当前位置:首页 >> >>

The Single Beam Dynamics Study at CESR 1


CBN 97-2

The Single Beam Dynamics Study at CESR 1
Alexander Temnykh2 Laboratory of Nuclear Studies Cornell University, Ithaca NY 14953, USA January 10, 1997
The described in 1 nonlinear eld components of CESR elements have been included in model structure. Then using program MAD 2 the single beam dynamics was explored. The calculated variation of betatron tunes versus orbit position and the excitation of the nonlinear betatron resonances found with tracking are in good agreement with measurements. The elements given the main contribution into tune variation as well as into the nonlinear coupling resonances excitation have been detected.

Abstract

Model Nonlinearities
The CESR nonlinear eld components included in model have been described in 1 . The e ective nonlinearity of a element was represented by thin multipole placed in structure next after it. In the case of quads, sextupoles and horizontal separators the strength of these multipoles was proportional to the main element strength. The nonlinearity of vertical steerings was incorporeted into mode as the following. First, using the save set data and vertical steerings calibration, its de ection angles have been calculated, then its nonlinearities were determined according to magnetic measurements. The
1 2

Work supported by the National Science Foundation On leave from BINP, Novosibirsk

1

corresponding multipoles were placed in model structure in the location of these steerings. Below is the list of expressions used for the calculation of the nonlinear magnetic eld components of the di erent types of CESR elements. It is based on the data described in 1 and is given in the form of MAD de nition. Quads. The table below is the summary of coe cients used for calculation of quads mutipoles strength. Quad type REC West REC East Q1 West Q1 East Q2 West Q2 East Mark II

K 2L=K 1L 1=m 1:752  10,3 1:630  10,2 4:026  10,3 1:270  10,2 1:469  10,2 5:106  10,3 0:0

K 5L=K 1L 1=m4 1:090  104 5:192  104 1:170  104 1:170  104 6:042  103 8:397  103 1:094  105

K 9L=K 1L 1=m8 ,3:676  1013 ,6:423  1013 ,1:505  1012 ,1:505  1012 ,1:131  1012 ,1:222  1012 ,7:541  1013

Here K 1L is the quad strength and all multipoles are normal type. Sextupoles. To calculate nonlinearity generated by sextupoles the following equations have been used.

K 8L 1=m8 = 7:2  1010  K 2L 1=m2
Here K 2L is the sextupoles strength and the corresponding multipoles are normal type. Horizontal separators. The following expressions have been used to calculate its e ective nonlinearity.

K 2L 1=m2 = 5:674  101  rad K 4L 1=m4 = ,1:703  106  rad
Here is the de ection angle, corresponding multipoles are normal type. 2

Vertical steerings in sextupole magnets. The e ective nonlinearity of these elements was calculated as:

K 4L 1=m4 = ,1:824  106  rad
Here is the de ection angle. Corresponding multipoles are skew type. It must be mentioned that the completed information about nonlinear components of the all CESR elements is in not available. In the presenting simulation only the precise known nonlinearities of magnetic elements have been used. In addition, there is another kind of magnetic eld nonlinearity described in 3 . It is the eld perturbation due to the magnetic material contained in CESR IR beam pipe. The strength of it was estimated but not exactly known. Thus it has not been included in model. This type of eld distortion as well as the other unknown parasitic magnetic elds may cause the di erence between measurements and calculations.

Tunes Variation
The most obvious evidence of the CESR leading magnetic eld nonlinearity is the betatron tunes variation with orbit displacement. Figure 1 shows the measured and calculated dependence of betatron tunes versus so called PR1 orbit distortion. This type of distortion made with electrostatic plates is been using at CESR to separate counter rotating beams in horizontal plane at the parasitic interaction points. Horizontal axes on plot gives the distortion amplitude in relative units, 1000 of it corresponds to approximately 6mm of horizontal orbit o set in arcs. Vertical axes is the tunes shift. One can see negligible di erence between measurement and calculation for horizontal tune shift and the bigger di erence for vertical tune shift. One of the reason for this di erence may be the ignored magnetic eld distortions in REC quads caused by known imperfection of vacuum chamber, see 3 . As at these quads location the vertical beta-function is much bigger than horizontal one, the e ect of the eld perturbation on beam dynamic in vertical plane is much bigger than in horizontal. The next plots 2 and 3 show the calculated e ect of the two main contributors into the betatron tunes variation. These are Mark II quads used in regular structure in arcs and the horizontal separators. 3

The total horizontal tune shift is about 8kHz at 3000 units of PR1 and depends on pretzel as PR14. The quads nonlinearity gives approximatly 5kHz of tune rise, it is 60  of the total amount. Horizontal separators give additional 3 kHz or 40 . The e ect of these nonlinearities on vertical tune is 2.7 times smaller, see gure 3.

Resonance Excitation and Coupling Di erence
Another e ect caused by nonlinear components of leading magnetic eld is the resonances excitation and the coupling di erence between two beams due to its orbit di erence. Resonance excitation results in the vertical beam emittance growth as well as in the beam lifetime reduction. The coupling di erence may cause the vertical size variance between two beams and its misaligment in IP. All that may hurt the machine performance. The tune scanning technique developed in 4 has been used to observe the resonances at CESR. Figure 4 and 5 show the result of two tune plane scans made with atten orbit and with orbit distorted with pretzel. The both gures show vertical beam size. In the case of atten orbit, gure 4, one can see fh , fv + fs = nf0 resonance line crossed tune plane from lower left corner to the right upper. Is is manifested by the vertical beam size increase. At the right lower corner there is trace of fh , fv = nf0 resonance. In the case of pretzeled orbit, gure 5, the resonance fh , fv + fs = nf0 became stronger and the additional resonance 3fh , fv = nf0 is appeared. The simulation of these two tune plan scans was done with tracking using MAD program. The quads nonlinearities have been calculated from designed quads strength. The sextupoles nonlinear eld components were determined using CESR save set and the coe cients described above. The same save set was used to determine the vertical steering de ecting angels and the steering nonlinearities. In model structure the steering angels was omitted and only its nonlinearities were included. The particles starting points for tracking were chosen nearby the beam center: 1:5 v in vertical, 1:5 h in horizontal plane and 1:5 e for synchrotron motion. It is in so called the beam core region, where the particles motion distortion must be seen as beam size change. The 40x40 grid in tune plane has been used to scan. In each of the 1600 points, 800 turns was tracked and the maximums of vertical and horizontal amplitudes have been recorded. Figure 6 and 7 show the 4

results of tracking. The rst one been for the case of atten orbit does not indicate any resonances in range fh 205kHz where is the CESR working point is usually placed and where is experimental tune scan was done. In contrary with it the tracking with pretzeled orbit, see gure 7, shown the fh , fv + fs = nf0 and 3fh , 3fv = nf0 resonances appearance. It is consist with experimental facts. The next step was to identify the elements caused these resonances excitation. Figure 8 is result of tracking under the same condition as previous but without vertical steering nonlinearity. Here is no fh , fv + fs = nf0 and 3fh , fv = nf0 resonances. So, one can conclude that vertical steering nonlinearity is responsible for these resonances excitation. Note that in the real machine experiments, it is not easy make such identi cation. Any change of vertical steering nonlinearity is been accompanied with the change of its de ection angle leading to the beam orbit displacement. Due to that there are many other parasitic e ects masking wanted ones. Let's rate CESR vertical steerings according to its resonance driving strength. It may be done in the following way. Consider ring with only one vertical steering. The Hamiltonian, H , describing particals motion may be written as:

H = H0 + S z  y  x4 1 Here the second term on the right side is for vertical steering nonlinearity e ect. It describes vertical kick, which depends on horizontal coordinate as x4 . Function S z gives the strength of steering and its locations. The term H0 on the right side of 1 is for the rest of the elements. The horizontal coordinate x may be written as: x = x + xs + p 2 Here x describes betatron motion, xs is for synchrotron, p is orbit o set due to pretzel. Make standard transformation to action-angle variables and substitute equation 2 into 1 one can obtain:
1 1 H = H0 + S    1=2 y =2  cos y    1=2 x=2  cos x + e cos s + p4 3 y x Here x, y and are horizontal, vertical betatron functions and dispersion in the steering location; x, y and e are action variables corresponding

5

to the horizontal, vertical and synchrotron motion; x, y , s are phase variables. After simple algebraic manipulation one can determine terms responsible for the discussed above resonances driving and for the e ect of the coupling di erence between two beams. The coupling di erence driving term.

Vcoupl 

x

1=2 1=2 p3
y

4 5 6

The fh , fv + fs = nf0 resonance driving term.

Vfh,fv +fs=nf0  V3fh,fv =nf0 

x

1=2 1=2 p2
y

The 3fh , fv = nf0 resonance driving term.
x

3=2 1=2 p
y

So, in order to rate the vertical steerings according to its resonance driving strength, the combination of functions, dispersion and pretzel on the right side of equations 4,5 and 6 must be used as criteria. The table below shows this rating.

6

Coupling di erence 3fh , fv = nf0 driving fh , fv + fs = nf0 driving V. str rel. strength V. str rel. strength V. str rel. strength V34 2.381e+05 V34 16564 V34 1.93252e+04 V10 1.313e+05 V10 8853.0 V10 9.61728e+03 V33 6.309e+04 V08 6445.9 V08 8.63999e+03 V08 5.585e+04 V33 5380.8 V33 5.40950e+03 V09 4.030e+04 V09 2550.6 V09 3.99708e+03 V17 3.240e+04 V17 2347.9 V17 3.94023e+03 V43 2.629e+04 V43 1836.2 V43 3.00401e+03 V39 2.247e+04 V19 1661.9 V29 2.36570e+03 V35 2.053e+04 V39 1458.9 V19 2.17108e+03 V19 2.046e+04 V13 1366.5 V35 2.09400e+03 V13 2.028e+04 V35 1279.0 V39 1.87708e+03 V29 1.653e+04 V29 1047.2 V23 1.46777e+03 V23 8.807e+03 V27 934.67 V25 1.04609e+03 V25 4.363e+03 V25 875.35 V27 9.36444e+02 V27 3.758e+03 V23 620.02 V13 4.55201e+02 V45 2.308e+03 V37 479.43 V37 3.33242e+02 V37 1.313e+03 V11 238.82 V45 3.20943e+02 V11 5.185e+02 V21 156.71 V21 2.36794e+01 V21 2.241e+01 V45 149.06 V15 1.96517e+01 V15 1.413e+01 V15 52.460 V11 1.39940e+01 V31 5.269e-01 V31 36.740 V31 1.83767e+00 V41 3.150e-01 V41 23.296 V41 1.18577e+00 One can see that the same set of vertical steerings given the main contribution into all three e ects. All they are located in the positions with big pretzel amplitude. The use of it must be avoided.

Conclusion
The known nonlinear eld components of CESR elements have been included into the model structure. Using program MAD the dependence of betatron tunes versus beam orbit displacement around ring was calculated. It is in good agreement with measurements. It was found that the main sours of this tune shift is the nonlinearity of Mark II quads used in regular structure of CESR arcs and 7

the horizontal separators electric eld nonuniformaty at big horizontal o set. The analysis of tune scan data and its comparison with tracking simulation made with MAD allowed to identify elements resposable for nonlinear coupling resonance excitation near CESR working point at pretzeled orbit. These are vertical steerings located at positions with big pretzel amplitude. Realizing these facts one can plan e ective measures to improve CESR perfomance.

Acknowledgments
I wish to thank David Rice, David Rubin and James Welch for their useful discussions and helpful support.

References
1 Alexander Temnykh, On Nonlinear Field Components of CESR Elements, CBN 96-19, Cornell 1996. 2 The MAD Program, Version 8.1, CERN SL 90-13 AP 3 A. Temnykh and D. Rubin, E ect of Q1-Q2 and REC Vacuum Chamber Welds on Dynamic Aperture, CBN 96-1, Cornell 1996 4 A. B. Temnykh, Observation of Beam-beam E ects on VEPP-4, Third Advanced ICFA Beam Dynamics Workshop on Beam-Beam E ects in Circular Colliders, Akademgorodok, Novosibirsk, 5-111989

8

Tune shift versus pretzel Measured and Simulated with MAD using known nonlinearities.
8000 dfh measured Sept 3 1996 dfv measured Sept 3 1996 dfh simulated by MAD dfv simulated by MAD

6000

4000

2000

df [Hz]

0

-2000

Pr1
-4000 -3000 -2000 -1000 0 1000 2000 3000

Figure 1: Measured and calculated betatron tune shift versus PR1 orbit distortion

9

Horizontal Tune shift versus pretzel simulated by MAD. Separeted effect from Mark II quads and Hseps.
8000 Without Hsep and MARK II quads nonlin Hsep nonlin effect Mark II quads nonlin effect Total effect

6000

4000

df [Hz]

2000

0

Pr1
-2000 -3000 -2000 -1000 0 1000 2000 3000

Figure 2: The contribution of the Mark II quads nonlinearity and horizontal separators eld nonlinearity into horizontal tune shift.

10

Vertical Tune shift versus pretzel simulated by MAD. Separeted effect from Mark II quads and Hseps.
1000

df [Hz]

0

-1000

-2000 Without Hsep and MARK II quads nonlin Hsep nonlin effect Mark II quads nonlin effect Total effect

-3000

Pr1
-4000 -3000 -2000 -1000 0 1000 2000 3000

Figure 3: The contribution of the Mark II quads nonlinearity and horizontal separators eld nonlinearity into vertical tune shift.

11

Figure 4: Real tune plane scan at CESR with atten orbit. Vertical beam size is shown. Line crossing the tune plane from the left bottom to right up is fh , fv + fs = nf0 resonance.

Figure 5: Real tune plane scan at CESR with pretzeled orbit. Vertical beam size is shown. In addition to fh , fv + fs = nf0 resonance one can see the appearance of 3fh , fv = nf0 resonance. 12

Figure 6: The simulation of tune plane scan with atten orbit. Maximum of vertical amplitude during 800 turns is shown. Starting points for tracking are in the beam core.

Figure 7: The simulation of tune plane scan with pretzeled orbit. Maximum of vertical amplitude during 800 turns is shown. Starting points for tracking are in the beam core. 13

Figure 8: The simulation of tune plane scan with pretzeled orbit without vertical steering nonlinearity. Maximum of vertical amplitude during 800 turns is shown. Starting points for tracking are in the beam core.

14


相关文章:
更多相关标签: