Design of a High-Precision Fast Wire Scanner for the SPS at CERN

Abstract

Studies are going on of a new wire scanner concept. All moving parts are inside the beam vacuum and it is specified for use in all the machines across the CERN accelerator complex. Key components have been developed and tested. Work is now focussing on the installation of a prototype for test in the Super Proton Synchrotron (SPS) accelerator. This article presents the specification of the device and constraints on the design for integration in the different accelerators at CERN. The design issues of the mechanical components are discussed and optimisation work shown. Finally, the prototype design, integrating the several components into the vacuum tank is presented.

Full text

DESIGN OF A HIGH-PRECISION FAST WIRE SCANNER FOR THE SPS
AT CERN
R. Veness, N. Chritin, B. Dehning, J. Emery, J. Herranz Alvarez, M. Koujili, S. Samuelsson,
J.L. Sirvent, CERN, Geneva, Switzerland
Abstract
Studies are going on of a new wire scanner concept.
All moving parts are inside the beam vacuum and it is
specified for use in all the machines across the CERN
accelerator complex. Key components have been
developed and tested. Work is now focussing on the
installation of a prototype for test in the Super Proton
Synchrotron (SPS) accelerator.
This article presents the specification of the device and
constraints on the design for integration in the different
accelerators at CERN. The design issues of the
mechanical components are discussed and optimisation
work shown. Finally, the prototype design, integrating the
several components into the vacuum tank is presented.
INTRODUCTION
Wire scanners are installed in the LHC and all circular
machines in the injector chain as a means to measure the
transverse beam profile and hence emittance. The
motivation for the development of a new scanner design
has been described in a previous article [1], along with the
concept with the rotor of the motor and wire position
measurement system inside the beam vacuum [see
Figure 1]. Development of key components, in particular
the motor and control system, are well advanced [2].
Work is now focussing on the integration of all the
required components with the aim of producing a scanner
capable of 20 ms-1 scanning speed combined with 2 µm
position precision.
A number of mechanical components require careful
optimisation. These include the motor housing, shaft,
bearings, fork and wire. In addition, the design concept
includes an in-vacuum optical position encoder in order to
reach the required precision. Development of these
components is described in the following sections.
Figure 1: Fast Wire Scanner concept.
INTEGRATION CONSTRAINTS
Wire scanners are currently installed in the PS, Booster,
SPS and LHC at CERN. It would greatly simplify
operation and maintenance if the same basic design could
be implemented for all of these machines. To this end, the
main constraints in terms of machine physics, operation
and environment have been analysed for each machine.
These are summarised in Table 1.
Table 1: Summary of Integration Constraints from the
CERN Accelerator Complex
Machine Scan
aperture
(mm)
RF
Screen
Bakeout Space
Constraint
PS
Booster
146x70 N N Axial,
Transverse
PS 146x70 N N Axial
SPS 152x83 Y N -
LHC 65x65 Y Y Transverse
The scan aperture is the horizontal and vertical space
that must be cleared by the wire. RF screens are required
in some machines to minimise impedance and RF heating
effects. Integration of new scanners into existing
machines must take into account machine geometries and
equipment. Axial space constraints occur in machines
with a tight lattice whereas transverse constraints are seen
with parallel equipment on the beamline (eg, the
cryogenic distribution line in the LHC). It can be seen
from table 1 that each of the machines brings constraints
to the design. A solution has been adopted where the main
components can be integrated into designs for the PS,
SPS and LHC. Each machine will require a different fork
geometry and a different flange interface, but other main
components and principles will be common. The layout of
the Booster with 4 rings in very close proximity mean that
it has not yet been possible to integrate the design into
this machine.
Combining these constraints leads to a design with
aperture range up to 152 by 152, with the option to
include RF screen and to be bakeable to 200°C in order to
activate a low emission yield getter coating used in the
LHC vacuum system.
DESIGN OF COMPONENTS
Motor Housing
The motor housing has the function of separating the
rotor in-vacuum and stator on the atmospheric side of the
Proceedings of IBIC2012, Tsukuba, Japan MOPB79
Transverse Profiles, Screens & Wires Monitors
ISBN 978-3-95450-119-9
1
Copyright c
○2012 by the respective authors — cc Creative Commons Attribution 3.0 (CC BY 3.0)

electric motor. The housing needs to be thin to fit in the
relatively small air gap (approximately 0.8 mm) between
the rotor and stator. The required wall thickness of the
housing has been determined using finite element method
(FEM) analysis. The analysis shows that elastic instability
is the critical failure mode. The stability of the structure
increases strongly with wall thickness. The dependence of
stability on the length of the structure is weaker, where a
shorter structure increases the stability. Furthermore, a
thicker housing (within the limits of this problem) is
easier to manufacture. When all the above mentioned
considerations are put together, the optimum wall
thickness of the motor housing is determined to be 0.4
mm.
Figure 2: Graph of motor housing collapse load for
different lengths and wall thicknesses.
Shaft and Bearings
One principal of this design is to support forks and
position measurement system on one rigid structure to
maximise precision. The shaft forms the core of the
mechanical structure. It is driven by the motor and
supports the rotor, forks and optical disc. The main
constraint on the shaft is its deformation in torsion. The
torsional deformation appears due to the acceleration of
the shaft and the inertias of the components mounted on
the shaft. This deformation must be kept small to ensure
accuracy of the measurements. Although the shaft is not
accelerating when the wire passes through the beam, the
shaft will vibrate with a maximum amplitude
corresponding to the torsional deformation due to the
acceleration. The shaft needs to be hollow to be able to
pass cables through it. Analysis of the shaft, using
analytical calculations and FEM simulations, show that
using a larger outer diameter has a strong effect on the
stiffness. It is also shown that the stress in the shaft is low
and the inertia of the shaft is of little importance
compared to the inertias of the components mounted on
the shaft. This means that the strength of the material is
not critical and that a stiff, relatively heavy stainless steel
is a better material choice compared to lighter, more
flexible alternatives such as aluminium or titanium. The
analysis also shows that it is the optical disc and the disc
holder which give the largest contribution to the shaft
twist. This is because they are mounted on the shaft end
opposite to the motor. Therefore effort should be put into
minimising the mass of these components. Figure 3
showsthe offset of the wire position relative to the
encoder which the shaft vibrations give rise to. The
calculations show that an outer diameter of 35 mm is
needed to keep the deformation below the tolerated limit
of 5 μm.
Figure 3: Graph of relative offset of wire vs. encoder for
different shaft diameters and thicknesses.
The bearings need to assure high precision in terms of
radial runout and the materials used must be UHV
compatible and radiation and bake out resistant. This
means that traditional lubricants such as oil and grease
cannot be used [3]. Instead one must rely on running the
bearings without any lubricant or possibly using solid
lubricants (such as molybdenum disulphide or tungsten
disulphide coatings). It is also recommendable to use
different materials for the races and the rolling elements
in the bearing, to avoid cold welding. One available
alternative for this is hybrid bearings which use steel
races and ceramic rolling elements.
Optical Disc and Support
The principle selected for the high precision
determination of the beam size is an optical system based
on a glass disc with a photo-lithographed μm pattern
made of high reflectivity chrome, placed inside the
vacuum chamber and fixed on the scanner shaft. This
incremental angle encoder uses single-mode optical fibre
and UHV fibre optic feedthrough (9/125μm) to drive
1310 nm laser light on a 1:1 lens system in-vacuum that
focuses the light on the disc surface with a 10 μm light
spot size. Using the reflectivity of the chrome pattern, the
reflected light is coupled back into the same fibre, and
through an optical circulator directed to the photodiode.
The laser diode, circulator, photodiode and subsequent
electronics will be located in the surface building and
only one optical fibre will go down to the accelerator
tunnel (250 m). The performance of this single fibre
angular position sensor has been tested and validated on
the bench shown in Fig 4.
MOPB79 Proceedings of IBIC2012, Tsukuba, Japan
ISBN 978-3-95450-119-9
2
Copyright c
○2012 by the respective authors — cc Creative Commons Attribution 3.0 (CC BY 3.0)
Transverse Profiles, Screens & Wires Monitors

Figure 4:
This fibre
with a track
using only
o
reached by
a
Heidenhain
eccentricity
minimized.
Figure 5:
E
applied (cen
t
made with
H
The final
immunity,
w
radiation tol
e
Tes
t
bench f
o
encoder pro
v
of 10 μm sl
i
o
ne channel.
a
ngular calibr
a
RON225 (
F
errors, and
E
rror before
t
er) and erro
r
H
eidenhein R
O
system is
w
orks with te
m
e
rant due the
o
r the optical
p
v
ides a resol
u
i
ts, and two
p
An accurac
y
a
tion with a c
o
F
ig. 5). With
partly grati
n
calibration
histogram (b
o
O
N225 angula
UHV com
p
m
peratures up
special fibr
e
p
osition sens
o
u
tion of 157
p
osition refer
e
y
of ±25 μ
R
o
mmercial en
c
this calibr
a
n
g errors c
a
(top), calib
r
o
ttom). Calib
r
r
position sen
p
atible with
to 200oC, an
d
e
used. Studi
e
or
.
μRad
ences
R
ad is
coder
a
tion,
a
n be
r
ation
r
ation
n
so
r
.
EMI
d it is
e
s are
on-
g
mec
h
inco
r
wor
k
p
roc
M
a
g
T
h
to
p
tran
s
case
U
n
aper
t
que
n
LH
C
own
no
m
T
h
of a
vac
u
othe
r
p
iec
e
alig
n
equi
p
osi
t
the
f
rota
t
Fig
u
syst
e
Fo
r
T
h
care
f
mai
n
defo
Ho
w
iner
t
dist
r
man
u
for
qua
n
Con
s
mak
i
man
u
met
h
usin
g
fro
m
p
ro
d
g
oing to verif
y
h
anical reso
l
r
poration of
o
k
ing axis to
edure.
g
netic Rest
r
h
e design wil
l
p
revent unpl
a
s
portation or
i
of power o
r
c
n
controlled
m
t
ure could
c
n
ching of sup
e
C
). This effec
t
unbalanced
w
m
inal speed re
q
h
e conceptua
l
magnetic cir
c
u
um has a pe
r
r
p
art inside
t
e
fixed to th
e
n
ed with th
e
l
ibrium posit
i
t
ions. When
t
f
iel
d
of the
p
t
e freely. This
u
re 6: Conce
p
e
m.
r
ks
h
e two fork
s
f
ully optimis
n
tain the wir
e
rmation and
w
ever the for
k
t
ial load on t
h
r
ibution ne
e
u
facturing vi
e
each machi
n
n
tities will be
s
idering all o
i
ng these
u
facturing [4
]
h
od will allo
w
g
FEM, and
m
3D CAD m
o
d
uced with co
n
y
the use of a
l
utions (aro
u
o
ther calibrati
o
perform a
r
aint Syste
m
l
integrate a
m
a
nne
d
move
m
i
nstallation a
n
c
ontrol syste
m
m
ovement of
c
ause the
m
e
rconducting
m
is enhanced
a
w
eight woul
d
q
uired fo
r
saf
e
design of th
c
uit in two pa
r
r
manent mag
n
t
he vacuum
c
e
shaft. This
external m
a
i
on correspo
n
t
he electric c
p
ermanent m
a
concept can
b
p
tual design
which sup
p
e
d. They ne
e
e
under tens
i
vibration
d
k
s are also
a
e shaft and
m
e
d to be
e
wpoint, ther
e
n
e due to t
h
small.
f these facto
r
components
]
(also referr
e
w
complex g
e
produced d
i
o
dels allowin
n
ventional m
a
5 μm track t
o
u
nd 70 μR
a
on methods
o
more reliab
l
m
m
agnetic res
t
m
ents of the
n
d also durin
g
m
failure.
the wire wit
h
m
elting of t
h
magnets (in t
h
a
s
t
he wire s
p
d
be much sl
o
e
opera
t
ion o
f
h
e restrain
t
s
y
r
ts: One
p
art
n
et and elect
r
c
onsists of a
f
has two
p
ol
e
a
gnetic circ
u
n
ding to the
c
oil is energi
z
a
gnet allowin
g
b
e seen in Fi
g
for the mag
n
p
ort the wir
e
ed to be sti
f
i
on and prev
e
d
uring wire
a
major com
p
m
otor, so the
m
optimised.
e
will be dif
f
h
e different
r
s, it is logic
a
using m
e
e
d to as 3D
p
eometries to
i
rectly in s
m
n
g for forms t
h
a
chining.
o
reach bette
r
a
d), and th
e
o
n the scanne
r
l
e calibratio
n
t
rain
t
in orde
r
fork durin
g
g
operation i
n
h
in the bea
m
h
e wire an
d
h
e case of th
e
p
eed due to it
s
o
wer than th
e
the scanner.
y
stem consist
s
o
utside of th
e
r
ical coil; th
e
f
erromagneti
c
e
s that, whe
n
u
i
t
, define a
n
‘parked’ for
k
z
ed it cancel
s
g
the shaft t
o
g
ure 6.
n
etic restrain
t
e
need to b
e
f
f enough t
o
e
nt excessiv
e
acceleration
.
p
onent of
t
h
e
m
ass and mas
s
From th
e
f
erent design
s
apertures, s
o
a
l to conside
r
e
tal additiv
e
p
rinting). Thi
s
be optimise
d
m
all quantitie
s
h
at cannot b
e
r
e
r
n
r
g
n
m
d
e
s
e
s
e
e
c
n
n
k
s
o
t
e
o
e
.
e
s
e
s
o
r
e
s
d
s
e
Proceedings of IBIC2012, Tsukuba, Japan MOPB79
Transverse Profiles, Screens & Wires Monitors
ISBN 978-3-95450-119-9
3
Copyright c
○2012 by the respective authors — cc Creative Commons Attribution 3.0 (CC BY 3.0)

Studies of the wire are on-going and will be presented
separately.
STATUS AND NEXT STEPS
All major components have been integrated into a 3D
model with an envelope which would allow installation
into the PS, SPS or LHC. Figure 7 shows a section
through this model. The scanner is assembled as one self-
contained ‘cartridge’ that will be inserted into the
accelerator vacuum chamber. This will protect the wire
from damage during insertion – a common problem with
existing designs. A prototype will be constructed and
tested in the coming months to verify the operation of the
scanner assembly and performance. It is then planned to
produce a first production model for test in the SPS
accelerator. This will be installed in the forthcoming
‘Long Shutdown 1’ of all CERN accelerators in 2013-14.
The plan is then to produce a series of scanners for
installation in the second Long Shutdown scheduled for
2018-19.
Figure 7: 3-D model section through the scanner.
REFERENCES
[1] M.Koujili et al. ‘Fast and High Accuracy Wire Scanner’
DIPAC’09 Conference – 25-27 May 2009 /Basel-CH
[2] B.Dehning et al. ‘Vacuum Actuator and Controller Design
for a Fast Wire Scanner’ Proceedings of BIW2012,
Newport News, Virginia, USA
[3] D. Ramos ‘Bibliographic search on bearing technology for
ultra-high vacuum applications’ Internal communication
(2008).
[4] L.E. Murr et al. ‘Metal Fabrication by Additive
Manufacturing Using Laser and Electron Beam Melting
Technologies’ J. Mater. Sci. Technol., 2012, 28(1), 1-14.
MOPB79 Proceedings of IBIC2012, Tsukuba, Japan
ISBN 978-3-95450-119-9
4
Copyright c
○2012 by the respective authors — cc Creative Commons Attribution 3.0 (CC BY 3.0)
Transverse Profiles, Screens & Wires Monitors