Minutes of 08 Nov 2006 SWMD Telephone Meeting

Minutes of 25-Oct-2006 meeting (note correction to alpha1 in collim. 13 of Design considerations, from GE/JDAS).

Brief reports on:

1. Review timescales on critical path - any schedule changes?
   • 13 Nov for "physicist" designs for ESA run 3.
     
   • 13-17 Nov. for drawing up these designs and checking, ready to sign off at RAL and send to outside manufacturing company.
     
   Dates for first 2007 run now likely to be delayed by ~6 weeks, from 7-26 Mar.. Agreed to make use of very likely delay, allow ~1 week extra for more simulation studies in GdfidL, which already look promising - small amount extra time may improve understanding significantly.

2. T-480 Data analysis et al.
   • Survey results from sandwich 1 - Steve
     Slopes fitted to profiles are in good agreement with design, c. 324mrad. Offset in z for slot 1 sandwich 1 of approx. 680microns - discussed whether some margin for longitudinal misplacement of collimators (small screws, large holes?), differing in upper and lower sandwiches. Justin and George checked drawings for collimators, some inconsistencies noted, see JG email. Note that from sandwich 1 slot 3 measurements - more precise because of easier access with collimator tapers being close to end of sandwich - no indication that the upper and lower collimators are not parallel to one another, something we had discussed during summer when trying to understand data.
   • Combined results for each collimator - Luis talk
     • Statistical consistency of results? - talk - Nigel
     • Bunch length variation
     • Limiting effects on kick factor resolution?
     
     Luis has combined results from most runs now per collimator, presented kick factor measurements (kick, scaled by charge including variation within a run), with reference orbit taken from same slot as scan run and/or slot 0 (if available). Results fairly consistent in both cases. Relative statistical uncertainties vary from few per cent to 80 %. Errors from individual runs do need to be scaled by sqrt(chisq), to account for fact that simple 3 paramater fit to observed deflection/charge data does not describe all runs well. Other factors changing in experimental conditions during the (long) runs may be the underlying cause of this, e.g. bunch length changes. Need to look for correlation with other recorded information e.g. in MCCdata, energy BPMs. NKW suggested we have to discard runs which have crazy chisq. values, even when using chisq. scaling of fit parameter errors, e.g. cut on prob(chisq)>1.e-4.

     With the combined results so far, noted that difference (collim. 3 - collim. 1) is consistent with analytic expectation of ~2V/pC/mm from analytic calculation of resistive wake. Collim. 5 gives kick consistent with analytic calculation, collims. 7, 8 give larger kicks than collim. 6, as expected. Collims. 1, 2 give comparable kicks, both smaller than collims. 7-8, contrary to expectation. Note in all observations above, statistical significance of statements is low given currently estimated parameter errors.

3. Wire Tests - update

4. Wakefield Simulations
   • Numerical Simulations of Collimator Insertions using MAFIA, EUROTeV-Report-2006-103, C.D.Beard, R.M.Jones
   • Run 3 Collimators - Jonny
   • Effects of mechanical imperfections
     Effect of increased gaps as noted by survey estimated by Jonny's GdfidL simulations to be ~10-20% reduction in kicks, varying with mesh size. Effect of misalignment in z for collim. 5 found to be approx. linear, roughly 6% increase in kick per mm offset in z (over the region 0-1.5mm studied). Future check would be to test impact of moving beam position in opposite direction in y, as the geometry is no longer symmetrical.
   • (Possibly) rough estimates for run 3 designs?
   • Jonny alt. figure
     Has made predictions with 500um bunch and 12 cells/sigma, suggest that collim. 15 would have geometric kick comparable to collims. 6-8, while collim. 13 would be some 20% larger.

5. Merlin Simulations
   • Higher order mode wakefields in the ILC-BDS collimators - Adriana
   
   Adriana tracked particles from end of linac to IR with optics deck from Frank, examined vertical/horizontal emmittance and divergance at IP with/without wakefields, for two different transverse offsets at entrance to BDS (0.5, 1.5mm). Will repeat with different bunch length, see contribution to emmittance from each collimator and look at implementation of rectangular collimators.

6. Collimator designs and manufacturing
   • Decision on candidate designs for run 3
     1. Options so far: from GE, from NKW
     2. candidate #1 and justification
     3. candidate #2 and justification
     4. ...
     
     Discussed candidate designs. Given experience so far with T480 analysis, does not look likely that effects below 10% on typical kick of few V/pC/mm would be detectable. Working proposal is collimators A-C from George's slides (includes collim. 8, already tested in 2006); collimators 13-15 from NKW slides; "semi-circular collimator" and some curved section, t.b.d.. Surface finish effects would be implicit in comparison of existing collim. 8 with GE's B/C; Consider also possibility of making intentionally roughened surface on one collimator, to R_a ~10um. This would give surface finish ~0.1sigma_y at ESA, which would be comparable to surface finished achieved using two pass EDM at ILC where sigma_y ~10um.

7. Next meeting - to be decidedby email discussions, depends on progress with simulations/analysis.