Sunday morning: Beam Dynamics Issues Session highlights and outstanding questions M. Pivi and D. Rubin ECLOUD10 October 8-12 Cornell University
Summary -Electron cloud induced instabilities observed for long (ns) bunches at SPS and short (33 ps) bunches at CesrTA -The same instabilities appear (more or less in simulations) -Investigation of feedback as mitigation for long bunches suggests that it is a possible but challenging solution requiring high bandwidth and a strong kicker -Instabilities observed at CesrTA are characterized as head-tail with signature synchrotron sidebands of the vertical tune -Threshold inferred from tune shift to be about 10 12 electrons/m 3 in reasonable agreement with simulation - Vertical emittance is observed to increase with amplitude of sidebands Simulation indicates that feedback is marginally effective for positron bunches and that the bunch to bunch correlation length is short (a few bunch lengths)
Claudio Rivetta Feedback system for e-cloud/tmci
Claudio Rivetta Feedback system for e-cloud/tmci
J.L. Vay: Comparison with experimental measurements Experiment Bunch 119, Turn 100-200 Warp-Posinst Bunch 24, Turn 0-100 Fractional tune Fractional tune Nominal fractional tune=0.185 head tail head tail Z slice Separation in two parts, core and tail, with tune shift in tail higher than in core Similar tune shift for the core of the beam but higher with simulations for tail Effect appears earlier and is stronger in simulations, suggesting that the electron density is higher than in experiment
J.L. Vay: Tests with 5 pts FIR filter Experiment will use FIR filter (see talks from C. Rivetta and R. Secondo)
Raffaello Secondo: Feedback System Performance - III Limiting the Electric Field in the Kicker reduces the possibility of stabilizing the bunch via feedback system. Max Electric Field of the Kicker limited at 300 kv/m Single Slice vertical displacement in the TAIL Single Slice vertical displacement in the HEAD Vertical po osition [μm] Vertical po osition [μm] turns turns Normalized power [a.u.] Fractional Tune Normalized power [a.u.] Fractional Tune ECLOUD10 Workshop, Cornell University, Ithaca, USA, October 8-12, 2010 E-cloud SPS feedback simulations R. Secondoet al. 15
G. Dugan: Bunch-by-bunch power spectrum Run 166 49th ICFA Advanced Beam Dynamics Workshop (H,V) chrom = (1.33,1.155) Avg current/bunch 0.74 ma.
Run 166 G. Dugan:Detailed features of horizontal and vertical lines 49th ICFA Advanced Beam Dynamics Workshop Run 126 Lower frequency (~3 khz) shoulder in the horizontal tune spectrum is attributable to known dependence of horizontal tune on the multibunch mode. Bifurcation of the vertical tune spectrum (peak at ~ 1.5 khz higher frequency), which starts to develop at the same bunch number as the head-tail lines, is not understood. September 24, 2010 ECLOUD`10 - Cornell University 9
Vertical HT m=-1 G. Dugan: Precursor bunch dependence 49th ICFA Advanced Beam Dynamics Workshop Spectrum of bunch 1 in 30-bunch train 0.75 ma/bunch Vertical m=0 No precursor 0.75 ma precursor bunch: 182 ns before bunch 1 (1960 ns after bunch 30) September 24, 2010 ECLOUD`10 - Cornell University 10
J. Flannagan: xbsm bunch by-bunch at CesrTA
J. Flannagan: xbsm bunch by-bunch at CesrTA
J. Flannagan: xbsm bunch by-bunch at CesrTA
Ohmi-san Simulated Unstable spectra Lower sideband is dominant for high ωeσz/c (low emittance). 2 GeV Upper sideband is dominant for 5GeV 5 GeV
Ohmi-san Slow growth lower than the 2GeV threshold ρth=1.2x10 1 2 5GeV Slower than radiation damping time ρth=5x10 12 15
Session highlights and outstanding questions Claudio Rivetta - Measured the instability of ns length bunches in SPS - Observe tune shifting along the length of the bunch - Measurements of bunch spectrum suggest that 4Gsamples/sec is required to damp instability - Installation of Feedback in SPS in 2013 - --------------------------- Slides #: 4, Have you excluded beam loading in rf cavities? (yes) Timescale for kicker installation in SPS is 2 years which is short! Need more manpower help. What kind of kicker you expect to use? Not yet defined.
J-L.Vay: Session highlights and outstanding questions Numerical modeling of instability and feedback - Based on WARP (benchmarked against POSINST) - Simulation of 72 bunch train at 26GeV (SPS) with 10 interactions stations/turn - Instability appears at bunch 19 along with emittance growth and tune shift - Simulated bunch spectrum roughly agrees with measurement Effective feedback will require > 450MHz 10 kicks/turn is enough to resolve properly the instability? Slide 23: The bunch is 1Ghz, 0.5 Ghz is too low. Slide 27: cloud density too low: start adjusting the knobs to fit data Change filter characteristics with bunch number?
Session highlights and outstanding questions R. Secondo Simulation of feedback for instability observed in SPS - If kicker strength is limited (realistic), effectiveness of feedback is compromised. Slide 15 : how is the kicker field limit compared to the kicker noise? Slide 15: first the oscillation is damped and then the instability is excited again? This is to be understood better. It is either a numerical problem or problem with the feedback system? Development is for a dipole kicker, How about a quadrupole kicker?
Session highlights and outstanding questions Gerry Dugan: CesrTA Ecloud induced instabilities - Measure spectrum of individual bunches in a long train with gated spectrum analyzer including - Betatron tune - Amplitude of f v ± f s - Observe vertical m=±1 head tail lines - Typically synchro-betatron lines appear at bunch ~15 with corresponding increase in vertical emittance - Instability threshold inferred from tune shift at 8-9 X10 11 /m 3 near the beam - Threshold sensitive to chromaticity, beam current, number of bunches - Precursor bunch eliminates instability of the first bunch in the train
Session highlights and outstanding questions Are the tune line splittings similar to the PETRA III splitting? Can the splitting of the horizontal tune line due to the multi-bunch? And vertical splitting due to single-bunch. Observed head-tail line for 1 st bunch in CesrTA train! probably due to cloud surviving multiple turns in quadrupole and 0-field wiggler regions. Enough cloud left over to cause head-tail! Cured by precursor? Slide 26: the vertical tune doesn t move with increasing beam current. Why? Suggestion to try to characterize e-cloud with machine impedance, using 1 high current bunch.
Session highlights and outstanding questions John Flanagan: Measurement of size of individual bunches in the train indicates emittance growth in the same bunches that exhibit head tail instability - Dipole instability that appears with transverse feedback off makes it difficult to resolve coded aperture images we should avoid making e-cloud blow-up measurements with the (at least transverse) feedbacks off. Make measurements with FB ON (xbsm) and OFF (BPM). Or at least keep transverse FB on all the time? Fitting procedure (to be improved?) Suggestion: Fit BPM data with the coded aperture data. This is at the limit of the resolution. Can simulate the dependence of the fitting with the bunch displacement. Low photon statistics can limit the comparison
Session highlights and outstanding questions Ohmi-san: Agreement between simulations and experimental data and analytical model about instability threshold : although, appearance of 1 lower sideband instead of 2... Head tail instability threshold characterized by ω e σ z /c Tune sideband does correspond to synchrotron line? Check Chromaticity can cure linear growth below threshold?