Multibunch beam extraction using the strip-line kicker at the KEK Accelerator Test Facility

The International Linear Collider (ILC) damping ring (DR) injection and extraction kickers have a very special role: the bunch spacing 189–480 ns is compressed to 3–9 ns when injected into the DR and then decompressed to 189–480 ns when leaving the DR. The kickers act as a bunch-by-bunch beam manipulator to compress and decompress the bunch spacing into/from the DR. They require a fast rise/fall time (3–9 ns) and a high repetition rate (6–2 MHz). Among the candidate technologies, the multiple strip-line kicker system is the most likely to realize the specifications for the ILC reference design. A beam extraction experiment with a prototype strip-line kicker has been carried out at the KEK Accelerator Test Facility (ATF). The kicker is composed of two units of 60-cm-long strip-line electrodes. The multibunch beam (30 bunches spaced at 5.6 ns) stored in the DR was extracted successfully with a bunch spacing of 308 ns. The measured stability of the kick angle was $3.5\times {10}^{-4}$. Some, but not all, parameters of the tested kicker meet the ILC-DR injection/extraction kicker requirements.

Figure 1

Schematic layout of the strip-line kicker for the ILC DR: The kicker consists of 15 or more strip-line units. The length of the strip line is limited to realize the rise/fall time and the kick angle of each unit is small. Multiunit is required to realize the specified kick angle.

Figure 2

(a) The waveform of the kicker pulse (positive polarity); (b) the estimated kick field as a function of time when the kicker pulse is applied to a 60-cm-long strip line; (c) the expanded peak area of the estimated kick field: There is a continuous change in the kick angle as a function of time, and the gradient does not fall to zero for any more than an instant for a period of less than 200 ps.

Figure 3

Bunch structure in the DR and in the extraction line: The bunch stored in the DR is three trains and each train consists of ten bunches with 5.6 ns bunch spacing. The beam is extracted bunch by bunch from the last bunch of each train with 308 ns spacing. The bunch spacing changes to 302.4 ns after every third bunch is extracted.

Figure 4

Comparison of the beam extraction orbits (blue lines) in two systems: one with a pulsed magnet (upper) and one with a strip line (lower). The pulsed magnet kicker produces a 4.6 mrad kick angle and 25 mm displacement at the entrance of septum A. The strip-line kicker produces a 3.0 mrad kick angle. The stored orbit moves outside of the DR (local bump orbit, shown as red line), whereupon the strip-line kicker kicks the beam. The auxiliary septum expands the small orbit difference. The extracted beam has an 18 mm displacement and a 1.6 mrad angle at the entrance of septum A.

Figure 5

Time sequence of the beam injection and extraction: (1) the three trains of the beam are injected to the DR; (2) the local bump orbit is excited gradually after the beam is damped; (3) the beam is kicked bunch by bunch at the end of the flattop of the bump orbit; and (4) the bump orbit returns to zero.

Figure 6

Schematic layout of the strip-line kicker system for ATF.

Figure 7

Photograph of the strip-line electrode and diagram of the simulated electric and magnetic field.

Figure 8

Burst pulse waveforms of FPG10-3000KN.

Figure 9

Output timing feedback: the fine delay setting (green) and the output timing (red).

Figure 10

(a) Time delay from the reference timing as a function of the pulse number. (b) Compensated time delay from the reference timing as a function of the pulse number.

Figure 11

(a) Bunch current in the DR: The stored beam consists of three trains and each train consists of ten bunches with 5.6 ns spacing. The vertical scale is $1\mathrm{nC}/\mathrm{V}$. Vpeak is 400 mV and the charge is $0.4\mathrm{nC}/\mathrm{bunch}$. (b) Bunch current at the extraction line: A train consists of 30 bunches and each bunch has 308 ns (302.4 ns for every third pulse) spacing. The vertical scale is $22\mathrm{nC}/\mathrm{V}$. Vpeak is 18 mV and the charge is $0.4\mathrm{nC}/\mathrm{bunch}$.

Figure 12

The kick angle jitter distribution: Fit from the measured horizontal positions using 19 BPMs and the R12 of the design transfer matrix at each location estimates the angle distribution of the kicker.

Figure 13

The measured kick angle difference as a function of the kicker pulse timing.

Figure 14

(a) Kick angle as a function of the pulse number: The pulse-to-pulse variation of the kick angles of successive pulses is around 1%. (b) Kick angle jitter ratio as a function of the pulse number: The jitter ratio for the kick angle was distributed from $1\times {10}^{-3}$ to $4\times {10}^{-3}$.

Figure 15

The vertical beam size measurement with the wire scanner at the downstream of the focus point.