Presentation on theme: "1 Collimation Simulations and Optimisation Frank Jackson ASTeC, Daresbury Laboratory."— Presentation transcript:
1 Collimation Simulations and Optimisation Frank Jackson ASTeC, Daresbury Laboratory
2 Contents Collimation design (like BDS) lead by SLAC/FERMILAB Areas of input – collimation depths and collimation optimisation. Collimation Depths For recent and current BDS designs. BDSIM cross check (off energy calculation) Collimation Optimisation and Halo Tracking Method and History Latest results 2006e
3 Collimation Depths Overview Criteria, clean passage of halo SR through IR apertures Beamcal and extraction quads are limiting apertures Many complicating factors Multiple crossing angle 20,14,2,0 mrad Multiple detector designs, L* IP Parameter Sets (Nominal only here, as in BDS design) LDC small xing angle
4 BCD (2005) Collimation Depths Used original NAPOLY DBLT SR fan tracing code Linear, on-energy optics, head-on Play tricks for crossing angle situations. 20 and 2 mrad was baseline, similar depths
5 New BDS for RDR 14 mrad approved after Vancouver, now likely to become single IR Identical FD to 20 mrad but different IR design 20 mrad 14 mrad 2006”c”
6 BDSIM cross check for 2006c DBLT is linear on-energy envelope tracking BDSIM can track off-energy halo through FD SR profile at 1 st Extraction Quad (r=18mm) p=1%, Gaussian On-energy, p=1% Gaussian
7 Alternative BDS Head-on re-emerged as small-angle alternative
8 Other Parameter Sets? Rough estimation possible Smaller IP beam size larger IP angles Wider SR fan tighter collimation Example: high lumi parameters Beamsize ~ 2 smaller than nominal Collimation ~ 2 tighter
9 Collimation depths ideas for further work/collaborations The real RDR deck, 2006”e” incorporates reoptimisation of extraction quad apertures, need to check collimation depth All collimation depths assume ideal criteria clean SR passage Tolerances never really studied BDSIM/detector interface for effects of non-clean SR passage.
10 Collimation Optimisation Some history from pre-ILC days NLC and TESLA collimation efficiency compared in TRC report. NLC rather good collimation efficiency ILC evolved from NLC Survivable betatron spoiler optics Collimation performance poorer. Tighter apertures in y. SPEX as secondary betatron spoiler. Can we tweak ILC lattice to restore NLC performance?
11 More History and Basic Design ILC BDS gone through multiple revisions since 2005 But none have specifically addressed collimation optimisation Latest 2006e deck is for minimised length/cost PHASE ADVANCES not perfect for any design BANDWIDTH through final doublet not well controlled for any design
12 2006e Collimation Features 2 spoilers /2 apart, energy spoiler Phase advances not exact Does not deliver nominal performance, even on- energy Mitigate by aperture tightening (SPEX as betatron spoiler!) SP2 SP4 SPEX x = 0.38 y =0.59 x = 2.76 y =2.34 1.00 2
13 2006e Bandwidth Bandwidth figure of merit: Twiss vs. E at IP 2006e NLC
14 Optimisation Strategy Work done in 2006 Advice and tools from A. Seryi and M. Woodley Restore phase advances between SP4,SPEX,IP Optimise Bandwidth MAD fitting driven by optimisation routine
15 EPAC 2006 Results Starting with ‘FF9’ BDS Used explicit matching section 7 quads and 1 drift length Constrain Twiss at SP4 and phase advances SP4-IP No constraint on bandwidth Multiple solutions obtained, best bandwidth chosen x,y =n /2
16 EPAC 2006 Results Collimation performance checked by halo tracking in MERLIN ‘Black spoilers’ no secondaries Some improvement clear from phase restoration and bandwidth optimisation Design performance only delivered by spoiler tightening Halo ( p=1%) Tracking to FD entrance Original FF9 Performance New Performance
17 2006e Latest Work SPEX optimisation Use 4 quads + 2 drift lengths at end of betatron section match phase SP4/SPEX Then repeat IP-SP4 matching as before, optimising bandwidth Able to achieve matched solution with promising bandwidth
18 2006e Optimised Performance Tracking Results Several simulation parameters here different to EPAC06 Halo distribution and population Improved performance, suggest no longer need vertical SPEX collimator Same population in both halos at FD Original 2006e Performance New Performance
21 Conclusions Optimised lattice performance seems much improved for primary halo tracking Full simulation (secondaries) and losses along line needed BDSIM/STRUCT Lattice longer/costlier so ultimately this optimisation may not be feasible. Additional quads rather than longer length?