Jonathan Smith, COLSIM, 4 th December, CERN 1 GdfidL Simulations of ESA Prototype Collimators Jonathan Smith (Lancaster University/Cockcroft Institute)

Jonathan Smith, COLSIM, 4 th December, CERN 2  =  /2rad r=4.0mm 4, 4  324mrad r=1.4mm 3, 3  324mrad r=1.4mm 2, 2  =324mrad r=2.0mm 1, 1 Revised 4-May-2006 Beam viewSide view (“DESY sandwich”) Collim. #, slot h=38 mm 38 mm L=1000 mm 7mm  r=1/2 gap As per last set in Sector 2, commissioning Extend last set, smaller r, resistive WF in Cu cf. same r, tapered

Jonathan Smith, COLSIM, 4 th December, CERN 3  =  /2rad r=1.4mm 5, 4  166mrad r=1.4mm 6, 3  1 =  /2 rad  2 =166mrad r 1 =4.0mm r 2 =1.4mm 7, 2 r 1 =4.0mm r 2 =1.4mm   =289mrad   =166mrad 8, 1 Revised 4-May-2006 Beam viewSide view (“SLAC sandwich”) Collim.#, slot h=38 mm 38 mm 7 mm cf. collim. 4 smaller r 211mm cf. collim. 2, same r 31mm cf. collims. 4 and 6 133mm cf. collim. 7, and same step in/out earlier data

Jonathan Smith, COLSIM, 4 th December, CERN 4 EM Simulations with GdfidL

Jonathan Smith, COLSIM, 4 th December, CERN 5 Analytical estimate ● Stupakov says: ● Stupakov asserts that the h»b is met, but is this really valid?

Jonathan Smith, COLSIM, 4 th December, CERN 6 For SLAC collimators...

Jonathan Smith, COLSIM, 4 th December, CERN 7 (6 cells/sigma)

Jonathan Smith, COLSIM, 4 th December, CERN 8 (12 cells/sigma)

Jonathan Smith, COLSIM, 4 th December, CERN 9 Performance of these and others at 1mm bunch

Jonathan Smith, COLSIM, 4 th December, CERN 10 Mesh stability: Collimators 1&2 ● Only 3 decent points at 300µm for most collimators ● More at 500µm ● 1mm ~ OK – can use spline fit on data to get an estimate – not done so far – further analysis to see if this takes us closer to ECHO/PBCI.

Jonathan Smith, COLSIM, 4 th December, CERN 11 What do we do about it? (1)

Jonathan Smith, COLSIM, 4 th December, CERN 12 What do we do about it? (2)

Jonathan Smith, COLSIM, 4 th December, CERN 13 Further collimator designs semi-circle, with[9]/without[10] flat, opposing demi-circles[10], 8 with flat[11].

Jonathan Smith, COLSIM, 4 th December, CERN 14 More possible collimators 7 with flat [13], half exponential[14], 13 with shallower angle[15] exponential profile[16], 13 with ellipse connecting 4mm and 1.4mm aperture[17] 13 with ellipse connecting beam pipe radius and 1.4mm aperture (also see 9) [18] half cosine taper [19], raised cosine taper [20], tanh tapers [21] (set typically to the length of collimator 6)

Jonathan Smith, COLSIM, 4 th December, CERN 15  166mrad r=1.4mm 12  =166mrad r =1.4mm 11  =166mrad r =1.4mm 10  166mrad r=1.4mm (1/2 gap) 6 Revised 27-Nov-2006 Beam viewSide view Collim.# h=38 mm 38 mm Roughened surface, compare with 12 As 10, in Ti-6Al-4V, polished, cf. 12 As 10, in OFE Cu, polished, cf. collim. 6, 13 Runs 3, 2007 Exists, from 2006 runs. For reproducibility ~211mm 1.4mm  =21mm   

Jonathan Smith, COLSIM, 4 th December, CERN 16 non-linear taper r=1.4mm 16?  1 =  /2 rad  2 =50mrad r 1 =4.0mm r 2 =1.4mm 15  1 =  /2 rad  2 =166mrad r 1 =4.0mm r 2 =1.4mm 14  1 =  /2 rad  2 =166mrad r 1 =4.0mm r 2 =1.4mm 13 Revised 27-Nov-2006 Beam viewSide view Collim.# h=38 mm 38 mm cf. ? Polished, cf. collim. 13 Polished, cf. collims. 7, 11, 13 Polished, cf. collim. 7, 12, 13 Ti6Al4V OFE Cu =21 mm Form t.b.d. = 0.6    Ti6Al4V Runs 3, 2007 ~52 mm =21 mm  ~52 mm =21 mm   ~125 mm =21 mm

Jonathan Smith, COLSIM, 4 th December, CERN 17 Variation of collimator kick with bunch length

Jonathan Smith, COLSIM, 4 th December, CERN 18 Big gap?

Jonathan Smith, COLSIM, 4 th December, CERN 19 Dependence on Gap

Jonathan Smith, COLSIM, 4 th December, CERN 20 Misallignment

Jonathan Smith, COLSIM, 4 th December, CERN 21 GdfidL & PBCI ● TEMF working on ECHO/PBCI – 3D, moving mesh, conformal, non-dispersive solver s/σ σ/Δz=6 W ║ (s)/(V/pC) GdfidL PBCI With thanks to Mikko Kärkkäinen

Jonathan Smith, COLSIM, 4 th December, CERN 22 Conclusion EM simulations: MAFIA GdfidL ECHO Bench Tests (scaling?) Analytical Approach (Lancaster Theory Group) Beam Tests ESA Damage Studies ? Prediction of effect at IP (MERLIN) New designs Prediction of mechanical properties Now starting second iteration…

Jonathan Smith, COLSIM, 4 th December, CERN 23 Discussion. ● Is parameterization from GdfidL, MAFIA, ECHO, PBCI,ABCI in order to improve the models used in PLACET/BDSIM/MERLIN was something worth pursuing?