1. Highlights of the YAG measurements carried out October 9-13, 2007 Recycler Weekly Departmental Meeting October 10 th, 2007 L. Prost, A. Shemyakin, A. Warner, M. Sutherland, K. Carlson, A. Burov

2. 2 Thanks to all involved !  Instrumentation setup including installing the camera in the tunnel and proper timing in the control room took significant efforts from  Arden, Mary, Kermit, Randy  Additional problem with SPAGNI failure and Kautz power station outage  Great responsiveness from EE and Mechanical support Thanks to Ron, Jerry, Fernando, Rich, Jeff, Greg, Kermit  Taking and analyzing the data was a collective effort from  Mary, Arden, Kermit, Timofei and Alexey (& Sasha and I)  A lot had to happen in a short period of time to make these measurements fruitful  They were very successful Hopefully, it will directly translate into faster cooling !

3. 3 GOAL  Before the shutdown, YAG images were taken and showed that the beam at the exit of the cooling section was very elliptical, indicating a strong quadrupole component  Large angles  The goal of the measurements was to correct the ellipticity  Make the beam cylindrical in the cooling section i.e. round AND low envelope scalloping

4. 4 Camera setup (I)  Camera located at the exit of the cooling section (past the 180  bend magnet) Electron beam direction CID camera 180  bend magnet YAG crystal drive (towards the outside wall) Note: The constant focusing direction is y (vertical) for the camera but actually x (horizontal) for the beam. y x SPQ01I (not visible on the picture) is the last solenoid before the YAG Beam reference system

5. 5 Camera setup (II)  Camera settings  Screen = 6 kV  MCP = -600 V  Gate/Photo cathode = -200 V  Gate width = 100 ns Gate position: ~Max of the beam pulse YAG crystal Ti wires for calibration

6. 6 Timing  Ref. timing for the gate delay based on SPQ01I = 0 A image  Drift of the order of 2  s was observed over a shift Corrected by adjusting gate delay Acquisition window Gate pulse Control electrode pulse ‘Real’ gate position w.r.t. beam pulse depends on cable length… Ref. image (inverted) Adjusted timing to return to this image

7. 7  In the y-direction (constant focusing)  From geometrical considerations: 15 pixels/mm  Confirmed with measurements with inserted scraper and YAG motion Images of a wire placed in front of the YAG crystal overlap when the YAG is moved by 3 mm and the corresponding image shifted by 45 pixels in the YAG direction of motion (x-direction)  Trials to do direct measurements for calibration in the other (x) direction have so far failed !!  Data are inconsistent Calibration Sum of two images: YAG at position 1 and YAG at position 2, with the image shifted by 45 pixels Image of the wire in front of the YAG x y

8. 8 Distortion determination (I)  Ideally, the YAG crystal makes a 45  angle with respect to the axis of the camera  Gives distortion k ≡ cotan  between X and Y directions At small currents, an effect of a solenoid is reduced to a beam turn by. Due to a possible deviation of the YAG tilt from 45 degrees, the visible angle  is different from its ideal value , which is described as At  << 1, it is equivalent to  Use SPQ01I to verify/determine k A. Burov In the YAG setup, this is the X-direction of the camera which is affected by this angle    cotan 

9. 9  Put scrapers in to create a ‘line of reference’ which will rotate as a function of SPQ01I  Identify rotation of the whole beam by fitting an ellipse Distortion determination (II) Fitted ellipse angle Threshold image (auto) Slope is d  /dI

10. 10  Using and taking two scraper data sets, we find C = 2.4  /A (design: 2.3  /A) and k = 0.84 (i.e. 50  angle instead of 45  )  Additional data sets (i.e. SPQ01 scans and associated rotation of the fitted ellipse) agree rather well Distortion determination (III) ? d  /dI calculated with C = 2.4  /A and k = 0.84 Dependence of d  /dI on the apparent angle 

11. 11 Beam rounding procedure (A. Burov, Dpt Meeting July 25 th )  For two distinct values of SPQ01I  Record initial image  Change upstream quads successively (6 quads) Record associated images  Calculate ellipticities  Fit ellipse to threshold (binary) image  Extract semi-major and semi-minor  e = 2 (a –b)/(a+b)  Include effect of camera angle (i.e. distortion k )  Compute MULT  SVD algorithm  Use MULT to make the beam more round  Repeat… If e = 0 for two values of SPQ01I, then the beam is perfectly round This was done automatically (eventually) with a Java application written by Timofei Boshakov

12. 12 Note on the image analysis  Once the image is recorded, the first step of the analysis consists in creating a binary image, for which a threshold (0- 255) is set, somewhat arbitrarily  This threshold determines how much of the beam (‘core’, ‘core + halo’,…) is taken into account in the calculation of the best fit ellipse For the automation process we choose a threshold value of 55 i.e. selection of the core of the beam Original image (inverted) Threshold Fit ellipses Auto (here 22)33 (i.e. ‘core’) 4 (i.e. ‘halo’)

13. 13 Preliminary results (after multiple iterations) First visible image + 1.5  s 118 119 120 121 122123 Nominal New nominal

14. 14 Some untold details…  Once the beam for our nominal settings was considered round enough  Switched to DC mode and adjusted correctors such as to recirculate 200 mA  Verified that steering did not introduce different quadrupole components Made an additional scan with quadrupole –Ellipticities found were similar to before steering We did see smaller ellipticities than reported in previous table –Timing drift ?  Checked pepper-pot image Fairly rectangular and regular pattern

15. 15 Adjusting ‘cylindricity’  Additional SPQ01I scans are needed to determine /correct for the magnitude of envelope scalloping  The scan gives information about the beam radius, angular momentum and radial divergence (fit)  Back propagate to the entrance of B1 lens in OptiM Get initial conditions  Adjust B1 and B2 to have a cylindrical beam in the CS  Applied this procedure to our nominal file (Alexey)  Got new settings forSPB01I & SPB02I High current values obtained may be a problem (see further)  Repeated rounding procedure with new settings Analysis of the final results underway YAG screen

16. 16 Higher order multipoles seem to be generated in the first bend  Upstream (of the first bend) focusing and steering scans show very large effects on the beam shape while downstream scans primarily change the beam size and rotation.  Example of a scan with SPA06 (last solenoid before the first bend) at 190G in the cooling section (and corresponding field at the cathode) This effect likely more pronounced at 190G because of the larger effective emittance (thus beam size) SPA06I = 14.4 ASPA06I = 15.4 ASPA06I = 16.4 ASPA06I = 17.4 A Nominal

17. 17 Some (possible) consequence  Because the beam size increases for higher current (it actually has a maximum at ~250 mA), the multipole component of the beam may be different at different beam currents  Example for the nominal case ‘rounded’ Not obvious…  More visible on a non-optimal case U pulse = 3.2 kV I eq =0.1 A U pulse = 4 kV I eq = 0.3 A U pulse = 4.5 kV I eq = 0.6 A Ellipticity increases ~2×

18. 18 Quality of the solenoidal field appears to degrade at high currents  When the lens (SPQ01 or SPB01 for instance) current becomes relatively large (1/2 of its maximum?) the beam shape starts degrading (i.e. does not remain round or at least elliptical)  High non-linearities at high field, even near the axis ? SPQ01I scan, Upulse = 4.5 kV

19. 19 ‘Last minute’ analyses (from Alexey)  Fit of SPQ01A scan for the ‘new’ nominal file:  Give r = 2.4 mm,  = 0.01,  = 2.0 m for initial conditions at the entrance of the cooling section  Show that 10-15% envelope oscillation remain Improvements expected with SPB01I = 20.4 A & SPB02I = 8.0 A SPQ01I [A] Beam radius [cm] SPQ01 scan w/ fit Corresponding envelope in cooling section ~0.6 mm r = 2.4 mm

20. 20 Conclusion  Beam rounding converged and we achieved ellipticities of the order of 10-20%  Two files available: Before shutdown nominal corrected for its ellipticity Improved nominal (new SPB01 and SPB02 settings intended to make the beam more cylindrical in the cooling section) –SPQ01 scan indicates that ~15% oscillation may remain (other settings proposed by Alexey could be tried)  Did one set of rounding measurements at 190 G Not as successful as for 105 G –We seem to have too strong sextupole components  Several questions remain  Consistent calibration of the x-direction  Effect of steering/focusing upstream of the first bend  Possible non-linearities from lenses at high current i.e. any change to lenses with already relatively high current changes the multipole component in the beam