1. Philip Burrows ALCPG11, Eugene 21/03/111 Performance of FONT at ATF2 Philip Burrows John Adams Institute Oxford University

2. Philip Burrows ALCPG11, Eugene 21/03/112 Feedback On Nanosecond Timescales Beam-based FB R&D for future Linear Colliders Philip Burrows Glenn Christian Javier Resta Lopez Colin Perry Graduate students: Ben Constance Robert Apsimon Douglas Bett Alexander Gerbershagen Michael Davis Valencia, CERN, DESY, KEK, SLAC

3. Philip Burrows ALCPG11, Eugene 21/03/113 Outline Brief reminder of intra-train feedback system Implementation in ILC + CLIC IRs Prototype hardware development (FONT systems) Summary + outlook

4. Philip Burrows ALCPG11, Eugene 21/03/114 IP intra-train feedback system - concept Last line of defence against relative beam misalignment Measure vertical position of outgoing beam and hence beam-beam kick angle Use fast amplifier and kicker to correct vertical position of beam incoming to IR FONT – Feedback On Nanosecond Timescales (Oxford, Valencia, CERN, DESY, KEK, SLAC)

5. Philip Burrows ALCPG11, Eugene 21/03/115 CLIC FD region Elsner

6. Philip Burrows ALCPG11, Eugene 21/03/116 Spool tubes Reacting bars QF QD0 Cam/support ILC Final Doublet Region (SiD for illustration) Oriunno

7. Philip Burrows ALCPG11, Eugene 21/03/117 Remaining issues Engineering of real hardware optimised for tight spatial environment: BPM, kicker, cables … Further studies of radiation environment for FB: was studied for ILC, less so for CLIC; where to put electronics? need to be rad hard? shielded? off to side? EM interference: beam  FB hardware kicker  detector

8. Philip Burrows ALCPG11, Eugene 21/03/118 Prototyping status

9. Philip Burrows ALCPG11, Eugene 21/03/1199 KickerBPM Analogu e feedback Analogue BPM processor Drive amplifier e- FONT system prototype (schematic) CLIC

10. Philip Burrows ALCPG11, Eugene 21/03/1110 KickerBPM 1 Analogue BPM processor BPM 2 BPM 3 e- FONT3 CLIC prototype at KEK/ATF (2004-5) 56ns train of bunches separated by 2.8ns

11. Philip Burrows ALCPG11, Eugene 21/03/1111 KickerBPM Digital feedback Analogue BPM processor Drive amplifier e- FONT system loop (schematic) ILC

12. Philip Burrows ALCPG11, Eugene 21/03/1112 KickerBPM 1 Digital feedback Analogue BPM processor Drive amplifier BPM 2 BPM 3 e- FONT4 ILC prototype at KEK/ATF (2006-9) 300ns train of bunches separated by 150ns

13. Philip Burrows ALCPG11, Eugene 21/03/1113 FONT5 location

14. Philip Burrows ALCPG11, Eugene 21/03/1114 FONT5 schematic QD10XQF11XQD12XQD14XQF13XQF15XK1K2 P2P3P1 To dump FB board DAQ P2  K1 (‘position’) P3  K2 (‘angle’) P3  K1

15. Philip Burrows ALCPG11, Eugene 21/03/1115 KickerBPM Digital feedback Analogue BPM processor Drive amplifier e- Each FONT5 system loop 300ns train of bunches separated by 150ns

16. Philip Burrows ALCPG11, Eugene 21/03/1116 FONT5 beamline hardware 3 new BPMs and 2 new kickers installed in new ATF2 extraction line February 2009; BPM movers installed 2010

17. 6 New FONT5 digital FB board Xilinx Virtex5 FPGA 9 ADC input channels (TI ADS5474) 4 DAC output channels (AD9744) Clocked at 357 MHz phase-locked to beam 4x faster than FONT4

18. Philip Burrows ALCPG11, Eugene 21/03/1118 One damping ring cycle (463ns) data returned each pulse: RS232 over ethernet All BPM sum (charge) and difference signals Absolute sample time adjustable in 70ps taps: accurate peak sampling Ratio of difference to sum peaks gives y-position Pedestal subtraction w. on-board trim DACs (no latency gain) FONT5 DAQ

19. Philip Burrows ALCPG11, Eugene 21/03/1119 Outline of FB results Latency Basic loop performance Banana correction Coupled-loop FB results Next steps

20. 6 FONT5 latency: P2  K1 loop Latency 133ns

21. Philip Burrows ALCPG11, Eugene 21/03/1121 Latency estimate

22. 6 P2  K1 loop performance With banana correction

23. 6 P2  K1 loop performance Incoming position scan

24. Philip Burrows ALCPG11, Eugene 21/03/1124 P2  K1 loop jitter reduction Bunch 1 Bunch 2 2.1 um  0.4 um Factor of 5 jitter reduction

25. Philip Burrows ALCPG11, Eugene 21/03/1125 Bunch 2 jitter vs. gain

26. Philip Burrows ALCPG11, Eugene 21/03/1126 Bunch 1-2 correlations Bunch 1 Bunch 2 Feedback removes bunch correlations

27. Philip Burrows ALCPG11, Eugene 21/03/1127 Bunch 1-2 correlations vs. gain

28. Philip Burrows ALCPG11, Eugene 21/03/1128 0.4 micron jitter propagation P2 Y [micron]

29. Philip Burrows ALCPG11, Eugene 21/03/1129 Jitter propagation to ATF2 IP 4 Counts Assuming perfect lattice, no additional jitter sources (!)

30. Philip Burrows ALCPG11, Eugene 21/03/1130 Summary so far These spectacular results were obtained with beam of exceptional quality: Incoming train jitter: 2um Bunch 1-2 correlations:96% Bunch 2-3 correlations:80% This is NOT typical!

31. P2  K1 loop jitter reduction Bunch 1 Bunch 2 Bunch 3 13 um  5 um  3 um

32. Philip Burrows ALCPG11, Eugene 21/03/1132 FB simulation: P2-K1+P3-K2 coupled Bunch 1 Bunch 2

33. Philip Burrows ALCPG11, Eugene 21/03/1133 Further Feedback Tests K1 – P2 loop K2 – P3 loop K1 – P2 + K2 – P3 uncoupled K1 – P2 + K2 – P3 coupled

34. Philip Burrows ALCPG11, Eugene 21/03/1134Philip Burrows X-band 2010 Daresbury 2/12/1034 K1 – P2 loop gain scan

35. Philip Burrows ALCPG11, Eugene 21/03/1135Philip Burrows X-band 2010 Daresbury 2/12/1035 K2 – P3 loop gain scan

36. Philip Burrows ALCPG11, Eugene 21/03/1136 K1 – P2 + K2 – P3 coupled: K1 gain scan

37. Philip Burrows ALCPG11, Eugene 21/03/1137 K1 – P2 + K2 – P3 coupled: K2 gain scan

38. Coupled loop jitter reduction Bunch 1 Bunch 2 Bunch 3 9.4 um  5.6 um  4.4 um

39. Philip Burrows ALCPG11, Eugene 21/03/1139 Summary: FONT5 IP feedback concept well advanced Prototype meets ILC technical requirements in terms of BPM resolution, kicker drive and latency Future effort focussed on achieving ATF2 goals

40. Philip Burrows ALCPG11, Eugene 21/03/1140 IPBPM configuration

41. Philip Burrows ALCPG11, Eugene 21/03/1141 IPBPM configuration

42. Philip Burrows ALCPG11, Eugene 21/03/1142 Kicker location

43. Philip Burrows ALCPG11, Eugene 21/03/1143 Some working assumptions (1) Kicker centre ~ 0.5m upstream of IPBPM Kicker aperture 40mm (?) Kicker length ~ 15 cm (?) Matched 50 Ohm terminations  Half of current FONT5 sensitivity: 0.5 urad / Amp (can easily scale from above assumptions)

44. Philip Burrows ALCPG11, Eugene 21/03/1144 Some working assumptions (2) Dynamic correction range: Beam size 37 nm Beam y jitter ~ beam size (?) 2 sigma correction  70 nm @ IP = 140 nrad kick  drive current = 0.15/0.5 ~ 0.3 A (per strip)

45. Philip Burrows ALCPG11, Eugene 21/03/1145 Some working assumptions (3) Amplifier: Peak power = 0.3 **2 x 50 = 5 W per strip  eg. Minicircuits: 10W, 5  500 MHz Low latency (5 ns) Output can be pulsed for long bunch train No margin for kick

46. Philip Burrows ALCPG11, Eugene 21/03/1146 IPBPM electronics (Aeyoung Heo)

47. Philip Burrows ALCPG11, Eugene 21/03/1147 Digitisation of IPBM signal Digitise I and Q signals Derive amplitude and phase  charge-independent position signal FONT5 ADCs (TI ADS5474) clocked at 357 MHz Very high bandwidth sample point Sample time adjustment sensitivity c. 100ps

48. Philip Burrows ALCPG11, Eugene 21/03/1148 FONT Digitisation IQ Joshi

49. Philip Burrows ALCPG11, Eugene 21/03/1149 Digitisation of IPBM signal Digitise I and Q signals Derive amplitude and phase  charge-independent position signal FONT5 ADCs (TI ADS5474) clocked at 357 MHz Very high bandwidth sample point Sample time adjustment sensitivity c. 100ps Up- and downstream IPBPM signals needed

50. Philip Burrows ALCPG11, Eugene 21/03/1150 Latency estimate Amplifier (as described) 5ns Kicker fill (15cm) 0.5ns Beam flight time amplifier  IPBPM 2ns Cables (3 x 1.5m?) 23ns IPBPM electronics40ns? Digital processing60ns Total 131ns

51. Philip Burrows ALCPG11, Eugene 21/03/1151 Summary of ATF2 IP FB Conceptual design for IP FB system System parameters look feasible Critical parameters: dynamic correction range, bunch spacing Digitisation of IPBPM I and Q signals is easiest approach Technical details need to be finalised: locations of BPM + kicker, kicker aperture, cable runs …

52. Philip Burrows ALCPG11, Eugene 21/03/1152 Extra material Conceptual design for IP FB system System parameters look feasible Critical parameters: dynamic correction range, bunch spacing Digitisation of IPBPM I and Q signals is easiest approach Technical details need to be finalised: locations of BPM + kicker, kicker aperture, cable runs …

53. Philip Burrows ALCPG11, Eugene 21/03/1153 P2  K1 loop jitter reduction Bunch 1 Bunch 2 Bunch 3 2.1 um  0.4 um  0.8 um

54. Philip Burrows ALCPG11, Eugene 21/03/1154 Bunch correlations Bunch 1 Bunch 2 Bunch 3