ATF Emittance Studies - Analysis of March shifts (Mark Woodley) : The vertical bump generated in the ring around QM7 also induces some growth of the vertical beam size on the XSR monitor. Translated in terms of emittance, the growth is a factor 1.5 over the range studied (from 0.7 to +0.6 mm vertical offset in QM7). No correlation with bunch intensity was found, however some not insignificant vertical leakage of the bump in the rest of the ring was observed, which may be the source of this emittance growth (through x-y coupling and vertical dispersion + radiation effects). As a next step, the emittance growth from this leakage will be simulated. In addition, the effect of the vertical dispersion from the bump on the projected emittance in the EXT line will be computed. Mark also described his plans to (1) use BPMs near the extraction to infer the actual strength of QM7R, (2) simulate the dispersion measurement in the EXT line to understand whether the parabolic responses are reasonable, and (3) identify the source of the error in the EXT optics error. It was commented that from the fit of BPMs planned for (1), one could also evaluate the stability of the vertical trajectory in QM7 on different time-scales, which could be important to assess whether dedicated feedback should be planned.
Emittancegrowthstudiesusingstaticbumpsin theATF EXT line (Maria Alabau) : Measurements with the OTR and XSR from the shift on March 4 were analysed. In the range 0.1-0.5 mm for the QM7 bump amplitude, the vertical beam size is constant on the XSR (and hence presumably also vertical ring emittance). In this range, the beam size was measured to increase on the OTR by about 10%. A similar effect (of about 20%) was seen in the December measurement, with however less precision. The expected OTR beam size was also simulated as function of bump amplitude, for nominal Twiss parameters, and nominal vertical and horizontal input emittances of 12 pm and 1.2 nm, and for a vertical emittance increased by a factor 4, which is the typical kind of value measured in December in the ring. With such an increase, there a good agreement with the December data can be seen, which would tend to confirm the hypothesis that the source of the increase is indeed the skew quad component generated from the vertical offset in QM7. The curves also show that vertical emittance must be relatively small in the ring if sizeable increases of the beam size on the OTR are to be observed experimentally. It was commented (after the meeting) that these simulated plots should be redone with Twiss parameters consistent with those from the periodic solution with the instantaneous ring optics (rather than nominal values). They should also be re-expressed in terms of equivalent vertical emittance for easier comparison with the values measured with wire-scanners and the comparason should also be done with the March data. Finally, slightly different horizontal input emittances (increased by factors 1.2 or 1.5) could also be tried to check for the match of the slope of the increase with respect to bump amplitude.
Twiss fitting techniques and ellipse plot reconstructions (Tony Scarfe) : The 12 March multiwire beam size measurements have been analysed using a least-square minimiser to fit the Twiss parameters (referenced to MW0X) and emittance values, giving values nicely consistent within the errors with those obtained by Julien at LAL. Tracking the beam sizes in the region of the wire-scanners, based on these fitted parameters, gives values consistent with the measurements, which is indicates that the fit should be of reasonable quality. Tracking the design nominal parameters gives on the other hand large differences. The phase advance between wire-scanners is quite small for the fitted parameters. Twiss parameters were also back-propagated to the injection assuming constant 2x2 emittances. Fairly large differences are found with respect to the nominal values, though the significance cannot be assessed yet in the absence of an evaluation of the error propagation (a plan to calculate the measurement uncertainty was described by Tony and will be implemented in the near future). Ellipse plots similar to those made by SAD were computed, which confirm that in the present optics and with present input parameters, the three wire-scanner measurements are essentially at the same betatron phase.
Determination and back-propagation of Twiss parameters (Julien Brossard) : Twiss parameters and emittances determined during the March 12 shift by multi-wire and quad scan measurements were back-propagated to the entrance of the EXT line, along with their errors,
based on toy Monte Carlo generations. The optics had previously been fitted by Mark Woodley using as input parameters obtained from a quad scan measurement made on March 4, based on an existing optics used for laser wire measurements. The beam could be focused
in the diagnostic section such that the phase advances were relatively large between wire-scanner. From the March 12 measurements, the most precise back-propagated Twiss parameters are from the multi-wire scanners in the horizontal plane and from a quad scan for
the vertical plane. However, it was noted that the back-propagation may not be entirely correct as the error on gamma was treated independently from those of beta and alpha. This will be corrected for future estimations. In the vertical plane the back-propagated parameters from the quad scan were also compared with those from the quad scan from March 4 and found compatible within errors. Spot sizes and phase advances were also shown all along the EXT line for different input assumptions (quad scan, multiwire and damping ring periodic solution), indicating good compatibility with wire-scanner measurements in the vertical plane, though not really with the quad scans. For the horizontal plane, the errors in the back-propagation from wire-scanner measurements and the central values for the different assumptions show that results in the diagnostic section are not too sensitive to the input. The phase advances between wire-scanners and skew quads varies significantly, especially in the vertical plane. More work is needed to complete this analysis and some of the steps were discussed.
Attempts to correct coupling from 12 March measurements (Cécile Rimbault) : Beam size measurements were performed at MW3X wire scanner position as function of a quadrupole scan with QF5X and a skew quadrupole scan with QK3X. From these measurements parabolas were reconstructed, providing the Twiss parameters at QF5X.The parabola reconstructed from QK3X showed an explicit coupling since it was not centered to zero. Using simulation based on the real machine status, Twiss parameters were back-propagated to the entrance of the extraction line.Then we tried to reproduce the coupling measurements, assuming a source of coupling at QM7. To do that, we added in the simulation a skew quadrupole (with the same properties as QK3X) at QM7, for which we varied the strength to fit the QK3X parabola.The strength needed to reproduced measurement was found to be 0.018m-1, corresponding to a current of 3.5 A. A simulation based on combined variations of QK1X and QK3X was performed to attempt minimising the vertical emittance at MW3X. The
best value was found for QK1X=-2.2A and QK3X=-4.5A. We applied those coupling corrections and performed new measurements of the vertical beam size at MW3X with a new scan of QK3X, and QF5X. Parabola from QK3X showed a new minima corresponding to the correction given by the simulation.
Parabola from QF5X showed an explicit vertical size reduction at MW3X of 20%. More studies are needed to understand and reproduce the measurements of the 12th March.
There are minutes attached to this event.
Show them.
