Low-Level RF System Design for the Accelerator Test Facility (ATF) Damping Ring
M. Minty
LOW-LEVEL RF SYSTEM DESIGN FOR THE ACCELERATOR TEST FACILITY (ATF) DAMPING RING
M.G. Minty, SLAC, Stanford, CA 94309, USA K. Kubo, F. Hinode, S. Sakanaka, J. Urakawa, KEK, Oho, Tsukuba-shi, Ibaraki-ken, 305 Japan
Abstract
The ATF damping ring [1] was built to demonstrate the production of low emittance, high current beams for future linear colliders. To attain high beam currents, multiple high current bunch trains are required. The low-level rf system should be designed to minimize both steady-state and tran- sient beam loading effects in the accelerating cavities. In addition the design should be sufficiently flexible to allow for a variety of beam dynamics tests which require a wide range of beam currents and cavity voltages. The low-level rf system and stability boundaries for reduced power and full power operation are discussed in this paper.
1 INTRODUCTION
Control of the longitudinal beam parameters is just one as- pect of many exciting studies to be performed using the ATF damping ring. These include the use of damped cavities [2] for suppression of longitudinal coupled-bunch modes, the use of a sub-rf cavity to compensate for intra- train synchronous phase offsets [3], and beam-loading ef- fect minimization during normal operation using a single- turn beam injection/extraction scheme [1]. Many of the studies planned involve the use of a wide range of beam currents and bunch lengths (i.e. cavity voltages).
2 STABILITY BOUNDARIES
Table I shows the operating conditions at the design en- ergy of 1.54 GeV with a full 714 MHz (harmonic number
) rf system (250 kW klystron [1], 4 cavities) for different numbers of bunch trains. The cavity coupling pa-
rameter [2] corresponds to optimum coupling at full current neglecting higher order mode losses. The vari- ables listed are: the dc beam current , the beam energy
, the accelerating voltage , the radiation loss per turn per electron , the higher order mode loss per turn [4] , the synchronous phase1 , the synchrotron frequency , the longitudinal damping time , the natural energy spread
, the bunch length , the momentum compaction , the Robinson damping time , the total shunt impedance2
, the quality factor , the cavity fill time (without direct
feedback) , the cavity tuning angle (for minimum re- flected power) , the overvoltage3 , the rf bucket height
, the average klystron power , average dissipated
power , the average beam power , and the average reflected power .
Work supported by the Department of Energy, contract DE-AC03-
76SF00515
1
2
3
Table 1: RF parameters for design operation with 1 to 5, full current bunch trains.
Plots for various operating currents at different number of particles per bunch , number of bunches per train
, and number of trains are shown in Fig. 1 for the case of a single (top) and five (bottom) bunch trains. The solid curves are contours of constant total dc current. The expected threshold for transient bunch lengthening[1] is shown at assuming a 5 mm bunch length. The vertical lines indicate constraints imposed by the larger of the injection or extraction kicker rise and/or fall times ( ):
where is the bunch-to-bunch spacing. Assuming
ns, are shown in Fig.1 (b) a solid vertical line ( =1.4 ns), and a dashed vertical line ( =2.8 ns) . The solid verti- cal line in Fig.1 (a) corresponds to a maximum kicker flat- top time of 180 ns with ns.
The parameter space [5] for full current operation at 1.54
GeV is shown in Fig. 2. The horizontal axis is the tun-
ing angle , which is a measure of how far off resonance the cavity is being driven. The open circle designates the design operating point which lies along the line of zero
Ib(A) = 0.1
4
0.2
0.3
0.4
0.5
0.6
current limit (crosses below the power limit). For exper- iments requiring both high beam currents and low cavity voltages, direct feedback [6] will be required.
3
1.0
2 0.8
E = 1.30 GeV
1
(a)
0
0.6
0.4
0.2
3
0.8
E = 1.54 GeV
2
1 (b)
0
0 20 40 60
Nbpt
4-97
8297A2
0.6
0.4
0.2
0.0
0.1 0.2 0.3 0.4 0.5 0.6
Figure 1: Map of possible fill patterns for the ATF damping ring with a single train (a) and with the design fill of five bunch trains (b).
2.0
1.5
1.0
E = 1.54 GeV
loading angle ( ) for minimum reflected power. The shaded region shows a region of instability due to Robin- son' s high current limit. The region indicated by hatches is accessible as limited by the available klystron output power. In practice, the hatched region may be somewhat
0.5
0.0
0.5
1
Vc = (MV)
1.5
2
4-97
8297A4
reduced, particularly at high currents, if transient loading
in the accelerating cavities is not minimized.
Figure 3: Beam current limits. Plotted is the beam current versus the cavity voltage .
1.50
1.25
1.00
0.75
0.50
= 0 deg
Pgmax = 250 kW Vc = 1.0 MV
s = 78.9 deg
3 LOW-LEVEL RF SYSTEM
The design of the low level rf control system aims towards
These include compensation for radiation and higher order mode losses, provision of sufficient cavity voltage to en- sure an energy acceptance of 1 , regulation of the cavity
0.25
0.00
voltage and beam phase under steady-state operating con- ditions, and minimization of adverse effects arising from transient beam loading at injection. These requirements
-80
-60
-40
z [deg]
-20 0
4-97
8297A3
should be fulfilled while minimizing the required source power and the power reflected from the cavities.
A block diagram for the low level control system is
Figure 2: Parameter space for full current operation. Plot-
ted is the beam current ( ) as a function of tuning angle .
Steady-state limitations to the rf beam current (twice the dc beam current) are shown as a function of cavity voltage in Fig. 3. The top two plots are for initial commissioning at reduced power (45 kW) with 2 cavities and a radiative loss per turn of 79 keV at 1.30 GeV and 156 keV at 1.54
GeV. The bottom plot assumes 225 kW available klystron
power, 4 accelerating cavities, and a 1.54 GeV beam en- ergy. Shown for zero loading angle are two limits: the max- imum klystron output power (circles) and Robinson' s high
shown in Fig. 4. In the full rf system, the output of a single
714 MHz klystron is used to power 4 cavities. Conven- tional isolators are used after the klystron output power has
been divided by two. A 1428 MHz master oscillator pro- vides the phase reference for the S-band linac, the damp-
ing ring, and the extraction line bunch compressor klystron. The phase of the beam at injection and extraction is varied
using phase shifters upstream of the feedback loops. Us- ing the single-turn injection and extraction scheme [1] the injection and extraction phases may not be independently controlled. In this design, the damping ring rf phase is adjusted for optimum phase at injection; the phase at ex- traction is therefore fixed. To ensure proper phase in the
bunch compressor, the phase of the compressor klystron is adjusted via feedback using a measurement of the beam phase from the damping ring. Conventional feedback loops are used to regulate against changes in the cavity voltage and beam phase. Direct feedback [6] is included to facili- tate experiments at low cavity voltage.
avoided using either direct feedback or, for better regula- tion of the cavity voltage and beam phase, by changing the rf phase (in this case by from Eq. (3) at injec- tion of a bunch train. The latter option is particularly useful for maintaining a high duty cycle and is described further in Ref. [7].
To prelinac
& injector
x2
2
FINE
TUNE
Klystron phase regulatio
250 kW Klystron
Isolators
Re Interlocks
Phase lock to linac
REF
INJ
Direct FB amplitude boost
Limiter
Magic Tee
DRBPM
COARSE EXT
To compressor
VTUNE
VREF
Peak detector
Klystron amplitude regulation
Direct
Tuner control for 4 cavities
Direct feedback
Cavity amplitude feedback
Vector sum
Cavity phase feedback
Acknowledgements
Peak detector
4-97
8297A1
Figure 4: Block diagram of low level rf system.
4 RAMP TO FULL CURRENT
Due to the proximity of the design operating current to sta- bility boundaries (see Fig. 2), care must be exercised with injection of each bunch train. A suggested injection scheme involves detuning the cavities using the tuner feedback set- points prior to injection of each train such that after the train has been injected, the average loading angle is zero. The required tuning angle for a train of current is
where is the total loaded impedance. But since the tuning feedback loop measures the loading angle (not the tuning angle), the tuner setpoint required at train is
Note that conventional current ramping with a fixed tuner setpoint would result in beam loss at injection of the final bunch train due to the beam loading limit.
Numerical simulations of the complete rf system have shown that transient loading of the rf system may lead to beam loss at the highest operating currents. This may be
We gratefully acknowledge E. Paterson, G. Loew, and S. Takeda for their support during for the course of these studies. In addition we thank R. Siemann for insightful discussions.
5 REFERENCES
[1] F. Hinode , `ATF Accelerator Test Facility Design and
Study Report' , KEK Internal 95-4 (June, 1995).
[2] S. Sakanaka, K. Kubo, T. Higo, `Design of a HOM Damped Cavity for the ATF Damping Ring' , Proc. 1993 IEEE Part. Accel. Conf., Washington, DC (1993) 1027; S. Sakanaka
, `Low-power Measurement on a HOM Damped Cavity for the ATF Damping Ring' , Proc. 1994 Intl. Linac Conf., Tsukuba, Japan (1994) p. 281, KEK Internal 94-79 (August
1994); S. Sakanaka , `Design of a High-power Test
Cavity for the ATF Damping Ring' , Proc. 1995 Part. Acc. Conf., Dallas, TX (1995) p. 1788.
[3] K. Kubo, T. Higo, T. Higo, `Compensation of Bunch Posi- tion Shift Using Sub-RF CAvity in a Damping Ring' , 1993
IEEE Part. Accel. Conf., Washington, DC (1993) p. 3503. [4] K. Kubo, private communication.
[5] See, for example, M. Minty and R. Siemann, "Heavy Beam Loading in Storage Ring Radio Frequency Systems", Nucl. Instr. and Meth. A, 376 (1996) 301-318.
[6] F. Pedersen, IEEE Trans. on Nucl. Sci., NS-22, no.3 (1975), and NS-32, no. 3 (1985).
[7] M.G. Minty, `Low-level RF System Design for the Next
Linear Collider Damping Rings' , these proceedings.
Por: Tirso Ramírez C.I.: 18392099
CAF
No hay comentarios:
Publicar un comentario