Proceedings DIPAC 2003 – Mainz, Germany
DIAGNOSTIC CHALLENGES AT SNS*
M.A. Plum, Los Alamos National Laboratory, Los Alamos, NM, USA;
T. Shea and S. Assadi, Oak Ridge National Laboratory, Oak Ridge, TN, USA;
L. Doolittle, Lawrence Berkeley National Laboratory, Berkeley, CA, USA;
P. Cameron and R. Connolly, Brookhaven National Laboratory, Upton, NY, USA
Abstract
The Spallation Neutron Source now being built in Oak
Ridge, Tennessee, USA, accelerates an H¯ ion beam to
1000 MeV with an average power of 1.4 MW. The H¯
beam is then stripped to H+, compressed in a storage ring
to a pulse length of 695 ns, and then directed onto a
liquid-mercury neutron spallation target. Most of the
acceleration in the linac is accomplished with
superconducting rf cavities. The presence of these
cavities, the high average beam power, and the potential
for the e-p instability in the storage ring, provide unique
challenges to the beam diagnostics systems. In this talk
we will discuss these challenges and some of our
solutions, including the laser profile monitor system, the
residual gas ionization profile monitors, and network
attached devices. Measurements performed using
prototype instrumentation will also be presented.
INTRODUCTION
Most of the beam diagnostics [1] at the Spallation
Neutron Source (SNS) are fairly standard: beam position
monitors, beam loss monitors, wire scanners, beam
current monitors, slit and collector emittance stations,
Faraday Cups, etc. However, there are several aspects
about the SNS that create some special challenges, such
as the superconducting rf cavities, the high beam power,
and the potential for the e-p instability in the ring. The
high beam power and the superconducting rf cavity
challenges have led to the development of a laser profile
monitor system that replaces the originally-envisioned
wire scanner system in the superconducting linac (SCL).
The challenges associated with the e-p instability and the
expected beam loss in the ring have led to improvements
in the gas ionization profile monitor design. We have also
taken advantage of technology developments by basing
many of our diagnostics instrumentation designs on the
personal computer (PC) platform. A layout of the various
diagnostics systems is shown in Fig. 1.
LASER PROFILE MONITOR SYSTEM
The profile monitor system for the SCL was originally
envisioned to be a carbon wire scanner system. However,
linac designers were concerned about the possibility that
carbon wire ablation, or broken wire fragments, could
find their way into the superconducting cavities and cause
them to fail. A high reliability wire scanner actuator was
Operational
Operational
Next
Next 66 Months
Months
IDump
IDump
11 Position
Position
11 Wire
Wire
11 Current
Current
FY2004
FY2004
FY2005
FY2005
D-Plate
D-Plate
33 Position
Position 11 Loss
Loss 11 Current
Current
22 Wire
Wire 22 Faraday
Faraday Cup
Cup 11 Bunch
Bunch
11 Video
33 Neutron
Video 11 Halo
Halo
Neutron
11 Beam
Beam Stop
Stop Faraday
Faraday Cup
Cup
11 Emittance
Emittance (Slit
(Slit and
and Collector)
Collector)
MEBT
MEBT
66 Position
Position
22 Current
Current
55 Wires
Wires
11 Emittance
Emittance
33 Neutron
Neutron
DTL
DTL
10
10 Position
Position 55 Wire
Wire
55 Loss
Loss 55 Faraday
Faraday Cup
Cup
66 Current
Current 12
12 Neutron
Neutron
RING
RING
44
44 Position
Position 22 Ionization
Ionization Profile
Profile
87
11 Current
87 Loss
Loss
Current
55 Electron
Electron Det.
Det.
22 Wire
11 Beam
Wire
Beam in
in Gap
Gap
22 Video
11 Tune
Video
Tune
CCL
CCL
10
10 Position
Position 99 Wire
Wire
88 Neutron
Neutron
48
48 Loss
Loss 33 Bunch
Bunch
11 Faraday
Faraday Cup
Cup 11 Current
Current
SCL
SCL
32
32 Position
Position 58
58 Loss
Loss
88 Laser
Laser Wire
Wire 77 Neutron
Neutron
CCL/SCL
CCL/SCL Transition
Transition
22 Position
Position 11 Wire
Wire
11 Loss
Loss 11 Current
Current
EDump
EDump
11 Current
Current 22 Loss
Loss
11 Wire
Wire
RTBT
RTBT
17
17 Position
Position
43
43 Loss
Loss
44 Current
Current
55 Wire
Wire
11 Harp
Harp
HEBT
HEBT
29
29 Position
Position 11
11 Wire
Wire
49
49 Loss
Loss
44 Current
Current
LDump
LDump
66 Loss
Loss
66 Position
Position
11 Wire
Wire
Fig. 1. (color) Layout of the diagnostics in the SNS facility, color-coded to indicate the staged installation dates.
___________________________________________
*Work supported by the Office of Science of the US Department of
Energy.
Invited Talks
IT08
35
Proceedings DIPAC 2003 – Mainz, Germany
Beam Box Mount
• Scan actuator
• Lens actuator
• Mirror flipper
Toroid 2
Peak detector
Electrode
laser
Laser Monitor
Magnet
Optical Table
• Laser profile
• Laser attenuation
Integrator
Toroid 1
Fig. 2. (color) Schematic layout of a laser beam profile monitor.
developed [2] at Los Alamos National Laboratory
(LANL) in tandem with experiments [3,4] using a laser to
measure profiles of H¯ beams at Brookhaven National
Laboratory (BNL).
Once the laser profile monitor concept was proven by
experiments at BNL, and subsequently on the SNS MEBT
at Lawrence Berkeley National Laboratory, the decision
was made to replace the carbon wire scanner system with
the laser profile measurement system in the SCL. The
Fig. 3. (color) Horizontal beam profile in the SNS
MEBT, measured January 2003. Top: an example of the
electron collector signal. Bottom: the results of the
measurement, with a Gaussian fit plotted out to 2.5 σ.
36
advantages that the laser profile monitor system has over
the wire scanner system are: 1) profiles can be measured
during normal operations, as opposed to the 100 µs,
10 Hz duty factor restriction needed to prevent damage to
carbon wires; and 2) there are no moving parts inside the
vacuum system, thus reducing the possibility of a vacuum
system failure. A disadvantage is that the laser is not as
rad hard as a wire scanner actuator, but we have
overcome this issue by placing the laser far away from the
beam line.
The laser profile monitor concept is straightforward: a
tightly focused laser beam is directed transversely through
the H¯ beam, causing photo-neutralization. The released
electrons are either swept away by magnetic fields
normally present in the linac lattice, or directed by a
special dipole magnet to an electron collector that may or
may not be part of the laser profile monitor system. The
beam profile is measured by scanning the laser beam
across the H¯ beam and measuring the resultant deficit in
the H¯ beam current and/or, if the released electrons are
collected, by measuring their current. A simple schematic
of the concept is shown in Fig. 2.
The advantage of collecting electrons vs. measuring the
deficit in beam current are: 1) the signal to noise ratio is
better because of the large numbers of released electrons;
and 2) the simplicity of the electron collector, since the
electron energy is well defined and the electrons are well
collimated. The disadvantages are: 1) an external
magnetic field is required, 2) an in-vacuum electron
collector is required, and 3) the electron collector signal
may suffer from interference caused by beam loss. At the
SNS linac we will use both methods. Every laser station
will have an electron collector, and there will be beam
current measurements at the entrance and exit of the
superconducting linac.
Recent developments in laser technology have raised
laser powers to the point where a low-cost laser that can
be easily carried by a person is now powerful enough to
IT08
Invited Talks
Proceedings DIPAC 2003 – Mainz, Germany
Calibration
mode voltages
Vacuum
Electron source
-16.9kV
-15.9kV
-15.8kV
Chassis at beamline
Measurement mode
voltages
-15.9 kV
-16 kV
64 Channels through two 50-pin
type D vacuum feedthroughs
Beam
ground
Balanced line driver
into individually-shielded
twisted pair cable
-1 kV
-100 V
Collector board
MCP-Burle 3810
Current Sensitive Amp, 2V/300nA
Fig. 5. (color) Schematic of the IPM.
measurement. With this setup, a laser station can be
moved or added in an 8-hour shift.
Proof of principle tests were conducted at BNL and on
the SNS MEBT at LBNL. The most recent and most
complete tests were conducted last January on the SNS
MEBT at ORNL. Shown in Fig. 3 is an example of this
latest test, where a prototype system was installed at the
end of the MEBT using the final-design beam box, dipole
magnet, and mirror actuators.
The laser profile monitor can also be used to measure
any beam that might be in the gap between the 690-ns
long mini pulses. This gap is ideally void of any beam. By
adjusting the laser firing time to occur within the gap, any
signal on the electron collector must be due to beam in the
gap. As shown in Fig. 4, this concept was also tested last
January during the MEBT commissioning, and some
beam in the gap was in fact detected, with a magnitude of
about 2 parts in 1000. We eventually expect to achieve an
accuracy of about 1 part in 10,000. We expect to
complete the installation of the laser profile system by
September 2004.
Fig. 4. (color) Some waveforms from the laser profile
monitor tests in the SNS MEBT. Top: The laser is fired
near the center of the 32 mA peak current beam bunch.
Bottom: the laser is fired during the 310-ns gap
between the 690-ns minipulses.
IONIZATION PROFILE MONITOR
almost completely strip all the electrons from the portion
of the H¯ beam intercepted by the laser. The laser can be
mounted directly to actuators on the beam line, and this
was in fact the method used for some of the earlier work.
However, concerns about long-term radiation damage
have led us to install a single laser in a room above the
SNS linac, and to transport the laser beam to the profile
monitor stations using a system of mirrors.
The laser chosen for the SNS system is the Continuum
Powerlite Precision II, 600 mJ, 10 nsec, 1064 nm, 30 Hz
ND-YAG laser. The laser beam is transported down
through a hole in the ceiling of the beam tunnel at the
downstream end of the linac, and then along the length of
the linac to the various beam profile measurement
stations. Each of the 32 warm inter-segment regions will
contain a beam box with fused-silica view ports and an
electron collector. However, to contain costs, only the
first four inter-segment regions in the medium-beta
portion of the SCL and the first four inter-segment
regions in the high-beta portion of the SCL will be
instrumented with the actuators, the electron deflection
magnet, and the electronics needed to make a profile
Invited Talks
The SNS ionization profile monitor (IPM), to be
installed in two (one horizontal, one vertical) locations in
the ring, will be based on an improved version of the
IPMs installed [5] in the RHIC ring. In fact, some of the
improvements have already been tested on the RHIC
IPMs.
The SNS (and RHIC) IPMs are based on electron
collection in parallel electric and magnetic fields. The
electrons are amplified by a microchannel plate and
collected on a 64-channel gold-plated printed circuit
board. The resultant signals are then transported through
the vacuum chamber on 50-pin D-connectors to chargesensitive amplifiers mounted near the beam line. The
signal from each channel is transported to the equipment
building using balanced line drivers and individuallyshielded twisted pair cables. An electron source has also
been added to calibrate the instrument. Some
specifications are shown in Table 1, and a schematic is
shown in Fig. 5.
The modifications to the RHIC IPM were necessary
due to rf coupling to the beam, susceptibility to beam
loss, and possible interference from the e-p beam
IT08
37
Proceedings DIPAC 2003 – Mainz, Germany
Table 1. Some specifications of the SNS IPM.
Peak beam current range
15 mA to 60 A
Bandwidth
5 MHz
Profile monitor resolution ±2.8 mm
Profile measurement
±2.5% of nominal
accuracy
beam width
Aperture radius
100 mm
Micro channel plate
Burle 3810
Electron source
Burle electron
generator array
Magnetic field
2 kG
Electric field
75 kV/m
Pre amp
Current sensitive, 2 V /
300 nA
Fig. 7. (color) Photograph of the custom-designed PCI
card and Digital Front End for the BPM system. The
Analog Front End was designed and fabricated by
Bergoz, Inc.
instability. Beam loss in the vicinity of the IPM can cause
the primary beam and secondary particle showers to pass
through the micro-channel plate and the collector board,
thus causing large background signals. Also, as
demonstrated in the LANL Proton Storage Ring, the e-p
instability can create huge amounts of electrons that could
be collected by the IPM and possibly swamp the beam
profile signal.
To alleviate these concerns the detector components
were moved outside the beam aperture by moving the
electron sweep electrode and the microchannel plate
(MCP) away from the beam centerline and shielding the
MCP with a grounded wire grid. The beam pipe in the
vicinity of the IPM (in fact all the beam pipes in the SNS
ring) will also be coated with TiN to suppress secondary
electron creation. Additionally, the IPM’s electric and
magnetic fields will now extend upstream and
downstream of the active volume to prevent electrons
created outside the IPM from entering the active volume.
Finally, the IPM’s strong electric field will prevent
electron multipacting within the active volume. To
counteract the influence of the IPM’s fields on the ring
orbit, three electromagnetic dipole magnets will be added
to the ring lattice.
The SNS beam intensity will be high enough that it will
not be necessary to inject any gas into the IPM. This will
make the system simpler, more robust, and will reduce the
costs. A 10-8 Torr vacuum is expected in the ring, which
corresponds to an expected signal level of about 150
electrons collected per turn injected into the ring (a total
of 1000 turns will be injected during normal operations).
For example, to obtain 5 to 7 profiles along the length of
the beam bunch, and to collect at least 200 electrons per
profile, we must average over about 10 machine cycles (at
60 Hz during normal operations) to get the profile
information for turns 1 – 10. For turns 11 – 1000 no
averaging will be required.
The new RHIC IPM, which incorporates many of these
design changes, was tested by purposely causing a
substantial amount of beam loss by bumping the beam
into the beam pipe wall. The test was conducted using
gold beam in the yellow ring in March 2003. As shown in
Fig. 6, the profile using the improved IPM is much better
than the profile measurement using the unmodified unit.
The SNS IPMs are scheduled be delivered to ORNL by
May 2004.
NETWORK ATTACHED DEVICES
Fig. 6. (color) The RHIC ionization profile monitor
measurement before (top) and after (bottom) the
modifications.
38
At the SNS we have chosen to base many of our
diagnostics on the rack-mounted personal computer (PC)
platform, rather than the more typical VXI, VME, or
CAMAC platforms. Instead of implementing many BPMs
within, e.g. one VME crate, the Networked Attached
Device concept implements each BPM with its own
independent resources such as a processor, a timing
IT08
Invited Talks
Proceedings DIPAC 2003 – Mainz, Germany
decoder, and a network interface. The overall costs stay
the same using the cost-effective PC platform but the
software is simplified and common failure points are
reduced. The standard software suite [6] includes
Windows 2000 or XP embedded for the operating system,
LabVIEW for the signal acquisition and signal processing
software, Input-Output Controller (IOC) core software to
communicate with the EPICS-based control system, and a
Shared Memory Interface to connect LabVIEW to the
IOC. Bench tests on a prototype network attached device
demonstrated a 100-element (with 4 bytes /element)
waveform update rate of 1000 Hz generated by LabVIEW
and communicated over the network to a remote EPICS
Channel Access client. The 800-MHz Pentium CPU was
less than 5% busy. In some cases (e.g. the beam position
monitor (BPM) and the beam current monitor (BCM)
systems), custom PCI cards have been designed and
fabricated, so that the signal cables are connected directly
to connectors on the rear panel of the PC. In other cases
(e.g. the wire scanner and the energy degrader / Faraday
Cup systems) we use off-the-shelf PCI cards to control
actuators and acquire data. For example, shown in Fig. 7
is the PCI card for the BPM system [7], and Fig. 8 shows
how the PCI card fits into a rack-mounted PC.
The Network Attached Device concept was first tested
with the prototype BPM, BCM, and wire scanner systems
on the SNS MEBT at LBNL in February 2002. All these
systems were brought on line in one short week, and
performed well during this initial commissioning period.
We did however have some difficulties interfacing to the
EPICS control system because we did not at this time
have the IOC core software installed on the PCs. We plan
to have this software ready for the upcoming DTL
commissioning.
SUMMARY
A suite of diagnostics instrumentation has been
designed to meet the challenges offered by the SNS
project. Interesting developments include the laser profile
monitor for H¯ beams, the improvements to the RHIC
ionization profile monitor, and the network attached
devices based on the PC platform.
To date the SNS facility has been commissioned up
through the end of the MEBT at 2.5 MeV using prototype
BPM, BCM, wire scanner, and slit and collector emittance
systems. All of these systems have performed well,
although a few bugs remain to be worked out, like the
IOC core software for the PC systems. The laser profile
monitor concept was also tested on the MEBT, as well as
at a couple different beam lines at RHIC.
The next stage of diagnostics installation is now in
progress to prepare for DTL commissioning later this
year, followed by CCL commissioning in 2004. The SNS
is expected to be fully commissioned by early 2006.
REFERENCES
[1] T.J. Shea et al., “SNS Accelerator Diagnostics:
Progress and Challenges,” proceedings of the 2001
Particle Accelerator Conference, Chicago, Ill, USA,
June 18 – 22, 2001.
[2] R. Hardekopf et al, “Wire Scanner Design for the
SNS Superconducting-RF Linac,” proceedings of the
2001 Particle Accelerator Conference, Chicago, Ill,
USA, June 18 – 22, 2001.
[3] R. Connolly et al., “Laser Profile Measurements of
an H- Beam,” proceedings of the 2001 Particle
Accelerator Conference, Chicago, Ill, USA, June 18
– 22, 2001.
[4] R. Connolly et al., “Laser Beam-Profile Monitor
Development at BNL for SNS,” proceedings of the
2002 Beam Instrumentation Workshop, Upton, NY,
USA, 6 – 9 May 2002.
[5] R. Connolly et al., “Performance of the RHIC IPM,”
proceedings of the 2001 Particle Accelerator
Conference, Chicago, Ill, USA, June 18 – 22, 2001.
[6] W. Blokland et al., “Network Attached Devices at
SNS,” proceedings of this conference.
[7] J. Power et al., “Beam Position Monitors for the SNS
Linac,” proceedings of the 2001 Particle Accelerator
Conference, Chicago, Ill, USA, June 18 – 22, 2001.
Fig. 8. (color) Photos of the BPM system in a 1U rackmount PC.
Invited Talks
IT08
39