-001 Measurement concept technique first used by G. Franchetti et al. * at GSI in 2004 only beam loss considered to identify resonances  large normalized transverse emittances (ε x1σ ≈ 10 mm·mrad, ε y1σ ≈ 8 mm·mrad), small tune spread (ΔQ x ≈ -0.05, ΔQ y ≈ -0.07) tune in one plane kept constant, in the other dynamically ramped intensity recorded derivative calculated each peak normalized by intensity before the respective resonance

5. 4 A. HUSCHAUER and R. WASEF, MSWG, May 7, 2013 all steps repeated for different constant tune interpolation of whole set of data on equidistant grid color scaling informs about losses First measurement at 1.4 GeV LEQ active to control the tunes estimated tune spread: ΔQ x ≈ -0.05, ΔQ y ≈ -0.07 strongest observed resonance: 2q x +q y =1 vertical tune constant horizontal tune constant important for machine operation: 3q y =1 Reminder: identification of resonances

6. 5 A. HUSCHAUER and R. WASEF, MSWG, May 7, 2013 Measurements with PFW at 2 GeV only 4 out of the 5 available circuits powered  F8L fixed to 0 A (4 current mode) same resonances excited as in measurements on previous slide  no additional excitation by PFW Reminder: identification of resonances

7. 6 A. HUSCHAUER and R. WASEF, MSWG, May 7, 2013 Free space within the tune diagram Combination of both scans LHC beam today ≈ -0.28 HL-LHC demand > 0.34 tune spread operational area additional margin measurements suggest possibility to increase working point but even then: available area not large enough to accommodate HL-LHC beam  resonance compensation

8. 7 A. HUSCHAUER and R. WASEF, MSWG, May 7, 2013 Compensation of resonances resonances excited by the bare machine  effect of the main magnets 2D simulation campaign with OPERA to obtain magnetic errors (normal and skew) at 1.4 GeV due to mechanical errors of the yoke and alignment tolerances of the main coil errors implemented in MAD and PTC is then used to calculate driving terms of the resonances and the corresponding correction currents Main coils :  = 3 mm F8L:  = 1 mm PFW:  = 0.7 mm Iron yoke (in y-direction) :  = 0.02/3 mm 26 GeV/c

9. 8 A. HUSCHAUER and R. WASEF, MSWG, May 7, 2013 Compensation of resonances resonances 3q y =1 and 2q x +q y =1 are caused by skew sextupole components of the magnetic field installation of 4 skew sextupoles (independent power supplies) during the short winter shutdown cause of the magnetic errors not completely understood: contribution of fringe fields, fields in the junctions between the single blocks of the main magnet,…

10. 9 A. HUSCHAUER and R. WASEF, MSWG, May 7, 2013 Compensation of resonances at 2 GeV beforeafter Compensation of 3q y =1 additionally: reduction of 2q x +q y =1 vertical tune constant horizontal tune constant

11. 10 A. HUSCHAUER and R. WASEF, MSWG, May 7, 2013 Compensation of resonances at 2 GeV beforeafter Compensation of 2q x +q y =1 3q y =1 is clearly enhanced vertical tune constant horizontal tune constant

12. 11 A. HUSCHAUER and R. WASEF, MSWG, May 7, 2013 Intermediate conclusion suggested by the presented measurements: the resonance 3q y =1 constitutes the major limit for increasing the space charge tune spread resonance compensation successfully implemented for the shown cases

13. 12 A. HUSCHAUER and R. WASEF, MSWG, May 7, 2013 Space Charge at injection (1.4 GeV) Current injection energy: 1.4 GeV Estimated tune-spread of current operational beam~(0.2 ; 0.28) LHC double batch injection: Long flat bottom: 1.2s HL-LHC beams requirement: tune-spread~.34 -.37  Importance of the study of the integer resonance effect and the tune spread on beams 1.2s 1 st Injection 170ms 2 nd Injection 1370ms Current operation area

14. 13 A. HUSCHAUER and R. WASEF, MSWG, May 7, 2013 Measurement settings The tune-spread was varied using an adiabatic bunch compression (20ms). The effect of the integer is observed through the longitudinal and transverse profiles as well as losses. The maximum tune-shift due to space charge is estimated using: After a quick check of the losses and emittances after compression, (6.23 ; 6.255) has been chosen as starting working point. (Measured ~(6.228 ; 6.253) ) While setting the beam, significant losses have been noticed near 4qy=1 while none were noticed during the tune diagram measurement. The 4 th order resonance seems to be excited and to have an intensity threshold. Previous study of the integer (CERN/PS 93-18)

15. 14 A. HUSCHAUER and R. WASEF, MSWG, May 7, 2013 4 th order Resonance Testing if the 4qy=1 is excited by Space Charge: - Bunch compression @ C190 - Tune step between C500 and C800 Set tunes Measured tunes 4 different settings:  Beam 1: I=115 e10 ppb Tune-spread =(.22 ;.4) (for Q21Q23 optics)  Beam 2: I=80 e10 ppb Tune-spread =(.18 ;.37) (for Q21Q23 optics)  Beam 3:I=35 e10 ppb Tune-spread =(.08 ;.24) (for Q21Q23 optics)  Beam 4: I=115 e10 ppb Debunched Time[ms]

16. 15 A. HUSCHAUER and R. WASEF, MSWG, May 7, 2013 4 th order Resonance  The 4 th order resonance seems to be excited by space charge

17. 16 A. HUSCHAUER and R. WASEF, MSWG, May 7, 2013 Adiabatic Bunch Compression Injectio n @ C170 Beginning of compression @ C190 End of compression @C210 Measurements @ flat bottom of 1.2s RF Voltage [kV] Time [ms]

18. 17 A. HUSCHAUER and R. WASEF, MSWG, May 7, 2013 Horizontal emittance behavior Beam used: I=1.15e12 ppb; ε h,normalized =1.6μm; ε v,normalized =1.25μm; Δp/p(1σ)= 0.95E-3 ; full bunch length=185ns Since the horizontal detuning is always less than.23 (Qx=6.23), no relevant change has been noticed in the horizontal plane. Therefore, only the vertical emittance is shown in the following results

19. 18 A. HUSCHAUER and R. WASEF, MSWG, May 7, 2013 Vertical growth vs. Time vs. Tune-spread Working point : (6.23 ; 6.255)

20. 19 A. HUSCHAUER and R. WASEF, MSWG, May 7, 2013 Vertical growth vs. Time vs. Tune-spread Most of the growth happens during the first 200ms.  Single batch injection could be a possible solution for the LHC beams, the reduce the blow-up at injection.

21. 20 A. HUSCHAUER and R. WASEF, MSWG, May 7, 2013 Vertical growth vs. Tune-spread vs. Losses

22. 21 A. HUSCHAUER and R. WASEF, MSWG, May 7, 2013 Vertical growth vs. Tune-spread vs. Losses The beam tune-spread is trapped between the 4q y =1 and the integer.  If one increases the vertical tune to avoid growth due to the integer, the losses increase because of the 4 th order resonance  There are less losses with higher tune-spread because the proton population becomes smaller on the 4q y =1 after compression.  The choice of the working point is a compromise between losses and emittance blow-up

23. 22 A. HUSCHAUER and R. WASEF, MSWG, May 7, 2013 Conclusion The 4q y =1 resonance seems to be excited by space charge. With the current scheme and limitations it is very challenging to meet the HL- LHC requirements (5% blow-up, 5% losses budgets and ΔQ~.34 -.37) Potential solutions: Decrease the harm of resonances A compensation scheme of the 3q y =1 resonance has been successful. Correcting the 4q y =1, could then allow operation of a very large tune-spread beam. Shortening significantly the flat-bottom. Decrease the tune-spread Use of flat bunches to decrease the tune-spread. New optics with a larger dispersion to decrease the tune-shift. (since the beam pipe is much larger than the LHC beams)

24. THANK YOU FOR YOUR ATTENTION!

25. Backup Slides

26. Longitudinal profile before/after compression Before compression After compression

27. Effect of integer resonance Previous study of the integer (CERN/PS 93-18) With While I am using: In this study the tune spread has been estimated as following: The effect of the integer resonance on high space charge beams is very important since most operational beams are close to/on the integer and the HL-LHC Beams will have even a larger tune-spread. Raymond WASEF, Space Charge Workshop, 16/04/13, CERN

28. IV. MD: New Optics  New Optics during the flat bottom for the LHC double batch injection beams. Current optics for LHC beams (ε normalized =2.5μm; Δp/p= 1 E - 3): -Horizontal Size (1σ) < 4.5mm (while beam pipe size ~ 146mm) -Vertical Size (1σ) < 3.5mm (while beam pipe size ~ 70mm)  Changing the optics by using the transition triplets: Increase of the beam size and therefore decreased tune-spread For one of the future options for the High Brightness LHC-25 beam with: 3.35 E 12 ppb ; ε 1σ,normalized =2μm; Δp/p (1σ) = 1 E -3 ; full bunch length=180ns ; E=2GeV. Tune-spread for current optics = (0.28 ; 0.37) Tune-spread for suggested optics = (0.15 ; 0.28) 11

29. Examples of Operational Beams (1.4GeV) Currently no significant emittance blow-up nor losses are observed for operational beams that cannot be cured by increasing the vertical tune and adapting the horizontal to remain near the diagonal (recent change Qx: 6.21->6.235, Qv: 6.23-> 6.245) 6