Linac automated beam phase control system

Abstract

Adjustment of the rf phase in a linear accelerator is crucial for maintaining optimal performance. If phasing is incorrect, the beam will in general have an energy error and increased energy spread. While an energy error can be readily detected and corrected using position readings from beam position monitors at dispersion locations, this is not helpful for correcting energy spread in a system with many possible phase errors. Uncorrected energy spread results in poor capture efficiency in downstream accelerators, such as the Advanced Photon Source's (APS's) particle accumulator ring (PAR) or booster synchrotron. To address this issue, APS has implemented beam-to-rf phase detectors in the linac, along with software for automatic correction of phase errors. We discuss the design, implementation, and performance of these detectors and how they improved APS top-up operations.

Full text

LINAC AUTOMATED BEAM PHASE CONTROL SYSTEM
S. J. Pasky, R. M. Lill, M. D. Borland,
N. S. Sereno and L. L. Erwin
Argonne National Laboratory, Argonne, IL 60439, U.S.A.
Abstract
Adjustment of the rf phase in a linear accelerator is
crucial for maintaining optimal performance. If phasing
is incorrect, the beam will in general have an energy error
and increased energy spread. While an energy error can
be readily detected and corrected using position readings
from beam position monitors at dispersion locations, this
is not helpful for correcting energy spread in a system
with many possible phase errors. Uncorrected energy
spread results in poor capture efficiency in downstream
accelerators, such as the Advanced Photon Source’s
(APS’s) particle accumulator ring (PAR) or booster
synchrotron. To address this issue, APS has implemented
beam-to-rf phase detectors in the linac, along with
software for automatic correction of phase errors. We
discuss the design, implementation, and performance of
these detectors and how they improved APS top-up
operations.
INTRODUCTION
The Advanced Photon Source (APS), at Argonne
National Laboratory is a high-brightness, third-generation
synchrotron light source that operates in top-up mode
75% of the time to maintain a storage ring current of
102 mA to 1% tolerance. The APS operating availability
reached greater than 98% last year. The excellent machine
performance is due to many hardware and software
improvements including software automation of machine
operations. This paper will describe one of many
automated tools used by APS specifically for our new
linac beam phase control system. This system measures
and corrects the phase of the beam relative to each rf
system that provides acceleration in the particle
accelerator. We will also cover the beam position
monitors (BPMs), related electronics, and how the bpm
information is interfaced with our automated software to
maintain hands- free injector beam phasing.
APS LINAC LAYOUT AND PHASE
DETECTOR LOCATIONS
Linear accelerators consist of a linear sequence of many
accelerating structures where accelerating fields are
generated and timed such that particles accumulate energy
from each accelerating structure. The APS linear
accelerator, or linac, was designed with five accelerating
sectors known as Linac One through Linac Five or (L1 -
L5). Three of the five sectors are SLEDed [1] sectors—
L2, L4, and L5 support four accelerating structures each
for particle acceleration up to ~ 450 MeV. L1 and L3 have
the capability of driving one of our two thermionic rf guns
[2].
Phase detectors for the phase detection system are
configured so that the beam phase is actually measured by
a BPM upstream of the accelerating structure as shown in
Figure 1. This configuration ensures that the ability to
measure and correct phase errors is not dependent on
having beam transport through the linac structures that are
being phased. Hence, we can measure and correct the
beam-to-rf phase for a set of structures as soon as beam
arrives at the entrance to the first structure. The only
exception to this configuration is in the case of the rf gun
phase detector.
The first phase detector measures the phase of the beam
using linac BPM L1:P1 relative to the RG1 or RG2 rf and
the L2 rf measured at the first L2 accelerating structure.
The L4 and L5 phase detectors operate using the first
accelerating structures in L4 and L5 rf waveforms and
BPMs L3:P3 and L4:P1, respectively.
Figure 1: Linac layout.
BEAM-PHASE CONTROL OPERATION
Operation of the linac beam-phase control is
mathematically very similar to our linac trajectory
controls. For the case of the linac trajectory, the horizontal
or vertical beam position at each BPM is held fixed by
changing correctors along the linac depending on the
deviation of the measured BPM position from the desired
BPM position. An analogous longitudinal beam
“trajectory” is given by the output of each phase detector
along the linac. The linac beam-phase control acts to keep
the phase deviation at each detector fixed by changing the
linac sector phases or, in the case of the RFGun phase
detector, the RFGun power, depending on the deviation of
the measured phase at each detector from crest.
_______
* Work supported by U.S. Department of Energy, Office of Sciences,
Office of Basic Energy Sciences, under Contract No. W-31-109-ENG-38.
Proceedings of LINAC 2006, Knoxville, Tennessee USA TUP001
Applications
Beam Diagnostics and Instrumentation
241

The operation of the RFGun phase detector is worth
special mention here. In the case of the L2, L4, and L5
phase detectors, the phase detector output depends on the
position of the phase setpoint of the klystron powering
these sectors. This is apparent since the phase of each
linac sector waveform is independent of all the other
sectors and the RFGun. Since the beam is created at the
RFGun, phase is defined by the RFGun rf waveform, and
hence the time the beam arrives at the BPM is the same
for all L1 klystron phase setpoints. Changing the L1
klystron phase setpoint will therefore have no effect on
the output of the RFGun phase detector. The only way to
change the RFGun phase detector output is to change the
power of the klystron powering L1 or the current in the
alpha magnet between the gun and the phase detector.
Changing the L1 klystron power results in a different
RFGun beam energy and hence a different time-of-flight
from the RFGun to L1:P1. This time-of-flight difference
is directly related to the RFGun phase detector output.
The linac beam-phase control therefore uses the L1
klystron power to keep the phase detector output constant,
which has the beneficial effect of keeping the beam
energy out of the RFGun constant.
Initial setup of the system requires manual phasing and
setting of RFGun power to establish the desired operating
points for each detector. Once this is completed, a
PEMtool [3] application (Figure 2) is used to transfer the
phase detector readbacks to the feedback setpoints.
Figure 2: Phase detector transfer phase feedback readings
to setpoints tool.
Transferring the readbacks to the setpoints zeros the
errors. With the errors zeroed, the control program then
acts to keep phase errors zeroed and the beam’s
longitudinal trajectory fixed. The control-loop software
application is shown in Figure 3 and is similar to others
used in the injector. From this application, the operator
can start, resume, suspend, or abort the control-loop.
Figure 4, left display, shows a linac beam-phase control-
loop runcontrol screen with an adjacent display of
predetermined process variables (PVs) that have
minimum and maximum operating values defined. When
the entire list of test PVs are within their operating range,
the control-loop will function.
Figure 3: Linac phase control application.
Figure 4: Linac phase control runcontrol screens and test
condition screen.
LINAC BPM ELECTRONICS
We have just covered how the beam phase control
system software works; we now discuss some of the
hardware.
Signals from linac beam position monitors are carried
over 0.141-inch hard-line cable to a feedthrough bracket.
There they connect to ¼-inch helix cable for the rest of
the journey from the linac tunnel to electronics located in
the klystron gallery. The phase detection system is
basically partitioned into subsystems, as shown in Figure
5 [4]. The stripline detector and bandpass filters are
external to the phase detector receiver. The receiver
comprises a summing network, phase detector, and
control and regulation boards.
Figure 5: High-power rf-to-beam phase detector
electronics.
High-Power RF-to-Beam Phase Detector
Electronics
The summing network front-end board combines four
signals from the BPM stripline. The stripline signal blades
are combined to minimize position dependence. This is
accomplished by three Wilkinson 2-way power combiners
that are printed on the Rogers RO3006 microwave board
TUP001 Proceedings of LINAC 2006, Knoxville, Tennessee USA
242 Applications
Beam Diagnostics and Instrumentation

substrate. The dielectric constant of the ceramic PTFE
composite substrate is 6.15 and the loss tangent is 0.0025
@ 10 GHz. One of the design goals was to keep the rf
board construction process as simple as possible by
avoiding bonding substrates. This equates to a two-layer
board with a thickness of 0.025 inch to insure the trace
width of 0.036 inch for 50-Ω lines.
The summing network board also provides the gain and
self-test capabilities for the system. The two signals are
sampled via 15-dB directional couplers, which are printed
on the circuit board. This provides the ability to trouble-
shoot the system without disconnecting any cables. The
directional couplers also serve as feeds for the self-test
oscillators. In the self-test mode the coupler is switched
from a 50-Ω termination to a voltage-controlled oscillator.
The oscillator drives a two-way equal power divider that
is also printed on the circuit board and provides equal
inputs to the phase detector. The board employs a
selectable gain stage to shift the operating range by 20 dB.
This amplification will shift the input operating range
while maintaining the same system gain. This feature will
be used in some applications and has the effect of
extending the dynamic range.
The phase detector board also uses the Rogers 3006
ceramic PTFE composite substrate. The input signals are
fed into matching networks and then fed into the Analog
Devices AD8302. The AD8302, shown in Figure 6,
integrates two closely matched wideband logarithmic
amplifiers, a wideband linear multiplier/phase detector,
precision 1.8-V reference, and analog output scaling
circuits. The gain and phase video output signals are then
filtered and scaled to ± 1.0 into 50 Ω. The signals are then
fed into the digitizer.
The control and regulator board is constructed on
standard FR-4 board and provides conditioned input
power and housekeeping for the system.
The boards are housed in an EMI-shielded aluminum
case. The receivers are installed in a 19-inch-wide, 4-U-
height card crate where up to eight receivers can be
installed
Figure 6: AD8302 block diagram.
ENGINEERING TOOLS
Acquired data from the beam phase control system
receiver card is analyzed by a data acquisition and digital
I/O card and then graphically displayed. This graphical
display is known as the engineering phase detector
calibration screen, shown in Figure 7. Here phase and
amplitude waveforms are provided along with raw and
conditioned values of voltage and phase including values
smoothed for use by the phase control loop.
Figure 7: Phase detector calibration screen
CONCLUSION
Initial measurements and tests of the linac beam-to-rf
phase detector system were found to have good
performance over a wide range of linac operating
conditions. Once the beam phase control-law was
interfaced with the hardware, beam injection efficiency
dramatically improved for top-up operation, and manual
phasing is no longer needed by operations personnel. I
should note that additional changes will be made in the
near future to use a single feedback matrix with the
control loop instead of three separate matrices with
individual control loops. This should improve injector
phase response when power fluctuations of L1 occur.
ACKNOWLEDGMENTS
The authors would like to thank the following people
for their assistance: A. Brill, C. Gold, and R. Zabel for
hardware installation, E. Norum for software
development, and M. Varotto for MEDM screen
development.
REFERENCES
[1] Z.D. Farkas et al., “SLED: A Method of Doubling
SLAC’s Energy, Proc. 9th Int. Conf. on High Energy
Accelerators, SLAC, Stanford, California, May 2-7,
1974, pp. 1-9 (1974).
[2] J.W. Lewellen, S. Biedron, A. Lumpkin, S.V. Milton,
A. Nassiri, S. Pasky, G. Travish, M. White,
“Operation of the APS RF Gun,” XIX International
Linac Conference, Chicago, Illinois, ANL-98/28, pp.
863-865 (1999).
[3] M. Borland, “The Procedure Execution Manager and
its Application to Advanced Photon Source
Operation,” 1997 Particle Accelerator Conference,
Dallas, TX, pp. 2410-2412 (1998).
[4] R. Lill, “New Beam Position Monitor System Design
for the APS Injector,” Proceedings of the Beam
Instrumentation Workshop 2002: Tenth Workshop,
AIP Conference Proceedings 648, pp. 401-408 (2002).
Proceedings of LINAC 2006, Knoxville, Tennessee USA TUP001
Applications
Beam Diagnostics and Instrumentation
243