Vibrating Wire R&D for Alignment of Multipole Magnets in NSLS-II

Transcription

1 Vibrating Wire R&D for Alignment of Multipole Magnets in NSLS-II 10 th International Workshop on Accelerator Alignment February 11-15, 2008, Tsukuba, Japan Animesh Jain for the NSLS-II magnet team

2 Collaborators M. Anerella, G. Ganetis, P. He, P. Joshi, P. Kovach, S. Plate, J. Skaritka Brookhaven National Laboratory, Upton, NY, USA and Alexander Temnykh Cornell University, Ithaca, NY 1

3 Introduction For optimum performance, the magnetic axes of quadrupoles and sextupoles in NSLS-II should be aligned to better than ±30 microns. Optical survey accuracy (~50 microns) is inadequate to achieve the required tolerance. It is difficult, and expensive, to maintain the required machining and assembly tolerances in a long support structure (~5 m) holding several magnets. It is desirable to achieve the required alignment using direct magnetic measurements in a string of magnets. 2

4 Magnet Alignment R&D Several magnets, including multipoles and corrector dipoles, will be installed on a girder ~5-6 m long. Based on the accuracy required, and the overall length of the girders, the vibrating wire technique developed at Cornell was deemed most appropriate for this task. An R&D program was initiated to develop the technique at BNL and demonstrate the required accuracy. Preliminary work was carried out using a temporary setup with help from Cornell and staff from NSLS. A new R&D setup was designed and is now operational. 3

5 The Vibrating Wire Technique: Basics In this technique, an AC current is passed through a wire stretched axially in the magnet. Any transverse field at the wire location exerts a periodic force on the wire, thus exciting vibrations. The vibrations are enhanced if the driving frequency is close to one of the resonant frequencies, giving high sensitivity. The vibration amplitudes are studied as a function of wire position to determine the transverse field profile, from which the magnetic axis can be derived. Vibration amplitudes measured at many resonant modes can also give the axial distribution of field along the wire. 4

6 Vibrating Wire R&D Setup at BNL Wire Vibration detectors (~ 13 mv/micron) X-Y Stages (2.5 micron accuracy) Magnet position adjusters (1 micron resolution) Granite table for supporting magnets during R&D phase (to be replaced eventually with an actual girder) Unique feature: Vibration sensors are installed on both ends, giving two simultaneous measurements 5

7 Vibrating Wire R&D Setup: Wire Ends & Sensors Fiducial nests (earlier version) Ceramic V-notch X-Detector Y-Detector Camera to ensure that wire is seated correctly Wire Ends Details Fiducials relate the wire ends to the overall girder coordinate system. New version has 4 fiducials Wire Vibration Sensors 6

8 Vibrating Wire R&D Setup: Manual Magnet Movers Magnet Position Adjusters (Fine and coarse adjustments using differential screws) First version with stainless steel parts did not work very smoothly. New version with Silicon-Bronze parts works well. Dial indicators to monitor magnet motion. Mounting of horizontal indicators is now improved from an earlier version. 7

9 Survey Equipment Laser Trackers To be used for survey of girder and wire fiducials. Portable CMM machines Can be used to relate magnet positions to the girder fiducials with better accuracy than laser trackers. 8

10 SLS Magnets in the Vibrating Wire Setup SLS Quadrupole SLS Sextupole SLS Quadrupole at 80 A: db y dx dz = 3.19 T db y dx = 13.8 T/m Magnets can be run at currents up to 140 A, but saturation begins at ~80 A SLS Sextupole at 80 A: d dx 2 d dx 2 By 2 B y 2 = dz = 94.2 T/m T/m 9

11 Resonant Frequency and Wire Sag <A_x*I> or <A_y*I> A_x1 A_x1Calc A_y1 A_y1Calc A_x2 A_x2Calc A_y2 A_y2Calc Mode = 6 f f Y ( f ) = A 0 + Y 2 2 f ( f f ) + fγ Frequency (Hz) f x1= Hz f y1= Hz f x2= Hz f y2= Hz ~7.3 m long, mm dia. Cu-Be wire; 1.05 Kg weight f 0 = Sag Sag = 1 2L = T m l 2 mlgl 8T g 2 32 f 0 Correction for large wire sag (~ microns for ~7.3 m length) is very important, which in turn requires a very precise knowlege of resonant frequency. ±0.02 Hz ± 1 micron T = Tension L = Length m l = mass per unit length 10

12 Stability of Wire Sag Over Several Days Computed Sag (microns) SLS Quadrupole October 26-29, 2007 Using X Detector Using Y Detector SLS Sextupole Measurement Sequence Number Resonant frequency, and hence sag, is continuously derived as part of the measurements. Thus stability is strictly needed only at ~30 min. time scale. Oct. 30-Nov.1,

13 Quadrupole Measurements: Horizontal Scans X-detector Signals ( By) A_X1 40A_X1 60A_X1 80A_X1 0A_X2 40A_X2 60A_X2 80A_X2 14-Jan-2008 X Center is given by intersection with 0A line Current X1_Center X2_Center 40 A A A Wire X-Position (mm) 12

14 Quadrupole Measurements: Vertical Scans Y-detector Signals ( Bx ) A_Y1 40A_Y1 60A_Y1 80A_Y1 0A_Y2 40A_Y2 60A_Y2 80A_Y2 14-Jan-2008 Y Center is given by intersection with 0A line Current Y1_Center Y2_Center 40 A A A Wire Y-Position (mm) 13

15 Quadrupole Measurements Reproducibility Sensor #1 Sensor #3 16-Jan-2008 Std.Dev. = mm Horizontal Center (mm) Measurement Sequence Number 14

16 Quadrupole Measurements Reproducibility Vertical Center (mm) Sensor #2 Sensor #4 16-Jan-2008 Std.Dev. = mm Systematic difference seen between the two sensors! Measurement Sequence Number 15

17 Study of Measurement Resolution: Concept Measurements have been made in a SLS quadrupole and a SLS sextupole to study the measurement resolution achievable in each case. Magnetic center was first measured in the as-installed position of the magnet using vibrating wire technique. The magnet was then moved either horizontally or vertically by a known amount, as monitored by dial indicators. Vibrating wire measurements were made again and the results compared against dial indicators. 16

18 Correlation with Magnet Moves (Quad; Horiz.) (microns) Magnetic Center Change Ideal Line Measured Data Magnetic center follows magnet movement within < 1 micron! Data taken with only one set of sensors operational SLS Quadrupole QA42 at 80A 30-Oct Magnet Horizontal Movement (micron) 17

19 Correlation with Magnet Moves (Quad; Vertical) (microns) Magnetic Center Change Ideal Line Measured Data SLS Quadrupole QA42 at 80A 31-Oct-2007 Data taken with only one set of sensors operational Magnetic center follows magnet movement within < 5 microns Magnet Vertical Movement (microns) 18

20 Sextupole Measurements: Horizontal Scan Signal (arbitrary units) SLS Sextupole SR110 at 80 A (Mode = 6); 22-Jan-08 X1 Sensor (B_y) Y1 Sensor (B_x) X2 Sensor (B_y) Y2 Sensor (B_x) Parabolic fits -120 Horizontal center is defined as the point of zero slope in B_y Vs. X Wire Horizontal Position (mm) 19

21 Sextupole Measurements: Vertical Scan Signal (arbitrary units) SLS Sextupole SR110 at 80 A (Mode = 6); 22-Jan-08 Parabolic fits X1 Sensor (B_y) Y1 Sensor (B_x) X2 Sensor (B_y) Y2 Sensor (B_x) Vertical center is defined as the point of zero slope in B_y Vs. Y Wire Vertical Position (mm) 20

22 Sextupole Measurement Reproducibility Horizontal Center (mm) Systematic difference seen between the two sensors! 17 & 18-Jan-2008 Std.Dev. = mm (2nd Day Data Only) Measurement Sequence Number Sensor #1 Sensor #1 Next Day Sensor #3 Sensor #3 Next Day 21

23 Sextupole Measurement Reproducibility Vertical Center (mm) & 18-Jan-2008 Std.Dev. = mm (2nd Day Data Only) Sensor #1 Sensor #1 Next Day Sensor #3 Sensor #3 Next Day Measurement Sequence Number 22

24 Sextupole Measurements Using B x Instead of B y Obtaining centers from B_y vs. X and B_y vs. Y plots uses only one set of sensors, and requires quadratic fits. One could also use scans of B_x vs. X (or Y) for various values of Y (or X). These plots are expected to be linear with slopes proportional to offsets in Y (or X) direction. Doing three such scans allows to obtain centers from both B_x and B_y data. With 2 sets of sensors, one gets four values of magnetic center. B y = B 3 ( x x 0 ) 2 R ( y 2 ref y 0 ) 2 B x = 2B 3 ( x x )( y 0 2 Rref y 0 ) 23

25 Sextupole Measurements: B_x vs. X Scans 40 SLS Sextupole SR110 at 80 A (Mode = 6); 23-Jan-08 Y1 Sensor (Y = 0.5 mm) Y1 Sensor (Y = 0 mm) Y1 Sensor (Y = 0.5 mm) Y2 Sensor (Y = 0.5 mm) Y2 Sensor (Y = 0 mm) Y2 Sensor (Y = 0.5 mm) Bx (arbitrary units) Vertical center is defined as the point where B_x Vs. X has zero slope Wire Horizontal Position (mm) 24

26 Comparison of Sextupole Data Using B_x and B_y Quantity X_Center (micron) Y_Center (micron) SLS Sextupole at 80 A (23-Jan-2008) Sensors Used B_x Data B_y Data Mode=6 Mode=8 Mode=6 Mode=8 Pulley End Fixed End Pulley End Fixed End B_x data from the two sensors show better consistency. Systematic differences between results using two different modes are significant. 25

27 Issue of Background Fields in Sextupole Meas. There is a significant quadrupole background field from quadrupole magnet(s) even when these are unpowered. Based on rotating coil data, the remnant integrated quadrupole field is ~0.02 T in the SLS quadrupole. For a sextupole with integral field of 94.2 T/m at 80A, this could amount to a change in horizontal center by hundreds of microns, depending on quad position and mode used. The vertical center measurement is not affected because B y (or B x ) is independent of y (or x) in a quadrupole field. Corrections must be made for this background field. 26

28 Correlation with Magnet Moves (Sextupole; Horiz.) (microns) Magnetic Center Change Ideal Line Measured Data Magnetic center follows magnet movement within < 6 microns Data taken with only one set of sensors operational SLS Sextupole SR110 at 80A 26-Oct Magnet Horizontal Movement (microns) 27

29 Correlation with Magnet Moves (Sextupole; Vertical) (microns) Magnetic Center Change There is a systematic offset of ~13 microns. Possible reasons could be an error in the reference measurement, or temperature variations, which were not monitored. This needs further study. Measured Data Ideal Line Shifted Line SLS Sextupole SR110 at 80A 29-Oct-2007 Data taken with only one set of sensors operational Magnet Vertical Movement (microns) 28

30 Procedure for Multipole Alignment on a Girder Install magnets on a girder, install vacuum chamber, and carry out a rough alignment. Set up girder on a vibrating wire test stand in a temperature controlled environment, and wait for steady temperature. Determine center of each magnet relative to the wire coordinates. Move the magnets to locate the magnetic centers on a line joining the two end points of the wire. Lock the magnets in place, while monitoring the magnet positions using displacement gauges. Survey the wire ends, and all girder and magnet fiducials (?) using laser trackers and portable CMM machines. 29

31 Magnet Movers for Alignment on a Girder Actuator Gauge head to monitor motion Spring to prevent backlash in horiz. motion Magnet movers will be installed on the girder, and then removed after the magnets are aligned and locked in place. A test was performed to demonstrate ability to easily lock a magnet in place within 5-10 microns. 30

32 Future Work Study (and improve) the absolute accuracy of the measurements (better than ~10 microns desirable). Resolve various inconsistencies in detector responses. Survey of wire end V-notches relative to fiducials on test stand. Prototypes for motorized magnet movers. Integrate various components of hardware and software needed for measurements into a single, fully automated system for multipole alignment on a girder. 31

33 Conclusions A new vibrating wire R&D system has been designed, built and assembled. The R&D system has been used to measure a quadrupole and a sextupole received on loan from the Swiss Light Source. Unique feature of dual sensors allows extensive checks of consistency and systematic accuracy. Good correlation between magnet position and magnetic center has been shown (well within the required tolerance) for both quadrupole and sextupole magnets. Work is underway to further improve the accuracy of the system, and to automate the entire measurement and alignment sequence. 32

34 Acknowlegements J. Cintorino, S. Dimaiuta, J. Escallier, D. Harder, W. Louie, J. Mc Caffrey, D. Oldham, S. Ozaki, G. Rakowsky, P. Ribaudo, A. Sauerwald, S. Sharma, D. Sullivan, P. Wanderer, F. Willeke Brookhaven National Laboratory L. Rivkin, D. George Swiss Light Source 33