CERN, BE-ABP (Accelerators and Beam Physics group) J. Pfingstner Beam jitter studies at ATF and ATF2 Jürgen Pfingstner, Hector Garcia Morales 15 th of February 2013

CERN, BE-ABP (Accelerators and Beam Physics group) J. Pfingstner Beam jitter studies at ATF and ATF2 Content 1.Measurement and analysis setup 2.Search for beam jitter sources 2.1 Damping ring 2.2 ATF2 beam line 3.Additional data analysis with MIA 3.Tracking studies 4.Conclusions Many thanks to Y. I. Kim, J. Snuverink, G. White and M. Woodley for supporting us with all necessary information at ATF and for the many very helpful discussions !

CERN, BE-ABP (Accelerators and Beam Physics group) J. Pfingstner Beam jitter studies at ATF and ATF2 1. Measurement and analysis setup

CERN, BE-ABP (Accelerators and Beam Physics group) J. Pfingstner Beam jitter studies at ATF and ATF2 Intention of the jitter studies Intention: 1.Study beam motion in damping ring and ATF2 beam line -> Beam jitter has to be reduced from 10-20% to below 5% 2.Locate beam motion sources, especially beam jitter sources 3.Study the performance of already operating orbit feedback systems and try to make suggestions for possible improvements if necessary Setup: 1.Some scripts to read BPM data in the damping ring and the ATF2 beam line have been written 2.Several data set have been taken under different conditions between the 13 th and 18 th of December 2012

CERN, BE-ABP (Accelerators and Beam Physics group) J. Pfingstner Beam jitter studies at ATF and ATF2 Used BPM systems and EPICS Used BPM systems 1.96 button-BPMs in the damping ring, resolution of 20μm, only 90 used since others did not work properly 2.12 strip-line BPMs at the beginning of the ATF2 beam line, resolution 5μm 3.34 cavity BPMs in the end of the ATF2 beam line, resolution 0.1μm EPICS system 1.The Experimental Physics and Industrial Control System (EPICS) is a collection of software tools to create distributed soft real-time applications for large scientific applications. 2.EPICS data (PVs) of BPMs have been accessed via Matlab commands

CERN, BE-ABP (Accelerators and Beam Physics group) J. Pfingstner Beam jitter studies at ATF and ATF2 Correlation coefficient: Power spectral density (PSD) estimated as: Integrated root mean square (IRMS): Full and differential motion: x k … full beam motion data … differential beam motion data Analysis tools … standard deviation … cross correlation … DFT of x k … acquisition time of x k … sample number in x k

CERN, BE-ABP (Accelerators and Beam Physics group) J. Pfingstner Beam jitter studies at ATF and ATF2 2. Search for beam jitter sources

CERN, BE-ABP (Accelerators and Beam Physics group) J. Pfingstner Beam jitter studies at ATF and ATF2 Method of finding beam jitter sources Step 1: Starting at BPM N, try to find correlation pattern by calculating the correlation coefficient r for all BPMs. Step 2a: If there is no significant correlation -> N=N+1. Step 2b: If there is significant correlation -> source has to be expected close to BPM N. Step 3: Remove motion that corresponds to found jitter source (see next slide). Step 4: Go to step 1 with N=N+1, or stop if N=N max. … standard deviation … cross correlation

CERN, BE-ABP (Accelerators and Beam Physics group) J. Pfingstner Beam jitter studies at ATF and ATF2 Removal of correlated beam motion -> See Hectors’ presentation

CERN, BE-ABP (Accelerators and Beam Physics group) J. Pfingstner Beam jitter studies at ATF and ATF2 2.1 Damping ring

CERN, BE-ABP (Accelerators and Beam Physics group) J. Pfingstner Beam jitter studies at ATF and ATF2 Full beam motion in the damping ring 1/3 Plots based on data from BPM MB52R (number 30)

CERN, BE-ABP (Accelerators and Beam Physics group) J. Pfingstner Beam jitter studies at ATF and ATF2 Full beam motion in the damping ring 2/3 Damping ring feedback system suppresses low frequencies, but creates new frequencies between 0.01 and 0.1 Hz due to a step-like feedback action. A bit more stronger feedback action and a smoother actuation would be beneficial. Mode-like excitation changes could be overcome with an adaptive controller based on a on-line PSD estimation. Plots based on data from BPM MB52R (number 30)

CERN, BE-ABP (Accelerators and Beam Physics group) J. Pfingstner Beam jitter studies at ATF and ATF2 Full beam motion in the damping ring 3/3 Strong correlation of the full beam motion in the damping ring. Correlation coefficient propagates with the phase advance. Only little correlation has been observed in the ATF2 beam line. Correlation coefficient of BPM MB46R (number 25) with all other BPMs

CERN, BE-ABP (Accelerators and Beam Physics group) J. Pfingstner Beam jitter studies at ATF and ATF2 Differential beam motion in the damping ring Differential beam motion is only slightly correlated in the damping ring (strong signal contribution of BPM noise). In the ATF2 beam line hardly any effect has been observed -> no orbit jitter from the damping ring (but only orbit data used) Correlation coefficient of BPM MB46R (number 25) with all other BPMs

CERN, BE-ABP (Accelerators and Beam Physics group) J. Pfingstner Beam jitter studies at ATF and ATF2 2.2 ATF2 beam line

CERN, BE-ABP (Accelerators and Beam Physics group) J. Pfingstner Beam jitter studies at ATF and ATF2 Source 1 Located between BPMs 97 and 102 (MQF7X and MQD12X) Position can only be determined before BPM 102, since signal to noise ration is too bad before. Cavity BPMs instead of strip-line BPMs would help The data of the strip-line had to be shifted by one time step w.r.t. the cavity BPMs to see correlation.

CERN, BE-ABP (Accelerators and Beam Physics group) J. Pfingstner Beam jitter studies at ATF and ATF2 Signal and noise levels BPM noise calculation from data as described in Kim et al. PRST Accel. and Beams 15, BPM 102 is the first BPM with sufficient signal to noise ratio. Location of the first source can only be determined before BPM 102. Better BPMs would help

CERN, BE-ABP (Accelerators and Beam Physics group) J. Pfingstner Beam jitter studies at ATF and ATF2 Source 2 Located at BPM 112 (MQM16FF) Very strong source Candidates: Wire scanner: MW3X, MW4X, Profile monitor: OTR3X, Correctors: ZH1FF, ZV1FF, QP: QM16FF

CERN, BE-ABP (Accelerators and Beam Physics group) J. Pfingstner Beam jitter studies at ATF and ATF2 Source 2 Measured vertical beam position data of the BPMs 111 (left) and 112 (right) versus BPM 134. A strong change in correlation can be seen.

CERN, BE-ABP (Accelerators and Beam Physics group) J. Pfingstner Beam jitter studies at ATF and ATF2 Source 3 Located at BPMs 118 and 119 (MQD10BFF and MQD10AFF) Candidates: Ref. cavity: MF3FF, Monitor MFB1FF QPs: QD10B, QD10AFF, QM11FF

CERN, BE-ABP (Accelerators and Beam Physics group) J. Pfingstner Beam jitter studies at ATF and ATF2 3. Additional data analysis with MIA

CERN, BE-ABP (Accelerators and Beam Physics group) J. Pfingstner Beam jitter studies at ATF and ATF2 Model Independent Analysis (MIA) Try to use methods described in paper by J. Irwin et al. PRL 82(8) about Model Independent Analysis (MIA) (thanks Young-Im) 1.Data cleaning via SVD decomposition: –Remove noisy BPMs –Remove BPM noise floor 2.Degree-of-Freedom plot (DoF-plot) –Connection of SVs for SVDs with increasing number of used BPMs. –Lines are the connections of largest, second largest, … SVs. –Change of slope indicates physical source. Methods all just try to find location of sources, but are not capable of determining the form of the according oscillation: “Note that each of the eigenmodes in Eq. (4) does not correspond uniquely to the physical pattern in Eq. (2).”

CERN, BE-ABP (Accelerators and Beam Physics group) J. Pfingstner Beam jitter studies at ATF and ATF2 DoF-plot for beam jitter data Change of slope indicates physical source. Only cavity BPM with good signal to noise ratio are used Change around BPM 118 (MQD10BFF), but also around 130 (MSD4FF) No clear localisation possible Observation of direction does not give good hinds of oscillation shape.

CERN, BE-ABP (Accelerators and Beam Physics group) J. Pfingstner Beam jitter studies at ATF and ATF2 DoF-plot of the jitter correlation matrix Much clearer localisation of sources is possible: BPM 112 (MQM16FF) and BPM 118 (MQD10BFF) Also a small change in slope around BPM 130 (MSD4FF) Result is the same as for manual method Tool for faster source localisation in the future!

CERN, BE-ABP (Accelerators and Beam Physics group) J. Pfingstner Beam jitter studies at ATF and ATF2 4. Tracking studies

CERN, BE-ABP (Accelerators and Beam Physics group) J. Pfingstner Beam jitter studies at ATF and ATF2 Kick tracking with Lucretia Best “guess” for shape of beam jitter along the beam line is the motion removed to de-correlate from a certain source. Try to localise sources better, by introducing dipole kicks along the beam line and compare shape of oscillations with removed correlated jitter motion. Using lattice version 5.1: 10x1 optics with non-linearities and new QF1 matching Track beam with perfect beam line to a certain position Apply dipole kick by adding const. Δy’ to beam. Use same beam to track to the end of the beam line. Compare BPM readings in the region of the cavity BPMs with good signal to noise ratio.

CERN, BE-ABP (Accelerators and Beam Physics group) J. Pfingstner Beam jitter studies at ATF and ATF2 Matching tracking results with removed correlated beam jitter motion Different kicks between BPM 111 and 112 fit the combination of both source 2 and source 3 reasonable well. Δy’ = -4e-7 (Lucretia units) Devices in this area: MW3X, OTR3X, TILTM, IPBPM, ATIPBPM, MW4X Separately sources were not fitted well

CERN, BE-ABP (Accelerators and Beam Physics group) J. Pfingstner Beam jitter studies at ATF and ATF2 PSD and IRMS of removed correlated beam jitter motion Source 1 and source 2 have obvious components at 0.945Hz and 0.963Hz These components do not contribute significantly to the overall jitter. Source 3 has no such components.

CERN, BE-ABP (Accelerators and Beam Physics group) J. Pfingstner Beam jitter studies at ATF and ATF2 5. Conclusions 1.No significant effect of beam motion in the damping ring on the beam motion in the ATF2 beam line has been observed. However, only orbit data have been used instead of single turn data. 2.Three beam jitter sources have been identified with correlation studies. 3.Source 1 can be not clearly localised due to limitations of the BPM system. 4.Complementary studies with MIA have been performed and results fit for correlation data. 5.Tracking simulations with Lucretia have been performed The removed jitter motion of the individual sources could not be fitted. But the combination of both motions can be fitted reasonable well. Some beam diagnostics devices are located in this region. No clear signature has been found however. 6.The removed jitter motion of source 1 and 2 show significant frequency peaks in PSD (0.945Hz and 0.963Hz). 7.No dispersive patterns have been observed. Therefore, the influence of the beam energy seems to be negligible. 8.Improvements of current instrumentation system have been proposed: smoother actuation of the the DR orbit FB and synchronisation problem of BPM system.

CERN, BE-ABP (Accelerators and Beam Physics group) J. Pfingstner Beam jitter studies at ATF and ATF2 More Details Hector’s talk CLIC note 986 will be very soon on-line available

CERN, BE-ABP (Accelerators and Beam Physics group) J. Pfingstner Beam jitter studies at ATF and ATF2 Thank you for your attention!