A Transverse Injection Damper at RHIC
A. Drees, M. Brennan, P. Cameron, R. Connolly, R. Michnoff,
C. Montag
Collider-Accelerator Department, Brookhaven National Laboratory, Upton NY 11973
Abstract. During the RHIC Au-Au runs in 2000 and 2001 as well as in the polarized
proton run in 2001 no transverse damping system existed. To overcome residual injection jitter and to maintain transverse beam emittances a transverse injection damping
system is foreseen for the 2002/2003 run. This report describes the realization of the
injection damping system and outlines the future upgrade into a bunch-to-bunch transverse instability damping system, which is expected to be required for increasing bunch
intensities.
INTRODUCTION
Since it is planned that the injection damping system evolves into a transverse
damper in future years, design and outline considerations will focus on this upgrade
as much as possible to avoid duplication of efforts. For a bunch-to-bunch transverse damper a linear amplifier instead of a power supply and fast pulsing switches
will be used. The choices of BPMs to be used for a transverse damper are much
more constrained. It has to perform under rapid-changing conditions such as the
acceleration ramp and with increased bunch intensities. RHIC accelerates from 10
GeV/u to 100 GeV/u for Au and from 23 GeV/u to 250 GeV/u for protons and
allows a /3-squeeze from /3* = 10 m at injection down to /3* = 1m at flattop at
certain interaction regions causing significant changes in local phase advance and
/3-function. Other scenarios as un-squeezing IRs is on the list for future runs as well.
Therefore, different power supply and BPMs are required for injection damping and
bunch-to-bunch feedback. However, the signal conditioning for the chosen BPM,
the trigger and the damper module can be used in either system. The damper
module to be used for the injection damper is based on a VME board designed
for the AGS damping system with a programmable FPGA (Field Programmable
Gate Array) [1]. For the future damper the existing transverse kickers, which are
currently used for a tune measurement system [2] and will serve for the injection
damper, will be replaced by several dedicated 1m stripline modules in the same
area to avoid conflict with the tune measurement system. For the upcoming year,
since only used at injection, the kicker hardware including power supply and fast
CP648, Beam Instrumentation Workshop 2002: Tenth Workshop, edited by G. A. Smith and T. Russo
© 2002 American Institute of Physics 0-7354-0103-9/02/$19.00
409
HV
HV switches
switches will
will be
be shared.
shared.
I
KICKER AND BPM
Each
2m-long stainless
stainless steel
steel striplines
striplines
Each ring
ring has
has two
two kicker
kicker modules with four 2m-long
mounted
1m apart
apart allowing
allowing both,
both, horizontal
horizontaland
andververmounted on
on ceramic
ceramic stand-offs
stand-offs spaced 1m
tical
provide 4m
4m of
ofstripline
striplinekickers.
kickers.
ticalkicks.
kicks. The
The two
two kickers
kickers are
are connected in series to provide
Each stripline
stripline subtends
subtends an
an angle of 70
Each
70°o at an
an aperture
aperture of
of 77 cm.
cm. The
The assembly
assembly isis
designed to
to give
give 50fi
50Ω impedance
impedance when opposing lines are
designed
are driven
driven in
in the
the difference
difference
mode. Figure
Figure 11 shows
shows one
one kicker module in the assembly area.
mode.
area. Each
Each of
of the
the four
four
FIGURE 1. One kicker module on a work bench in the assembly area.
FIGURE 1. One kicker module on a work bench in the assembly area.
planes can be powered independently. So far only pulsed power has been used. The
planes can be powered independently. So far only pulsed power has been used. The
kick pulses are generated by fast FET switches [3] producing an approximately 140
kick pulses are generated by fast FET switches [3] producing an approximately 140
ns long pulse. By centering this pulse on the measured bunch single bunch exns long pulse. By centering this pulse on the measured bunch single bunch excitation is possible with 60 (RHIC design) and even up to 120 bunches (RHIC
citation is possible with 60 (RHIC design) and even up to 120 bunches (RHIC
upgrade) per ring where the bunch spacing is about 110 ns. Only one out of the 60
upgrade) per ring where the bunch spacing is about 110 ns. Only one out of the 60
or 120 bunches respectively is kicked. All switches for all striplines in both rings
or
120 bunches respectively is kicked. All switches for all striplines in both rings
are charged by one 5kV/2A power supply. Most BPMs used in RHIC are realized
are
charged by one 5kV/2A power supply. Most BPMs used in RHIC are realized
by short circuited transmission lines of 23 cm length, with a design impedance of
by
short
transmission
linesThe
of 23
cm length,
with listed
a design
impedance
50Ω, andcircuited
an aperture
of 7 cm [5].
selection
of BPMs
in tables
1 to 4of
50fi,
and
an
aperture
of
7
cm
[5].
The
selection
of
BPMs
listed
in
tables
1 to 4
below is based on devices with analog signals available in the 1002 service building,
below
is
based
on
devices
with
analog
signals
available
in
the
1002
service
building,
close to the 2 o’clock interaction region (IR2).
close to the 2 o'clock interaction region (IR2).
410
II
KICKER LOCATION
As can be seen in table 1 and 3 the actual kicker location discriminates the
vertical plane by more than a factor of 10 and 20 for the f3* = 3ra and /3* = Ira
lattices. Thus, it would be favorable to move the kickers such that during the ramp,
when /3* is changed from 10 m up to 1 m, the kicker efficiency would not drop in the
vertical plane. To do so, the kickers have to be closer to Q3. However, because of
the increasing /3-functions, the kicker aperture (7 cm) becomes a concern. Figure 2
shows the vertical and horizontal beam profiles in terms of a around IR2 for beams
at 100 GeV with a normalized emittance of 40 Trmm mrad. In order to avoid the
kickers being the limiting aperture, they have to be at > 6 <7, allowing them to move
as close as -52 m from the IR. This move by approximately 10 m is actually planned
for the next run and would reduce the vertical /3-function by 25% at injection while
increasing it by a factor of 5 to 10 during the ramp and at storage. The vertical
kicker- strength at injection is affected by 10% only and therefore not a problem.
The required kick strength g for a linear amplifying system can be defined as:
(1)
with / = 78 kHz being the revolution frequency, t being the required damping time
and (3SQ, (3S being the (3- functions at the location of the kicker and at the location
of the BPM respectively. Equation 1 can be used as a conservative estimate for
the required bang-bang kick strength. After a certain number of turns, however,
the bang-bang damper would become counter-productive. Therefore the injection
damper will be limited to a few hundred turns for damping. For linear damping, a
given amplitude should generate a kick Ax' of:
and
(2)
Ax1
The achievable kick angle for a single 3 kV pulse is about 11 //rad at injection
for Au and p [6]. Using the /3-functions at the kicker and the Q3 (Ql) BPMs
(see table 1 and 3) and a damping time of 200 turns, i.e. 2.5 msec, results in an
approximate orbit amplitude of 32 mm vertically (Ql) and 75 mm horizontally
(Q3). The largest observed amplitudes at injection in the straight sections were
< 25 mm in both planes. Therefore, the existing kick strength should be sufficient
to antagonize injection oscillations in both planes.
Ill
SELECTION OF BPMS
The kick angle after one pulse with 3 kV received by an ion going through the
kickers is approximately 10 //rad at injection energy (7 w 10) [6]. The effect of such
a kick translates into a beam offset given by:
411
10
9
8
7
6
5
4
3
2
1
-80
-60
-40
-20
20
40
60
80
distance from IR2 (m)
FIGURE 2. Horizontal (solid line) and vertical beam profile (dotted line) in units of
sigma around IR2. The dashed line indicates the current kicker location.
(4)
where /3Xjy(so) is the /3-function at the location SQ of the kicker and /3Xjy(s) is
the respective horizontal or vertical /3-function and ty the relative phase advance
between location s and SQ. Table 1 and 3 summarize the values of the /3 functions
at the location of the kickers and BPMs for the blue and yellow ring respectively.
Phase advances, relative to 6 o'clock (IR6) for the two rings are listed in table 2
and 4. The actual lattice strongly favors the horizontal plane when resulting beam
displacements are calculated.
The typical phase advance in one turn in RHIC is approximately 0.21/27T to
0.23/27T for both planes, corresponding to about 80°. During the ramp the tune
would vary from 0.2 to 0.25 at most. In general, the tunes are fairly close and
separated by some 0.01 to 0.02 only. Therefore, after one full turn, a local phase
advance of close to 0° or 180° between BPM and kicker is most suitable for the
damper. With this configuration a total shift of close to 90° will be kept. The
neighboring Q3 BPMs offer 0° relative to the new position in the horizontal plane.
The Ql BPMs on the other side of the IR provide an approximate phase advance
of close to 180° relative to the new kicker location in the vertical planes.
412
TABLE 1. Approximate j3 functions at the location of the blue kicker and blue
BPMs around IR2 for various lattices. The DX BPM are closest to the IR,
indicated by the double line.
device
plane
kicker
Q3BPM
Q1BPM
DXBPM
DXBPM
Ql BPM
Q3BPM
Q4BPM
Q5BPM
Q6BPM
Q7BPM
Q8BPM
Q9BPM
Q10 BPM
HV
HV
HV
HV
HV
HV
HV
HV
V
H
HV
HV
V
H
s(m)
2491.7
2519.0
2530.9
2547.6
2564.2
2580.9
2592.8
2629.1
2636.9
2651.7
2668.9
2682.0
2699.4
2710.8
0*1Om
/Mm) A / M
34
16
141
48
73
72
17
17
17
17
73
73
49
140
36
20
8
56
42
15
13
49
45
13
12
43
44
11
f3\ 3m
/Mm)
86
426
211
26
26
211
154
18
8
38
11
46
12
44
/?* Lm
/Mm) Py (m)
118
5
633
231
314
314
37
37
37
37
314
314
230
634
17
38
62
11
37
20
10
47
47
10
12
43
44
11
Py (m)
6
153
210
26
26
210
424
32
59
18
48
11
42
11
TABLE 2. Approximate phase advances relative to IR6 at the location of the blue kicker
and blue BPMs around IR2 for various lattices.The DX BPM are closest to the IR, indicated
by the double line. Values for the new kicker location are in brackets.
device
kicker
Q3BPM
Q1BPM
DXBPM
DXBPM
Q1BPM
Q3BPM
Q4BPM
Q5BPM
Q6BPM
Q7BPM
Q8BPM
Q9BPM
Q10 BPM
plane
HV
HV
HV
HV
HV
HV
HV
HV
V
H
HV
HV
V
H
s(m)
2491.7
2519.0
2530.9
2547.6
2564.2
2580.9
2592.8
2629.1
2636.9
2651.7
2668.9
2682.0
2699.4
2710.8
/?*!Om
Ar/27r
My/2?r
18.52 (18.55)
18.97 (19.13)
18.58
19.24
18.60
19.26
18.68
19.34
18.90
19.56
18.98
19.64
19.00
19.34
19.39
19.71
19.82
19.91
20.06
20.14
19.66
19.78
19.83
19.90
20.02
20.11
20.23
20.32
413
/?* 3m
p* Lm
/W27T
My/2?r
Ar/27T
^ y /27T
18.55
18.58
18.58
18.62
19.01
19.05
19.05
19.43
19.51
19.71
19.85
19.95
20.09
20.18
19.05
19.25
19.26
19.29
19.68
19.72
18.53
18.55
18.55
18.58
19.00
19.03
19.03
19.43
19.51
19.67
19.82
19.92
20.06
20.15
19.05
19.21
19.21
19.24
19.66
19.69
19.69
19.73
19.76
19.82
19.92
20.03
20.17
20.26
19.73
19.78
19.81
19.87
19.98
20.08
20.21
20.30
TABLE 3. Approximate j3 functions at the location of the yellow kicker and
BPMs around IR2 for various lattices. The order is defined by the direction of
the beam. The DX BPM are closest to the IR, indicated by the double line.
device
Q10 BPM
Q9BPM
Q8BPM
Q7BPM
Q6BPM
Q5BPM
Q4BPM
kicker
Q3BPM
Q1BPM
DXBPM
DXBPM
Ql BPM
Q3BPM
plane
V
H
HV
HV
V
H
HV
HV
HV
HV
HV
HV
HV
HV
s(m)
2710.8
2699.4
2682.0
2668.9
2651.7
2636.9
2629.1
2620.1
2592.8
2580.9
2564.2
2547.6
2530.9
2519.0
0*1Om
/Mm) A / M
12
43
11
44
13
45
49
13
15
42
56
8
20
36
34
16
141
48
73
72
17
17
17
17
73
73
49
140
f3\ 3m
/Mm)
12
44
11
47
18
59
32
85
426
211
26
26
211
154
Py (m)
42
11
46
11
38
8
18
6
153
210
26
26
210
424
p Lm
/Mm) Py (m)
12
42
44
11
10
47
47
10
20
37
62
11
38
17
118
5
633
231
314
314
37
37
37
37
314
314
230
634
TABLE 4. Approximate phase advances relative to IR6 at the location of the yellow
kicker and BPMs around IR2 for various lattices. The order is reverted in s to reflect
the direction of the beam. The DX BPM are closest to the IR, indicated by the double
line. Values for the new kicker location are in brackets.
device
plane
/?*!<3m
s(m)
Q10 BPM
Q9BPM
Q8BPM
Q7BPM
Q6BPM
Q5BPM
Q4BPM
kicker
Q3BPM
Q1BPM
DXBPM
DXBPM
Q1BPM
Q3BPM
V
H
HV
HV
V
H
HV
HV
HV
HV
HV
HV
HV
HV
2710.8
2699.4
2682.0
2668.9
2651.7
2636.9
2629.1
2620.1
2592.8
2580.9
2564.2
2547.6
2530.9
2519.0
8.54
8.62
8.74
8.83
8.95
9.02
9.07
9.12 (9.16)
9.18
9.21
9.29
9.51
9.59
9.61
F 3m
My/2?r
|W27T
8.37
8.49
8.60
8.70
8.81
9.13
9.18
9.25 (9.4)
9.51
9.54
9.62
9.84
9.92
9.94
414
iW27r
8.57
8.65
8.78
8.88
8.99
9.05
9.09
9.12
9.14
9.15
9.18
9.57
9.61
9.62
F
Lm
jU y /27T
M*/27T
My /27T
8.33
8.41
8.56
8.66
8.80
9.00
9.08
9.26
9.46
9.46
9.50
9.89
9.92
9.93
8.60
8.68
8.82
8.93
9.03
9.09
9.12
9.14
9.16
9.16
9.19
9.61
9.64
9.64
8.36
8.45
8.60
8.70
8.85
9.01
9.09
9.33
9.49
9.49
9.52
9.94
9.97
9.97
IV
TRIGGER AND DATA ACQUISITION
Figure 3 sketches the signal processing and triggering of both, the BPMs and the
kickers, for a bang-bang injection damping system. The damper module is based
on the existing AGS module and needs adjustments for the RHIC damper in the
I/O area.
RHIC Injection Damping System
Hi Volt. Switches
Trigger
HV out -
HVout
Trigger
HVin
HV out +
FIGURE 3. Block diagram for the RHIC injection damper.
The V124 module [7] receives and decodes the beam synchronous event link [8].
The raw data acquisition from the two BPM planes will be triggered by two channels
of the V124 board where a total of 8 channels is available. Each channel for
BPM readout and the kicker trigger has the appropriate delay so, on turn-by-turn
acquisition, the same bunch will be observed on the BPM and then kicked. Start
turn number, total number of turns for acquisition and damping as well as time
delays are all parameters which can be remotely set from a console level computer.
In general, the V124 allows the system to be triggered by any event broadcasted
on the beam synchronous link such as the injection-event, start-acceleration-event
or on demand. However, to damp injection oscillations only the injection-event is
going to be used.
The raw (bipolar) signal from the BPM will be attenuated by a programmable
attenuator of 0-30 db. In the signal conditioning version as used by the ARTUS
tune meter two signal processing modules compute the two difference signals and
the sum of all stripline signals for each ring. We are currently working on an
415
upgrade of the signal conditioning, which will then be used by both systems, the
tune meter and the transverse injection damper. It will also be used for signal
conditioning in the future bunch-to-bunch transverse damper.
v CONCLUSION
The existing kicker modules, once moved by approximately 10 m, are suitable
to act as a transverse injection damper. A basic design of a signal conditioning
exists and is currently in use for the ARTUS tune meter system. However, it is
planned to upgrade this version for both, the damper and the tune meter, for future
runs starting in FY'03. The actual damping module is based on the existing AGS
damper module and needs some modifications. This module, together with the
signal conditioning design, will be kept for the future bunch-to-bunch transverse
damper. Relative to the new kicker location, the existing Q3 and Ql BPMs provide
a suitable phase advance after one turn of about 90° and 270° respectively. The
high /3-function at the BPMs of 140 m and 71 m eases an amplitude measurement
with good signal to noise ratio in both planes. For the future feedback system,
the injection damping kickers will be replaced by several dedicated 1m stripline
modules in the same area to avoid conflict with the tune measurement system.
REFERENCES
1. Michelle Wilinski et al., "Enhancements to the Digital Transverse Dampers at the
Brookhaven AGS", these proceedings.
2. A. Drees, R. Michnoff,M. Brennan, J. DeLong, "ARTUS: The Tune Measurement
System at RHIC" proceedings BIW2000, Boston, 2000.
3. Behlke Electronic GmbH, http://www.euretek.com/
4. J. Xu et al., "The Transverse Damper System for RHIC", Proceedings of the Particle
Accelerator Conference (PAC) in San Francisco, 1991.
5. P. Cameron et al., "RHIC Beam Position Monitor Assemblies", IEEE Proc., 1995
PAC.
6. P. Cameron, R. Connolly, A. Drees, W. Ryan, H. Schmickler, T. Shea, D. Trbojevic,
"ARTUS: A Rhic TUne Measurement System", RHIC/AP/98-125, internal note.
7. H. Hartmann, T. Kerner, "RHIC beam synchronous trigger module", Proc. PAC 1999
(p. 696).
8. T. Kerner, C. R. Conkling Jr., B. Oerter, "V123 Beam Synchronous Encoder Module",
Proc. PAC 1999 (p. 699).
416