PowerPoint Presentation

Simulations and measurements of the injection oscillations and the MTEA. HuschauerSupervisor: S. Gilardoni

Acknowledgements: H. Bartosik, H. Damerau, D. Gamba, M. Giovannozzi, C. Hernalsteens, M. Kowalska, K. Li, E. Mtral, N. Mounet, G. Rumolo, G. Sterbini, R. Wasef, PS/PSB operations team LIU-PS Student Day, April 21, 2015OverviewPart I Injection oscillations

MotivationMeasurement resultsSimulation studies with HEADTAILOutlook

Part II The Multi-Turn Extraction (MTE)

From Continuous Transfer (CT) to MTEThe trapping process-bucketsSimulations with PTC and PyORBITCommissioningOutlook#A. HUSCHAUER, April 21, 2015Injection Oscillations - Motivationslow losses observed for a few 100 s after injection

losses in the injection region constitute one of the limitations for high intensity beams access very difficult in case of septum breakdown

aperture in the injection region#A. HUSCHAUER, April 21, 2015Vertical intra-bunch oscillationswide-band pick-up used to observe the beam signal proportional to displacement and line density

in a linear machine, with a tune set to q = 0.25, one would expect the following signal

#A. HUSCHAUER, April 21, 2015Vertical intra-bunch oscillationswide-band pick-up used to observe the beam signal proportional to displacement and line density

in a linear machine, with a tune set to q = 0.25, one would expect the following signal

however, for high-intensity beams a fast build up of intra-bunch oscillations (few turns after injection) is observed

nTOFCNGSIp = 375 1010 ppbIp = 700 1010 ppb#A. HUSCHAUER, April 21, 2015Simulation studies with HEADTAILNeed to study collective effects with simulation tools

ImpedanceWake2D is used to obtain wall impedance and wake functions for a given vacuum chamber

for the case of the PS the wake computations are based on a continuous stainless steel chamber with the shown dimensions (about 70% of the PS is equipped with this type)

In HEADTAIL bunches are longitudinally sliced and macro particles receive a kick depending on wake field

#A. HUSCHAUER, April 21, 2015Simulation studies with HEADTAILsimulationmeasurement

very good agreement between measurements and simulations obtained

wake field based on simple elliptic geometry is sufficient to reproduce the measurements

reduce this oscillations by reducing the injection error or using the transverse feedback#A. HUSCHAUER, April 21, 2015Outlook Part IRedo simulations with newly available code PyHEADTAIL (bug in the wake interpolation fixed)

Experimental verification of simulation results:

Dependence of the oscillation on the steering in the BTP-lineDependence on chromaticityVerify linear relationship between intensity and tune shiftDependence on energy

PyHEADTAIL also offers the possibility to include direct space charge

Eventually try to improve the analytic model#A. HUSCHAUER, April 21, 2015From CT to MTE90 rotation per turnelectrostatic septum

Horizontal phase spacecontinuous (de-bunched) beamtune set to qx = 0.25local bumps to approach electrostatic and magnetic septa2 ejections over 5 turns to fill 10/11 of the SPS

#A. HUSCHAUER, April 21, 2015From CT to MTE

beam loss as seen from the control roommagnetic fieldbeam intensitysignal from the beam loss monitors (BLMs), which are located all around the machinelocation of the electrostatic septum#A. HUSCHAUER, April 21, 2015From CT to MTEPrinciplebeam is split in the horizontal phase space using non-linear elements (sextupoles and octupoles) stable islands

tune is slowly varied to cross a fourth order resonance

Advantages:no septum needed to slice the beam significant reduction of losses (approx. 2%)same parameters for all four beamletssame optical parameters (in contrast to CT)

installation in the PS#A. HUSCHAUER, April 21, 2015The trapping process

qx = 0.20

qx = 0.25

qx = 0.2505

qx = 0.254

program non-linear elements to appropriate values to excite resonance

ramp horizontal tune across the resonant value

decrease current in the elements while increasing the tune

extract the beam once islands are sufficiently separated

#A. HUSCHAUER, April 21, 2015Presence of -buckets in the longitudinal plane?

During an MD in 2008 a distortion of the longitudinal profiles of the beamlets was observed when sending beam to the SPS.

caused by -buckets?

Synchrotron HamiltonianT. Bohl, BE-Note-2009-015_MDnon-linear slip factor (important close to transition)#A. HUSCHAUER, April 21, 2015Presence of -buckets in the longitudinal plane?

#A. HUSCHAUER, April 21, 2015Presence of -buckets in the longitudinal plane?

Further increase of 1 leads to merging of both layers and eventually to -buckets Highly non-linear magnetic fields required for MTEPlateau at 14 GeV/c is too far from transition to allow an important effect of 1It would be interesting to go closer to transition and artificially create -buckets#A. HUSCHAUER, April 21, 2015Simulations with PTC in 6D

Longitudinal motion modulates horizontal phase space Observation of trapping and de-trapping phenomenaImportant to study the whole trapping process in 6Dsmall dp/plarge dp/p#A. HUSCHAUER, April 21, 2015Simulations with PyORBITSuccessor of ORBITUsed at CERN in combination with PTC for space charge studiesAllows to vary the strength of magnetic elements with time perfectly suitable for MTE

#A. HUSCHAUER, April 21, 2015Simulations with PyORBIT

Horizontal distribution 14.1%Significant differences observed between horizontal, transverse and 6D distributionPercentage of particles captured in the islands (ideally 20%) gradually decreasedTransverse damper is operationally deployed to increase trapping efficiencyTransverse distribution 13.3%6D distribution 8.5%#A. HUSCHAUER, April 21, 2015MTE commissioningSome important aspectsOptimize the settings of the non-linear elements during the capture and at extractionOptimize the settings of the transverse damperSet up the extraction for MTE and other operational beamsCorrect trajectories during the rise of the slow extraction bump

However, a problem present since 2006 retards the progress of the commissioning Periodic modulation of the captureWire scanner measurement#A. HUSCHAUER, April 21, 2015Capture oscillations

Capture efficiency significantly modulated from cycle to cycleWaterfall representation of the WS measurements reveals an oscillatory behaviorPeriod of this oscillations observed to change regularlyVast amount of parameters investigated to track down the source

#A. HUSCHAUER, April 21, 2015Capture oscillationsWe are confident to be on the right track:PFW used to set chromaticities and working point during acceleration and at 14 GeV/cFFT on the measured current reveals several important frequencies with the main line @ 5 kHz switching frequency of the power converterNot a problem per se, but: variation in time of the amplitude of this 5 kHz component

#A. HUSCHAUER, April 21, 2015Capture oscillationsOscillation frequency measured with the WS is correlated to the frequency of the amplitude variation of the component @ 5 kHz.

Investigation with colleagues from the power converters and magnets ongoing.

#A. HUSCHAUER, April 21, 2015But why does this affect the capture?The 5 kHz component of the power converter can also be measured on the magnetic field and the tune of certain beams.The (secondary) tune around the stable fixed point inside the islands is much lower than the linear tune of the machine. especially true for particles close to the SFPCurrently we assume a resonant population/depopulation of the islands caused by the amplitude variation of this spectral component.

#A. HUSCHAUER, April 21, 2015Outlook Part IISimulations with PyORBIT

Ongoing investigation of the trapping process in 6DTracking of particles in the islands during the rise of the extraction bumpDirect and indirect space charge simulations (PIC, frozen, PIC-frozen,)

Experimentally

Get the observed capture oscillations under control to be able to continue with the commissioning Create islands at low energy without the PFW?Optimize various parameters as mentioned earlier bucketsStudy the transfer to the SPS#A. HUSCHAUER, April 21, 2015THANK YOU FOR YOUR ATTENTION!