Feedback on Nanosecond Timescales maintaining luminosity at future
Feedback on Nanosecond Timescales: maintaining luminosity at future linear colliders Ben Constance John Adams Institute, University of Oxford Io. P HEPP/APP annual meeting 2010
Introduction • The problems associated with achieving high luminosity at linear colliders • Focus on ‘cold’ colliders – International Linear Collider bias… • Feedback system requirements • Description of the FONT interaction-point feedback system • FONT 5 installation at the ATF, KEK Japan • Winter 2009/10 beam test results • Plans for spring 2010 Ben Constance Io. P HEPP/APP annual meeting 2010 1
The International Linear Collider and its beam structure ~31 km • The ILC beam design is as follows: • Pulsed operation at 5 Hz, each pulse contains 2625 bunches • Bunches separated by 370 ns and contain 2 x 1010 particles • Structure driven by factors such as: • Klystron RF power sources and damping rings • Superconducting accelerating cavities • The long bunch spacing can be exploited for digital feedback Ben Constance Io. P HEPP/APP annual meeting 2010 2
The ILC luminosity goal • Luminosity target of 1034 cm-2 s-1 (same as the LHC target) • Single collision opportunity, therefore very small beams at IP • Just 5. 7 nm by 639 nm at the interaction point • Need synchrotron cooling with damping rings • We can use very strong focussing not possible in a ring • Focus the beam so tightly at the IP that it is unstable afterwards • The positions of the beams must be very stable • Careful machine and beam alignment plus feedback systems Ben Constance Io. P HEPP/APP annual meeting 2010 3
Fast IP feedback systems • Many effects can misalign the few nm beams, e. g. • Ground motion • Temperature fluctuations (cooling water, electronics) • Mechanical vibration • Bunch-by-bunch feedback • Determine the offset error of the first bunch after the IP • Calculate a correction to the next bunch’s position • Apply nm level correction before the next bunch reaches IP (< 370 ns) • Repeat for entire train • IPFB makes use of beam-beam deflection after the IP • As the beams pass each other, they each receive a large kick • This kick is determined by the beams’ relative offset • Trivial to measure >100μm deflection after IP Ben Constance Io. P HEPP/APP annual meeting 2010 4
Beam-beam kick simulation Ben Constance Io. P HEPP/APP annual meeting 2010 5
The FONT project at the JAI • Feedback on Nanosecond Timescales • At JAI we are working on a digital bunch-by-bunch IPFB system for ILC • Measure deflection of one beam and correct the other • Delay loop ensures we retain and improve previous correction Ben Constance Io. P HEPP/APP annual meeting 2010 6
ATF 2 • FONT is currently being tested at ATF 2 • The Accelerator Test facility in KEK, Japan • Designed to demonstrate ILC-like beam sizes • Single 1. 3 Ge. V electron beam, 150 ns bunch spacing • Single linac, damping ring, extraction line • New final focus system Ben Constance Io. P HEPP/APP annual meeting 2010 7
FONT 5 at the ATF 2 extraction line FONT 5 installation region Final doublet and beam dump Ben Constance Io. P HEPP/APP annual meeting 2010 8
ATF 2 design goals • Demonstrate a 35 nm beam size • New final focus installed • Beam characterisation and tuning very much in progress • Demonstrate nanometre–level stability • This needs micron stability at the entrance to final focus • FONT 5 system has been commissioned to provide this • We have micron-resolution processors • Our kickers can provide micron-level corrections at ~1 Ge. V • Resolution of FB system defined by smallest correction possible • A kick that provides 1μm correction at 1 Ge. V provides 1 nm correction at 1 Te. V • FONT goals at ATF and ILC are compatible Ben Constance Io. P HEPP/APP annual meeting 2010 9
FONT 5 ATF 2 beamline installation • Three stripline BPMs made to existing ATF ext. line specification • Two fast fixed-gap stripline kickers recovered from NLCTA (vertical orientation) • Next-generation FONT 5 digital feedback board installed and operational • Flexible configuration allows a variety of feedback algorithms P 1 P 2 P 3 To dump Control / feedback loops K 1 QD 10 X QF 11 X K 2 QD 12 X QF 13 X QD 14 X QF 15 X FB board DAQ Ben Constance Io. P HEPP/APP annual meeting 2010 10
FONT 5 digital feedback board • Single feedback board controls multiple loops and all DAQ • Based around a powerful Xilinx Virtex 5 FPGA • 9 ADC input channels at 357 MSPS (TI ADS 5474) • 4 DAC output channels (AD 9744) • Basic algorithm includes: • Removal of beam intensity fluctuations • Correction of extracted train shape • Compensation for kicker pulse droop • All signal mixing, DSP, amplification plus rise of kicker takes place in < 130 ns! Ben Constance Io. P HEPP/APP annual meeting 2010 11
ATF pulse structure and feedback algorithm • We extract a three-bunch train • Repetition rate 2/3 Hz • Spacing variable from 140 – 154 ns • Bunch charge from 1 x 109 up to 5 x 109 e • FONT 5 intra-train feedback • Aim is to reduce pulse-to-pulse jitter, i. e. correct jitter that is correlated between bunches • First bunch is a pilot bunch and is not corrected • 3 FONT BPM signals used to calculate intensity-independent position readings • Position values used to calculate 1 or 2 feedback signals and apply to kickers • Feedback signal applied to the 2 nd bunch • The corrected 2 nd bunch is then measured to produce a feedback for the 3 rd bunch, and so on Ben Constance Io. P HEPP/APP annual meeting 2010 12
Results from winter 2009/10 beam tests P 2 – to – K 1 feedback loop closed analysed • Position feedback performed (see simulations to the right) Ben Constance Io. P HEPP/APP annual meeting 2010 13
Beam test results (2) • Upstream corrector dipole magnet used to introduce a variety of beam offsets Ben Constance Io. P HEPP/APP annual meeting 2010 14
Latest beam test results (3) – jitter reduction • We allowed the feedback to operate on a nice beam 17/02/10 Ben Constance Io. P HEPP/APP annual meeting 2010 15
Latest beam test results (4) – jitter reduction • Repeated this experiment 18/02/10 for 25 minutes Ben Constance Io. P HEPP/APP annual meeting 2010 16
Plans for spring 2010 • Two-loop angle feedback • Intend to close K 1 and K 2 FB loops simultaneously • Interaction between the two kickers means we must couple the loops • Coupling parameters easily determined by calibration Ben Constance Io. P HEPP/APP annual meeting 2010 17
Summary • Any future linear collider will require nm level stability at the IP • IPFB will be required • The FONT 5 feedback system is installed and operational at ATF 2 • Single-loop position feedback demonstrated • We will move on to two-loop angle feedback demonstration in the spring Ben Constance Io. P HEPP/APP annual meeting 2010 18
- Slides: 19