Diagnostic Investigation of Tune and Tune Shift in the IPNS RCS
Diagnostic Investigation of Tune and Tune Shift in the IPNS RCS
J. C. Dooling, F. R. Brumwell, and G. E. McMichael
J. C. Dooling, F.
R. Brumwell,
and G. E. McMichael
Argonne
National Laboratory
Argonne National Laboratory
Abstract. The Intense Pulse Neutron Source (IPNS) Rapid Cycling Synchrotron (RCS) accelerates 50 MeV protons to
450
MeV 30The
times
per second
for spallation
Average
current (RCS)
from theaccelerates
RCS has recently
Abstract.
Intense
Pulse Neutron
Sourceneutron
(IPNS)production.
Rapid Cycling
Synchrotron
50 MeVexceeded
protons to
16
with30peak
instantaneous
current
approaching
A. The RCS
makescurrent
efficient
21 kV
RF accelerating
450µAMeV
times
per second for
spallation
neutron15
production.
Average
fromusetheofRCS
hasofrecently
exceeded
voltage
between
the two rf 15
cavities
to damp
verticalefficient
instabilities.
in the
16 uA and
withuses
peakphase-modulation
instantaneous current
approaching
A. The
RCS makes
use ofSplit-ring
21 kV ofelectrodes
RF accelerating
ring
suggest
an anomalous
tune shift
that increases
time to
in damp
the acceleration
cycle. Based
on a background
voltage
and uses
phase-modulation
between
the two rfwith
cavities
vertical instabilities.
Split-ring
electrodes ingas
the
pressure
of 1 an
µTorr,
the neutralization
timeincreases
for the beam
is approximately
0.5 mscycle.
at injection
the beam
ring suggest
anomalous
tune shift that
with time
in the acceleration
Based suggesting
on a background
gas
becomes
quickly
thethe
cycle.
of the
can lead
to a positive
tune
pressure fully
of 1 neutralized
(iTorr, the relatively
neutralization
timeinfor
beamOver-neutralization
is approximately 0.5
ms beam
at injection
suggesting
the beam
shift
that is
presumably
Studiesinare
to characterize theofionization
using tune
the
becomes
fully
neutralizedincoherent.
relatively quickly
theunderway
cycle. Over-neutralization
the beam within
can leadtheto RCS
a positive
existing
and Position
System (PAPS)
newly installed
Retarding
Analyzer
(RFA).
Alsousing
a newly
shift thatProfile
is presumably
incoherent.
Studies and
are aunderway
to characterize
theField
ionization
within
the RCS
the
installed
deep-memory
digitizing
theinstalled
entire history
of a single
cycle toAlso
be recorded
existing fast,
Profile
and Position
Systemoscilloscope
(PAPS) andallows
a newly
Retarding
Field acceleration
Analyzer (RFA).
a newly
from
all four
of the
split ring
electrodes allows
simultaneously
a rate of
MS/s.
installed
fast,components
deep-memory
digitizing
oscilloscope
the entireathistory
of 250
a single
acceleration cycle to be recorded
from all four components of the split ring electrodes simultaneously at a rate of 250 MS/s.
inin the
the latter
latter spectrum
spectrum and
and are
are believed
believed toto reflect
reflect the
the
vertical
vertical tune,
tune, QQy.y. Amplitudes
Amplitudes ofofthe
the sidebands
sidebands are
are aa
strong
strongfunction
function ofofcirculating
circulatingcharge.
charge.
INTRODUCTION
INTRODUCTION
The
Rapid
TheIntense
IntensePulsed
PulsedNeutron
NeutronSource
Source(IPNS)
(IPNS)
Rapid
12
Cycling
CyclingSynchrotron
Synchrotron(RCS)
(RCS)accelerates
accelerates3x10
3xl012protons
protons
from
from50
50MeV
MeVtoto450
450MeV,
MeV,30
30times
timesper
persecond.
second. The
The
bare
and Qy~2.35.
bare xx and
and yy tunes
tunes are
are QQx~2.2
x~2.2 and Q
y~2.35.
Observations
Observationsofofspectra
spectrarecorded
recordedfrom
fromcapacitive
capacitivepickpickup
electrodes
within
the
RCS
suggest
that
up electrodes within the RCS suggest thatfor
foraaportion
portion
ofofthe
thebeam,
beam,the
thetune
tunemay
maybe
beincreasing
increasingfor
forpart
partofofthe
the
accelerator
acceleratorcycle.
cycle. InIn1999,
1999,looking
lookingatatdata
datafrom
from these
these
“Pie”
"Pie" position
position electrodes[1]
electrodes[l] using
using aa gated
gated spectrum
spectrum
analyzer,
analyzer,aabeam
beamloss
lossphenomena
phenomenareferred
referredtotoasas“pre"prespill”
spill"was
wasclearly
clearlyseen
seen totobe
bethe
the result
result ofof aa resistive
resistive
wall
wall (vertical)
(vertical) instability
instability asas had
had been
been previously
previously
surmised[2].
surmised[2]. The
Theinstability
instabilitycauses
causesaa50
50percent
percentloss
loss
ofof beam
beam atat the
the end
end ofof the
the acceleration
acceleration cycle
cycle ifif left
left
unchecked.
unchecked. The
Theresistive
resistivewall
wallinstability
instabilityisiscontrolled
controlled
through
throughphase
phasemodulation
modulation(PM)
(PM)ofofthe
therfrfaccelerating
accelerating
voltage.
voltage. The
Thephase
phasemodulator
modulatororor“scrambler,”
"scrambler,"varies
varies
the
theposition
position ofofone
oneofofthe
thetwo
tworfrfaccelerating
accelerating cavities
cavities
by
bysinusoidal
sinusoidalamplitude
amplitudeofof100
100mrad
mrad(6°)
(6°)near
neartwice
twice
the
thesynchrotron
synchrotronfrequency.
frequency. The
Thescrambler
scramblercircuit
circuitalso
also
introduces
introduces amplitude
amplitude modulation
modulation (AM)
(AM) atat the
the same
same
frequency.
frequency. “Slow-Q”
"Slow-Q"(captured
(capturedcharge)
charge)signals
signals with
with
and
andwithout
withoutthe
thescrambler
scramblerare
arepresented
presentedininFig.
Fig. 1.1.
injection
w/PM
a
0
5
10
15
time after injection (ms)
FIGURE
FIGURE1.1. Captured
CapturedRCS
RCScharge
chargewith
withand
andwithout
withoutPM.
PM.
The
Thesidebands
sidebandscorrespond
correspondtoto(n±ν
(n±vyy)f)fo0 for
fornn>>QQyy, ,inin
agreement
with
a
resistive
wall
instability[3].
agreement with a resistive wall instability[3]. The
The
scrambler
scramblerdamps
dampsthe
thetransverse
transverseinstability
instability presumably
presumably
by
by increasing
increasing the
the energy
energy spread
spread of
of the
the bunch.
bunch. One
One
way
way this
this can
can occur
occur isis by
by establishing
establishing aa parametric
parametric
oscillation[4],
oscillation[4], where
wherethe
the bunch
bunchrotates
rotates inin longitudinal
longitudinal
phase
phasespace
spacewhile
whilemaintaining
maintainingaaconstant
constantemittance.
emittance.
RESISTIVE
RESISTIVEWALL
WALLINSTABILITY
INSTABILITY
FAST
FASTRING
RINGDATA
DATA
Since
Sincethe
themid-1980s,
mid-1980s, the
thescrambler
scrambler has
has been
been used
used
totostabilize
stabilizethe
thebunch
bunchduring
duringthe
thelast
lastfew
fewmilliseconds
milliseconds
ofofthe
theacceleration
acceleration cycle.
cycle. Figure
Figure22shows
showstwo
twospectra
spectra
taken
taken just
just prior
prior toto extraction,
extraction, 14
14ms
ms after
after injection,
injection,
from
from aa vertical
vertical pie
pie electrode
electrode using
using an
an HP4396B
HP4396B
spectrum
spectrumanalyzer
analyzer(SA).
(SA). The
The spectrum
spectrumpresented
presented inin
Figure
Figure 2a2a isis that
that with
with the
the scrambler
scrambler modulation
modulation on
on
from
from 9.5
9.5 ms
ms toto 11.5
11.5 ms,
ms, and
and that
that inin 2b
2b isis with
with the
the
scrambler
scrambleroff.
off. Significant
Significantside-bands
side-bandsare
areininevidence
evidence
One
One concern
concern regarding
regarding the
the SA
SA isis the
the degree
degree to
to
which
which itit averages
averages and
and possibly
possibly distorts
distorts the
the actual
actual
bunch
bunch signal
signal during
during the
the cycle.
cycle. To
To reduce
reduce noise,
noise, 10
10
sweeps
sweeps are
are averaged
averaged toto generate
generate aa single
single spectrum;
spectrum;
each
each spectrum
spectrum requires
requires approximately
approximately 66 minutes
minutes toto
produce
produceatat aaBW
BWofof30
30kHz.
kHz. Thus
Thusatat30
30Hz,
Hz,we
weare
are
sampling
over
10,000
acceleration
cycles;
i.e.,
10,000
sampling over 10,000 acceleration cycles; i.e., 10,000
bunches.
bunches. The
Thepresent
presentgeneration
generationofoffast,
fast, deep-memory
deep-memory
CP642, High Intensity and High Brightness Hadron Beams: 20th ICFA Advanced Beam Dynamics Workshop on
High Intensity and High Brightness Hadron Beams, edited by W. Chou, Y. Mori, D. Neuffer, and J.-F. Ostiguy
© 2002 American Institute of Physics 0-7354-0097-0/02/$ 19.00
77
oscilloscopes
oscilloscopes makes
makes itit possible
possible to produce fast spectra
from
from aa single
single 14 ms acceleration period. Using aa
Tektronix
fast
Tektronix TDS7254
TDS7254 oscilloscope with 16 MB of fast
memory,
memory, the
the entire
entire RCS
RCS cycle
cycle can be recorded
simultaneously
simultaneously on
on four
four channels at a rate of 250
MSamples
MSamples per
per second.
second. The
The penalty for speed is
significantly
significantly higher
higher noise
noise with respect to the SA due to
the
the greater
greater oscilloscope
oscilloscope bandwidth
bandwidth (2.5 GHz).
itf
io"J
10
15
f(MHz)
(a)
0
5
20
10
15
f(MHz)
FIGURE
FIGURE 3.
3. Comparison
Comparison of
of top
top pie
pie electrode
electrodedata
datarecorded
recorded
with
with a)
a) SA
SA and
and b)
b) oscilloscope
oscilloscope 0.3
0.3 ms
msafter
after injection.
injection.
SCRAMBLER
SCRAMBLER STUDIES
STUDIES
The
The scrambler
scrambler helps
helps to
to damp
damp oscillations
oscillations that
that are
are
indicative
of
destructive
instabilities;
these
oscillations
indicative of destructive instabilities; these oscillations
can
can appear
appear as
as spectral
spectral features
features such
such as
as the
the sideband
sideband
structure
shown
in
Figure
1b.
While
the
structure shown in Figure Ib. While the scambler
scambler
removes
removes one
one type
type of
of narrow
narrow band
band (coherent)
(coherent)structure
structure
from
the
spectra,
it
introduces
other
from the spectra, it introduces other features
features which
whichare
are
broader
broader in
in nature.
nature. Figure
Figure 44 presents
presentspie
piespectra
spectraatattwo
two
times
times in
in the
the RCS
RCS cycle,
cycle, 10
10 ms
ms and
and 11
11 ms
ms after
after
injection.
The
latter
curve
has
been
shift
downward
injection. The latter curve has been shift downward
two
two orders
orders of
of magnitude
magnitude to
to make
make itit more
morevisible.
visible. The
The
scrambler
begins
at
9.4
ms
after
injection;
scrambler begins at 9.4 ms after injection; however,
however,
only
only the
the primary
primary bunch
bunch harmonics
harmonics are
arevisible
visibleatat10
10ms.
ms.
One
millisecond
later,
the
spectrum
has
changed
One millisecond later, the spectrum has changed with
with
aa pair
pair of
of broadband,
broadband, transverse
transverse modulation
modulation features
features
present
on
either
side
of
the
bunch
present on either side of the bunch harmonics.
harmonics. The
The
modulation
modulation features
features are
are thought
thought to
to be
be transverse
transverse
because they do not appear on longitudinal diagnostics
because they do not appear on longitudinal diagnostics
such as the beam current toroid (CT) or resistive wall
such as the beam current toroid (CT) or resistive wall
monitor (RWM) in the RCS. Thus, coupling from
monitor (RWM) in the RCS. Thus, coupling from
longitudinal motion into the transverse plane may also
longitudinal motion into the transverse plane may also
help to damp the resistive wall instability. The broad
help to damp the resistive wall instability. The broad
sidebands are not present when the scrambler is off.
sidebands are not present when the scrambler is off.
Knowing that the scrambler generates sidebands,
Knowing that the scrambler generates sidebands,
an effort was made to see if the these spectral features
an
effort
was made to see if the these spectral features
could be observed elsewhere in the acceleration cycle.
could
be
observed
elsewhere in the acceleration cycle.
The scrambler electronics allows for two modulation
The
scrambler
electronics
for Keeping
two modulation
periods at different times in allows
the cycle.
one at
periods at different times in the cycle. Keeping one at
FIGURE 2.
2. Bunch
Bunch Spectra
Spectra near
near extraction
extraction a)
with scrambler
FIGURE
a) with
scrambler
andb)
b)without
without scrambler.
scrambler.
and
To evaluate
evaluate the
the above
above concern,
concern, aa comparison
comparison has
has
To
been
made
of
spectra
obtained
from
both
the
SA
and
been made of spectra obtained from both the SA and
oscilloscope with
with the
the data
data given
given in
in Figure
Figure 3.
oscilloscope
3.
Approximately
0.3
ms
after
injection,
the
Approximately 0.3 ms after injection, the top
top pie
pie
electrode signal
signal recorded
recorded with
with the
the SA
SA exhibits
exhibits strong
electrode
strong
sidebands
as
shown
in
Figure
3a.
The
sample
sidebands as shown in Figure 3a. The sample period
period is
is
0.1 ms.
ms. Likewise,
Likewise, sideband
sideband structure
structure can
can also
also be
be seen
0.1
seen
in the
the spectra
spectra generated
generated from
from TDS7254
TDS7254 oscilloscope
oscilloscope
in
data
and
presented
in
Figure
3b.
The
time
data and presented in Figure 3b. The time data
data was
was
recorded at the same point in the cycle from an
recorded at the same point in the cycle from an
adjacent pie electrode during an 80 microsecond
adjacent
pie electrode during15 an 80 microsecond
period (20 kS interpolated to 215
values for the FFT).
period (20 kS interpolated to 2 values for the FFT).
The individual bunch shape recorded with the
The individual bunch shape recorded with the
oscilloscope differs slightly from that of SA but both
oscilloscope differs slightly from that of SA but both
represent an average. Note that the noise floor is
represent an average. Note that the noise floor is
several orders of magnitude higher for the oscilloscope
several orders of magnitude higher for the oscilloscope
generated spectrum. The noise level in the SA data is
generated
spectrum. The noise level in the SA data is
typically in the range of 10-12 W. Sidebands indicate a
typically in the range of 10"12 W. Sidebands indicate a
vertical fractional tune of 0.28; the horizontal tune is
vertical
fractional
tune of 0.28; the horizontal tune is
0.20.
Irrespective of the noise level, the two
0.20.
Irrespective
the spectral
noise level,
techniques do indicateof
similar
features.the two
techniques do indicate similar spectral features.
78
.01
S(10 ms)
10
S(llms)x.01 -
10
15
f(MHz)
20
FIGURE
FIGURE
4.
Bunch
spectra
from
top
pie
electrode
10
ms
and
FIGURE4.
4.Bunch
Bunchspectra
spectrafrom
from top
toppie
pie electrode
electrode10
10ms
msand
and
11
11
ms
(x0.01)
after
injection.
Scrambler
begins
at
9.4
ms.
11ms
ms(x0.01)
(xO.Ol)after
after injection.
injection. Scrambler
Scramblerbegins
beginsat
at9.4
9.4ms.
ms.
the
the
normal
time
for
stability
while
varying
the
other, itit
the normal
normal time
time for
for stability
stability while
while varying
varying the
the other,
was
possible
to
stimulate
the
bunch
and
observe
was
was possible
possible to
to stimulate
stimulate the
the bunch
bunch and
and observe
observe
sidebands.
sidebands.
The
scrambler
study
was
first
conducted
in
sidebands. The
The scrambler
scrambler study
studywas
wasfirst
first conducted
conducted in
in
December
December
1999
and
repeated
again
in
October
2001.
December 1999
1999 and
and repeated
repeated again
again in
in October
October 2001.
2001.
The
The
second
test
generally
repeated
the
results of
of
the
The second
second test
test generally
generally repeated
repeated the
the results
of the
the
earlier
study.
Scrambler
study
measurements
earlier
study.
Scrambler
study
measurements
are
earlier study. Scrambler study measurements are
are
given
given
in
Fig.
5.
Background
gas
pressure
in
the
RCS
given in
in Fig.
Fig. 5.
5. Background
Background gas
gas pressure
pressure in
in the
the RCS
RCS
is
isis on
on
the
order
of
µTorr.
Assuming
background
of
on the
theorder
order of
of111µTorr.
jiTorr. Assuming
Assumingaaabackground
backgroundof
of
N
;
the
neutralization
time,
τ
=
0.5
ms
at
injection
and
2
n
N
;
the
neutralization
time,
τ
=
0.5
ms
at
injection
and
N22; the neutralization time, Tnn = 0.5 ms at injection and
1.0
1.0
ms
atatextraction;
extraction;
therefore,
the
beam
becomes
fully
1.0ms
msat
extraction;therefore,
therefore, the
thebeam
beambecomes
becomesfully
fully
neutralized.
The
tune
shift
introduced
by
partially
neutralized.
The
tune
shift
introduced
by
partially
neutralized. The tune shift introduced by partially
compensated
compensated
space-charge,
may
be
expressed
as[5],
compensated space-charge,
space-charge,may
maybe
beexpressed
expressed as[5],
as[5],
(
)
IILp ,(T)R(l(τ)R 1 − γ 222((ττ)f
)feee ((T))
(ττ))
p ( τ)R 1 − Yγ (T)f
∆ν
((ττ)) ===−−—
∆ν
AV(T)
II0 εεn ββ22((ττ))γγ22((ττ))
0 n
10
12
pulsed quadrupole
current waveforms
0
2
4
6
8
t(ms)
10
14
-
12 14
FIGURE
FIGURE
5.
Extended
scrambler
tune
measurements
FIGURE5.
5.Extended
Extendedscrambler
scramblertune
tunemeasurements
measurements
collection
grid
at
the
bottom
of
one
straight
collection
grid
at
the
bottom
of
one
section,
collection grid at the
one straight
straight section,
section,
clearly
shows
the
generation
of
electrons
in
the
clearly
shows
the
generation
of
electrons
in
the
beam.
clearly shows the
of electrons in the beam.
We
have
recently
installed
a
Retarding
Field
Analyzer
We
have
recently
installed
a
Retarding
Field
Analyzer
We have recently installed a Retarding Field Analyzer
(RFA)
(RFA)
of
the
type
used
in
the Proton
Proton Storage
Storage Ring
Ring at
at
(RFA) of
of the
the type
type used
used in
in the
LANL
and
the
Advanced
Photon
Source
here
LANL
at
LANL and
and the
the Advanced Photon Source
Source here at
ANL[6].
ANL[6].
With
we
hope
to
learn
about
the
electron
ANL[6]. With
With ititit we
wehope
hope to
to learn
learn about
about the
the electron
density
density
and
energy
spread
near
the
beam. We
We look
look
density and
and energy
energy spread
spread near
near the
the beam.
look
forward
to
having
RFA
data
during
our
June
run.
forward
forward to
to having
having RFA
RFA data
data during
during our
our June
June run.
run.
(1)
(1)
(1)
where
where
isis the
the
fractional
neutralization.
If
overwhere fffeee is
the fractional
fractional neutralization.
neutralization. If
If overoverneutralization
is
allowed
and
f
(t)
=
t/τ
,
where
e
n
neutralization
is
neutralization isis allowed
allowed and
and ffee(t)
(t) == t/τ
t/Tnn,, where
where ttt is
is
time
time
measured
from
injection,
then
the
tunes
predicted
timemeasured
measured from
from injection,
injection, then
then the
the tunes
tunes predicted
predicted
by
by
Eq.
are
as
presented
in
Fig.5.
Some
of
the
by Eq.
Eq. 111 are
are as
as presented
presented in
in Fig.5.
Fig.5. Some
Some of
of the
the
structure
in
the
measured
tunes
may
be
the
result
of
structure
in
the
measured
tunes
may
be
the
result
of
structure in the measured tunes may be the
the
pulsed
quadrupole
magnets
in
the
RCS,
the
current
the
pulsed
quadrupole
magnets
in
the
RCS,
the
current
the pulsed quadrupole magnets in the RCS, the
waveforms
waveforms
of
which
are
given
at
the
bottom
of
Fig.5.
waveforms of
ofwhich
whichare
aregiven
givenat
atthe
the bottom
bottom of
ofFig.5.
Fig.5.
ACKNOWLEDGMENTS
ACKNOWLEDGMENTS
ACKNOWLEDGMENTS
Many
Many
thanks
to
the
IPNS
Accelerator Operations
Operations
Many thanks
thanks to
to the
the IPNS
IPNS Accelerator
Operations
Group
that
makes
this
work
possible.
Group
that
makes
this
work
possible.
Group that makes this work possible.
DISCUSSION
DISCUSSION
AND
FURTHER
WORK
DISCUSSION AND
AND FURTHER
FURTHER WORK
WORK
REFERENCES
REFERENCES
REFERENCES
Neutralization
Neutralization
through
ionization
means
electrons
Neutralization through
through ionization
ionization means
means electrons
electrons
and
ions
are
generated
in
the
background
and
gas,
perhaps
and ions
ions are
are generated
generated in
in the
the background
background gas,
gas, perhaps
leading
leading
to
the
formation
of
plasma.
A
plasma
sheath
leading to
tothe
the formation
formation of
ofaaaplasma.
plasma. A
Aplasma
plasma sheath
sheath
forming
forming
around
the
beam
would
both
shield
the
space
forming around
around the
the beam
beam would
would both
both shield
shield the
the space
space
charge
charge
potential
of
the
beam
from
the
wall
as
well
as
charge potential
potential of
of the
the beam
beam from
from the
the wall
wall as
as well
well as
as
carry
carry
image
current.
Plasma
formation
would
be
carry image
image current.
current. Plasma
Plasma formation
formation would
would be
be
enhanced
enhanced
in
the
combined-function
magnet
sections
of
enhanced in
in the
the combined-function
combined-function magnet
magnet sections
sections of
of
the
RCS
which
cover
just
over
half
of
the
beam
path.
the
RCS
which
cover
just
over
half
of
the
beam
path.
the RCS which cover just over half of the beam path.
Charge
Charge
confinement
times
in
the
magnets
should
be
Charge confinement
confinement times
times in
in the
the magnets
magnets should
should be
be
longer
then
in
the
field-free
straight
longer
sections.
longer then
then in
in the
the field-free
field-free straight
straight sections.
sections.
Liberated
Liberated
charge
is
restricted
to
follow
magnetic
field
Liberated charge
charge is
is restricted
restricted to
to follow
follow magnetic
magnetic field
field
lines
which
lead
quickly
above
and
below
lines
the
beam to
to
lines which
which lead
lead quickly
quickly above
above and
and below
below the
the beam
aaa stainless
stainless
steel
liner.
Secondary
emission
from
the
stainless steel
steel liner.
liner. Secondary
Secondary emission
emission from
from the
the
liner
liner
produces
more
electrons
which
are
again
forced
liner produces
produces more
more electrons
electrons which
which are
are again
again forced
forced
to
follow
field
lines
back
to
the
beam.
An
electron
to
follow
field
lines
back
to
the
beam.
An
electron
to follow field lines back to the beam. An electron
1.
1.
A.
V.
Rauchas,
et
al.,
IEEE
Trans.
Nuc.
Sci.,
28(3),
2338
1. A.
A.V.
V.Rauchas,
Rauchas,et
etal.,
al.,IEEE
IEEETrans.
Trans.Nuc.
Nuc.Sci.,
Sci.,28(3),
28(3),2338
2338
(1981).
(1981).
(1981).
2.
2.
C.
Potts,
et
al.,
IEEE
Trans.
Nuc.
Sci.,
32(5),
3107
2. C.
C. Potts,
Potts, et
et al.,
al., IEEE
IEEE Trans.
Trans. Nuc.
Nuc. Sci.,
Sci., 32(5),
32(5), 3107
3107
(1985).
(1985).
(1985).
3.
L. J.
Laslett, V.
K. Neil,
and A.
M. Sessler, Rev. Sci.
3.
3. L.
L. J.J. Laslett,
Laslett, V.
V. K.
K. Neil, and
and A. M. Sessler, Rev. Sci.
Instrum.,
Instrum.,
36(4),
1965(436).
Instrum.,36(4),
36(4),1965(436).
1965(436).
4.
M. Bai,
et al.,
Phys. Rev.
ST Accel.
Beams 3,
4.
4. M.
M. Bai,
Bai, et
et al.,
al., Phys.
Phys. Rev.
Rev. ST
ST Accel.
Accel. Beams
Beams 3,
3,
064001(2000).
064001(2000).
064001(2000).
5.
5.
M.
Reiser,
Theory
and
Design
of
Charged
Particle
5. M.
M. Reiser,
Reiser, Theory
Theory and
and Design
Design of
of Charged
Charged Particle
Beams,
Wiley,
New
York,
1994,
p.
274.
Beams,
Wiley,
New
York,
1994,
p.
274.
Beams, Wiley, New York, 1994,
6.
6. R.
R. A.
A. Rosenberg,
Rosenberg, K.
K. C.
C. Harkay,
Harkay, NIM
NIM A,
A, 453,
453, 507(2000)
507(2000)
6. R. A. Rosenberg, K. C. Harkay, NIM A, 453, 507(2000)
79