1. @ Fermilab ILC bunch-compressor and linac rf requirements Sergei Nagaitsev Fermilab Feb. 9, 2006

2. @ Fermilab ILC@FNAL - Nagaitsev 2 The Baseline Machine (500GeV) Nick Walker – GDE meeting Frascati 8.12.05

3. @ Fermilab ILC@FNAL - Nagaitsev 3 Gradient (yet to be demonstrated) Cavity type Qualified gradient Operational gradient Length*energy MV/m KmGeV initialTESLA3531.510.6250 upgradeLL4036.0+9.3500 * assuming 75% fill factor Total length of one 250 GeV linac  10km Need for each linac: 7872 Nb cavities 984 cryomodules 328 Klystrons …many km’s of waveguides

4. @ Fermilab ILC@FNAL - Nagaitsev 4 About the ILC Bunch Compressor (from PT)  Two-stage bunch compressor  Stage 1 @ 5 GeV (DR extraction energy) – 2 Klystrons (?)  Stage 2 @ ~15 GeV -- 20 Klystrons(?)  Necessary to get large compression factor with large longitudinal emittance from DR  Two final bunch lengths to consider  1 psec RMS (nominal)  0.5 psec RMS (LowQ / HighLum)  Two configurations for each final length  “A” configurations Total 180° longitudinal phase rotation (90° per stage) Easier transverse tolerances (lower energy spreads) at expense of tighter RF and longitudinal tolerances  “B” configurations Undercompress in stage 1 for 90° total rotation Easier longitudinal tolerances at expense of tighter transverse tolerances  “B” configurations are baseline, but being able to achieve “A” might be useful as well

5. @ Fermilab ILC@FNAL - Nagaitsev 5 Luminosity vs. IP Offset for Nominal ILC Parameters (from M. Church) calculated with Guinea-Pig

6. @ Fermilab ILC@FNAL - Nagaitsev 6 2% integrated luminosity reduction; BC300B

7. @ Fermilab ILC@FNAL - Nagaitsev 7 Luminosity distribution for 30000 bunches

8. @ Fermilab ILC@FNAL - Nagaitsev 8 PT Slide from Snowmass Conference 0.25 ?? 0.12 ??

9. @ Fermilab ILC@FNAL - Nagaitsev 9 Time scales needed for rf specs 1.Bunch-to-bunch (100s of ns) 2.Train-to-train (few Hz) 3.Long term (Seconds to days) Bunch patterns: 1.One pilot (1% intensity) bunch alone – should be able to get all the way through 2.One pilot (1%) bunch followed by a single (nominal) bunch (10 us) apart

10. @ Fermilab ILC@FNAL - Nagaitsev 10 RF specs  Bunch compressor (BC) may require a separate rf spec (nominal 300B: 0.25 deg rms, 0.15% (??) rms).  Main Linac (all energy jitter):  Each bunch has an rms energy spread of 0.1%  Colliding bunches have luminous region rms energy spread of 0.39 GeV (0.15%)  Bunch-to-bunch energy jitter adds in quadrature  Factor of 50 due to SQRT(N_klystrons)  >1 deg, 1% (rms) uncorrelated rf phase/amplitude  0.5 deg, 0.1% correlated  Still need to look at extracted bunch (from DR) energy jitter.

11. @ Fermilab ILC@FNAL - Nagaitsev 11 Groups of specs  20% energy error can get the beam thru the Linac  2% energy error to get to the IP  Two groups of specs: inside-the-feedback-loop, outside-of-the-loop  Outside: MO, Timing distr, Reconstruction -- SNS attained:.1 deg btw adjacent cavities, Tesla prototype: 0.5 degree rms over 15 km over long time scale  Inside: LLRF, HLRF and rf distribution  Assumes we can measure and sum up 24 cavity probes without errors  Some cables (from cavities to LLRF crate) will be 100 m long

12. @ Fermilab ILC@FNAL - Nagaitsev 12 What About Beam-Based Feedback?  This slide from Marc Ross  Energy:  Certainly will have train-to-train feedback Takes care of variation on time scale of seconds or longer  Intra train feedback dubious Takes ~10 usec to change cavity voltage ~1% Takes ~50 usec for signal to reach front of linac from IP May not be possible to make an intra-train energy feedback which has speed and range required  Arrival time:  Will have train-to-train feedback if we can make reliable IP instrument to measure arrival time difference Otherwise, we’ll rely in a dither feedback In either case, need seconds to many minutes to detect and correct arrival time problems  Intra-train has similar issues to energy