Optical Transition Radiation Monitor for the T2K Experiment

2.3 MB·English

by  Sampa Bhadra

Optical Transition Radiation Monitor for the T2K Experiment
S.Bhadrac,M.Cadabeschia,P.dePerioa,V.Galymovc,M.Hartza,c,∗,B.Kirbyc,1,A.Konakab,A.D.Marinoa,2,J.F.
Martina,3,D.Morrisb,L.Stawnyczyc
aUniversityofToronto,DepartmentofPhysics,Toronto,Ontario,Canada
bTRIUMF,Vancouver,BritishColumbia,Canada
cYorkUniversity,DepartmentofPhysicsandAstronomy,Toronto,Ontario,Canada
2
Abstract
1
0
An Optical Transition Radiation monitor has been developed for the proton beam-line of the T2K long base-line
2
neutrinooscillationexperiment. ThemonitoroperatesinthehighlyradioactiveenvironmentinproximitytotheT2K

n target. It uses optical transition radiation, the light emitted from a thin metallic foil when the charged beam passes
a
through it, to form a 2D image of a 30 GeV proton beam. One of its key features is an optical system capable of
J
transportingthelightoveralargedistanceoutoftheharshenvironmentnearthetargettoalowerradiationareawhere
9
itispossibletooperateacameratocapturethislight. Themonitormeasurestheprotonbeampositionandwidthwith

] aprecisionofbetterthan500µm,meetingthephysicsrequirementsoftheT2Kexperiment.
t
e
Keywords: Opticaltransitionradiation,protonmonitor,T2K
d
-
s
n
1. Introduction neutrino beam consists predominantly of muon neutri-
i
s. nos,withaverysmallcomponentofelectronneutrinos.
c The properties of neutrinos continue to puzzle and Near detectors ND280 and INGRID are located on
i
s challenge scientists in spite of tremendous progress the J-PARC site 280 m from the target. The INGRID
y
on both the theoretical and experimental fronts. T2K detector is placed on the axis defined by the proton
h
(Tokai-to-Kamioka) is a long-baseline neutrino exper- beam direction and monitors the neutrino beam direc-
p
[ iment [1] searching for neutrino flavour changes in a tion. ND280islocated2.5◦off-axistoacceptanarrow-
neutrino beam that is generated at the Japan Proton bandneutrinobeamwithpeakenergyaround600MeV
1
Accelerator Research Complex (J-PARC). The main and measures the neutrino energy spectrum and inter-
v
goalistomeasurethelastunknownleptonsectormix- action rates in the unoscillated state. The far detector,
2
2 ing angle θ13 by observing νe appearance from a νµ Super-Kamiokande (SK), 295 km away from J-PARC
9 beam. In addition, from the disappearance measure- and also at 2.5◦ off-axis, studies changes in the beam
1 mentν →ν oscillationparameters∆m2 andsin22θ aftertravel.
. µ τ 32 23
1 willbemeasuredwithaprecisionofδ(∆m2 )∼10−4eV2
0 and δ(sin22θ )∼0.01. The neutrino beam32 at J-PARC 1.1. Motivation for the Optical Transition Radiation
2 23 Monitor
isinitiatedby30GeVprotonsstrikingagraphitetarget
1 A precision measurement of the oscillation parame-
: resulting in many secondary particles, especially pions
v ters depends on an accurate determination of the posi-
andkaons. Afterbeingfocussedbymagnetichorns,the
i tion, profile and angle of the proton beam which pro-
X short-livedpionsandkaonsareallowedtodecayfreely
duces neutrinos. Due to the point-to-parallel nature of
ar insidea100mlongheliumfilledregion. Theresulting thehornfocussingoftheproducedmesons,theoff-axis
angleandthereforetheneutrinospectrumintheSKde-
∗Correspondingauthor tector is affected by the primary proton beam position
Emailaddress:[email protected](M.Hartz) and angle. The position and direction of the proton
1NowatUniversityofBritishColumbia,DepartmentofPhysics beam needs to be measured at the target with a preci-
andAstronomy,Vancouver,BritishColumbia,Canada
sionof1mmand0.5mrad,respectively,inordertoen-
2Now at University of Colorado at Boulder, Department of
surethatthecontributionfromthiseffecttothesystem-
Physics,Boulder,Colorado,U.S.A.
3AlsoatInstituteofParticlePhysics,Canada atic errors of the T2K physics measurements is small.
PreprintsubmittedtoNuclearInstrumentsandMethodsA January11,2012
and transports the OTR light through channels in the
ironandconcreteshielding,throughafusedsilicawin-
dowin theheliumvessel lid, toa camera. Thecamera
and readout electronics sit outside of the target vessel
inanenvironmentthathasaradiationdose5ordersof
magnitudesmallerthantheareanearthefirstmirror.
2. TransitionRadiation
When a charged particle travels between two differ-
ent media, the fields of the charged particle induce a
polarizationonthesurfaceofthenewmedium,andthey
combinecoherentlytoformtransitionradiation.Asdis-
cussedin[7],theformationdepthDfortransitionradi-
ationisoftheorderof
γc
Figure1: ThisdrawingshowstheT2Ktargetarea. TheOTRfoilis D= , (1)
placedbetweenthebeamcollimatorandthefirsthorn. ωp
whereω ,theplasmafrequencyofthemedium(withan
p
electronnumberdensityn ),isdefinedas
The beam size must be measured with a precision of e
about10%tokeepwithinthelimitsrequiredforradia-
4πn e2
tionlossesandtargetprotection. AnOpticalTransition ω2 = e . (2)
p m
Radiation(OTR)monitormeasuresthebeamprofileand e
positionjustupstreamofthetargettotherequiredpre- For example, in solid titanium (ω ≈ 1.34×1016s−1),
p
cisionandisthesubjectofthispaper. with a 30 GeV proton (γ ≈ 32), D is of the order of
Transitionradiationisproducedwhenchargedparti- 1µm. Thereforeonlyathinlayerofmaterialisneeded
clestraverseaboundarybetweenmaterialswithdiffer- toproducetransitionradiation.
entdielectricconstants. Thiswasfirsttheoreticallypre- The general expression for the number of photons,
dicted by Ginsburg and Franck [2] and experimentally N,thatareemittedinafrequencyrangedω,intoasolid
verifiedforvacuum-metalboundariesbyGoldsmithand angledΩ,whenachargedparticlepassesatnormalin-
Jelley[3]. Sincethenthistechniquehasbeenusedata cidencefrommaterial1(withdielectricconstant(cid:15) )to
1
numberofacceleratorsformeasurementsofbeamchar- material2(withdielectricconstant(cid:15) )is[8]
2
acteristics[4],[5],[6]whereingeneralathinfoilisin-
√
troducedinthepathofthebeamandtheresultingOTR d2N 2e2β2 (cid:15) sin2θcos2θ (cid:12)(cid:12) 1 (cid:12)(cid:12)2
= 2 ×(cid:12)(cid:12) (cid:12)(cid:12) ×
lightisrecordedwithacamera. dωdΩ πhcω (cid:12)(1−β2(cid:15) cos2θ)(cid:12)
2
For the T2K experiment, the expected dose near the (cid:12)(cid:12) (cid:12)(cid:12)2
(cid:12) (cid:112) (cid:12)
ttbahereagmedtodsfoeirreec7vt5eion0nkaWitsa0o.pd8eis×rtaat1nio0cn4eSi1svm/5h.r4.pe×Trph1ei0sn8dpiSrcevuc/lhlaurr,detwoshttihhleee (cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(1−β((cid:112)(cid:15)1(cid:15)1−−(cid:15)2(cid:15))2(s1in−2θβ)2(cid:18)(cid:15)(cid:15)21−coβsθ+(cid:15)1(cid:113)−(cid:15)(cid:15)12(cid:15)s2in−2θ(cid:15)2)2sin2θ(cid:19)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12) ,
placement of electronics nearby. A long optical path (3)
is needed to transport OTR light out of the harsh radi-
whereθistheanglebetweentheparticle’svelocityand
ation environment near the target to one more suitable
thephotonvector. Forthecaseofarelativisticcharged
forcameraplacement.
particle (β ∼ 1) moving from a material with |(cid:15) | > 1,
Fig. 1 shows a side view of the T2K target area. A 1
intovacuum((cid:15) =1),thisreducesto[9]
beamcollimator,threefocusinghorns,thetarget(inside 2
the first horn), concrete and iron shielding and the de- d2N 2e2 θ2
cayvolumeareallenclosedinalargevesselfilledwith dωdΩ = πhcω × (θ2+γ−2)2. (4)
helium. The collimator, target and horns are mounted
underlarge,steel-framedsupportmodules. Thesupport Notethatthelightisemittedinanarrowforwardcone,
modulesandtheshieldingsitonmountsattachedtothe with the maximum of the angular distribution at θ ∼
sides of the helium vessel. A series of mirrors focuses 1/γ. Forthecaseofaparticletravellingthroughathin
2
Reflection surface. The light diverges from the foil to mirror 1,
axis
travelsasaparallelbeamtomirror2andcomestoanin-
Backward light termediatefocushalfwaybetweenmirrors2and3(see
Proton target
Foil Fig.3).Thispatternrepeatsusingmirrors3and4before
thefinalfocusatthecameraposition.A25cmdiameter,
fused silica window in the aluminum lid of the helium
Proton
beam vesselallowstheOTRlighttoemergeforcaptureatthe
45deg camerasituatedontopofthelid.
Forward light
It is desirable to have the first mirror as far away as
possiblefromthefoiltoreduceradiationexposure,and
Figure2:Thisfiguresillustratesthedirectionoftheforwardandback-
alsotheaperturesizemustbelargeenoughtocollecta
wardOTRlightfromafoilorientedat45degreeswithrespecttothe
beam. large fraction of the light. The mirror diameter is lim-
itedbythemaximumallowablesizeforthechannelsin
the shielding. Given these considerations, mirror 1 is
foil, OTR is also produced in the backward direction
placed 110 cm from the foil (requiring an effective fo-
when the particle enters the foil from vacuum. In this
cal length of 110 cm) and has a diameter of 12 cm.
case, (cid:15) = 1 and (cid:15) = (cid:15). Equation 3, with β ∼ 1, now
1 2 Mirrors2and3havethesamefocallengthandsizeas
reducesto
mirror 1. However mirror 4 has a shorter focal length
ddω2dNΩ = π2hec2ω ×(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)√√(cid:15)(cid:15)−+11(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)2× (θ2+θ2γ−2)2. (5) othfe3s0izcemof(etffheecftoivileimfoacgaellaetntghtehcoafm6e0racmby).4T5h%istroedaullcoews
theimageofthe5cmdiameterfoiltofitwithina4cm
Thegeometrychangesslightlyforthecasewherethe diameterfibertaperconnectedtothefaceofthecamera.
particle is not at normal incidence to the surface. The Thefibertaperreducestheimagediameterfrom4cmto
forward lobe, produced by the charged particle exiting 1.1cmtofitontothesizeofthecamerasensor.
thefoil,isstillorientedaroundthelineofmotionofthe
charged particle. However, the backward, or reflected
60 cm
lobe,nowsurroundstheaxisofreflectionfromthefoil Fiber taper
(4 cm diameter)
surface. For a thin foil oriented at 45 degrees with re-
Mirror 4
specttoabeamofchargedparticles,thebackwardlobe Radiation
tolerant camera
will be reflected at 90 degrees from the original beam
direction, as shown in Fig. 2. It is this backward lobe He vessel lid Quartz window
thatwetransportanddetectintheT2KOTRmonitor.
Concrete shielding
3. OpticalSystem 110 cm
Mirror 2 Mirror 3
3.1. Overview ~7m
Fig. 3 shows the optical layout of the OTR system. 110 cm
The beam strikes a foil oriented at 45 degrees with re-
spect to the beam line. As described in the previous
section,thebackwardlobeoftheOTRisemittedaround
Iron shielding
thereflectionaxis,inthiscase90degreesrelativetothe
incident proton beam direction. The foil sits immedi-
Foil (5 cm diameter)
ately upstream of the target, and is downstream of the
Mirror 1 Proton beam
beamcollimator. into the page
The OTR light must travel through several bends in 110 cm
theshieldingtoavoidadirectpathfortheradiationfrom
thetargetregion. Aseriesof4parabolicmirrorstrans- Figure3:ThisfigureshowsaslicethroughtheopticalpathoftheOTR
portthelightthroughthispath. Theseparabolicmirrors systemwheretheprotonbeamisgoingintothepageandstrikingthe
are90degreesoff-axis,givinganeffectivefocallength foil.Threelightraysillustratethefocussingpropertiesoftheoptics.
(the distance from the centre of the mirror to the focal
point) twice the focal distance of the parent parabolic
3
3.2. Mirrors 50
Since there are four mirrors in the system, a mirror
s 40
surfacewithhighreflectivityisrequiredtoreducelight nt
u
loss. The parabolic mirrors used in this system were Co
fabricatedbyB-ConEngineeringInc. outofsolidalu- C 30
D
minumandcoatedwithauniform400nmthicklayerof A
Al2O3,whichhasareflectivitycloseto100%. Asmall ured 20
testmirrorwiththiscoatingwasirradiatedwithaproton as
e
M 10
beamatTRIUMFfortheequivalentradiationdosethat
isexpectedfor130yearsofoperationatthelocationof
mirror 1. No significant change in the reflective prop- 00 10 20 30 40 50
erties of the test mirror was observed. This provided Expected ADC Counts
confidence in the long term reflective properties using
thiscoating. Figure4:ThemeasuredcameraresponseinADCcountsisshownasa
functionoftheexpectedADCcountsassuminglinearoperation.The
responseisshownherebeforetheintroductionoftheambientlight
3.3. Camera
andisaveragedoverthepixels.
The radiation dose level at the camera is estimated
to be ∼1 kGy/year at 750 kW beam power. A charge
injection device (CID) camera (Thermo Fisher Scien-
Eq. 5. Based on ray tracing simulations, the light col-
tific 8710D1M) is used as it is radiation tolerant up to
lection efficiency for OTR light emitted at the centre
10kGy. Thesensorpixelmatrixconsistsof755×484
of the foil with perfectly reflective mirrors is approxi-
sites with pixel dimensions of 12.0 µm × 13.7 µm.
mately15%.Sincelightthatreachesthecamerareflects
An analog signal is sent from the camera over 50 m
off four aluminum mirrors, the generated OTR photon
of shielded cable to the data acquisition system at the
spectrummustalsobemultipliedbythewavelengthde-
groundlevelofthetargetbuilding.
pendentreflectioncoefficientforaluminumraisedtothe
Duringtheinitialoperationoftheneutrinobeam-line,
4thpower. Aftermirror4,thephotonspassthroughthe
the low intensity of the delivered proton beam led to
fiber taper, which has a transmittance of roughly 45%,
smallOTRlightyields. Attheselightlevelsthecamera
tostrikethephotocathodeofthecamera.
hasanon-linearresponseasshowninFig.4. Thiswas
A combining of all the acceptances and efficiencies
understoodtobecausedbyimpuritiesinthesiliconthat
(includingthequantumefficiencyofthecamerasensor,
trappedthecollectedcharge. Byaddinguniformambi-
which peaks at 30% at 570 nm) yields the number of
entlightthesetrapscouldbepopulatedandtheresponse
electrons recorded by the CID camera as a function of
ofthecameracouldbemovedintothelinearregion.
the wavelength of the initial OTR photons. For a tita-
Thecameraoutputsaninterlacedvideosignalwhere
niumalloyfoil,integratingoverthevisiblerangeresults
evenlinesarereadout∼17msecbeforeoddlines. The
in ∼ 2.1×10−5 electrons in the camera (∼ 2.1×10−8
camera pixels collect charge at the arrival of the OTR
ADCcounts)per30GeVproton.
light, immediately before the readout of the first even
During beam commissioning and early data taking
line. Duringthetimebetweenthechargecollectionand
periods, the number of protons per pulse varied from
itsreadout,apixelmaylosechargethroughleakagecur-
rents. Tostudythiseffect,theratioofneighboringeven 1×1011to1×1013,equivalentto0.13to13ADCcounts
at the pixel sampling the peak of the light distribution.
andoddpixelsistakenforimagesofOTRlight. Acon-
Atthelowerendofthisrange,theOTRlightyieldfrom
stantratioof0.81isobserved,indicatinganexponential
decay of the charge with time constant τ = 77 msec. titanium is too low to image. An aluminum foil gives
2−4timesmorelight,dependingonthesurfacerough-
The measured ADC count for each pixel is corrected
ness, due to higher reflectivity. The light yield from
for this charge decay based on the readout time of the
aluminum is, however, still insufficient for very low
pixel.
beamintensities,soaceramicAl O :Cr3+100µmthick
2 3
wafer producing fluorescent light is used for very low
3.4. PredictedPhotonYield
intensitybeamrunning. ThisisaDemarquestAF995R
The number of generated OTR photons for a mate- wafermachinedbytheNikiGlassCompany. Thefluo-
rial with dielectric constant (cid:15) can be estimated from rescenceisproportionaltotheenergydepositedbythe
4
proton beam (apart from nonlinear effects discussed in 4.1. TheFoilDiskSystem
Section7.1). Accountingforopticalandquantumeffi-
ciencies,thecameracollects6.0×109 electronsper30 TheOTRfoilsaremountedonadiskcarouselwhich
GeVproton,sufficienttoimagethebeamatverylowin- haseightholepositionsofdiameter50mm,asshownin
tensities. Aremotelycontrolleddiskofneutraldensity Fig.6. Withthediskmountedat45degreestothebeam
filtersinfrontofthecameracanberotatedtooptimize axis, the foils cover the full beam acceptance defined
thelightintensityforanygivenbeamintensity,andthis by the 30 mm diameter hole in the upstream collima-
feature will be particularly useful as we move to high tor. Seven hole positions are occupied as described in
beamintensitieswheresaturationatthecameracanoc- Table1andtheeighthpositionisempty. Twopositions
cur. are filled with a ceramic wafer and an aluminum foil,
asdiscussedinSection3.4. Havingfourtitaniumalloy
foils allows for replacement in case of foil damage at
3.5. PrototypeSystem
highintensity.Afifthtitaniumalloyfoilhasapatternof
A prototype system was assembled to demonstrate preciselylaser-machinedholesandisusedonlyforcal-
thedetector’sabilitytoobservetransitionradiationand ibration when there is no beam. Titanium was chosen
measurethepositionandwidthofaparticlebeam. The after various foil materials were studied using the pro-
prototype consisted of four parabolic mirrors with fo- gramsMARS[10,11,12,13]andFemLab,inaddition
cal lengths and relative distances scaled to 13.8% of toapproximatetheoreticalcalculations. Althoughtem-
the full system size, test foils held by a fixed support peratureriseisnotaproblemforseveralmaterials,only
and a charge-coupled device (CCD) type photosensor titaniumalloyshavesufficientyieldstrengthtosurvive
readoutdirectlytoacomputer. Thesystemwastested thestressesofthefull-intensityT2Kbeam,withasafety
in a NRC (Ottawa) electron linear accelerator capable factorofaboutfour. Basedonourstudies,thetitanium
of producing a continuous electron beam with similar alloy 15V-3Cr-3Sn-3Al (the numbers refer to the per-
Lorentz factor γ to the J-PARC proton beam, resulting centagesoftheelementsinthealloycomposition)was
in similar transition radiation angular and spectral dis- chosenfortheOTRfoils.
tributions. Electron beam induced transition radiation Each metal foil is stretched by a clamping ring (see
wasobservedfortestfoilscomposedoftitanium-alloy, Fig.7)withamachinedridgewhichforcesthefoiledge
aluminum and graphite, with calibration holes placed into a corresponding circular groove on the disk. The
inthetitanium-alloyfoilalsovisible. Inordertoiden- sizeandshapeoftheridgeandgrooveweredetermined
tifypotentialbackgroundlightsourcestheJ-PARCtar- byiterativedesignandtestinginordertoprovidethere-
getstationenvironmentwassimulatedbyenclosingthe quiredtension. Finiteelementanalysisindicatedthata
prototypeinansealedbagfilledwithhelium, atwhich stress of somewhat less than 100 MPa will result from
pointnoadditionalbeam-inducedlightwasobserved. theexpectedmaximumintensitybeamof3.3×1014pro-
Theprototypesystembeamwidthandpositionmea- tonsperpulse. Inordertokeepthefoilsflat, atension
surement resolutions were estimated to be 15% and stresscomfortablygreaterthanthisisrequired. Thefi-
0.2 mm, respectively, by comparing the CCD camera nalclampingmechanismproducesameasuredstressof
results to the NRC wire detector measurements of the 190MPa.
beam. This provided confidence to proceed to the de- Thediskismountedonan“arm”heldbytwo“legs”
signandconstructionofthefull-scalesystem. (seeFig.5)thatareattachedprecisely(dowel-pinned)to
alargealuminumplatewhichispartofthetarget/horn
assembly. The disk, arm and legs are made of tita-
4. MechanicalDesignandSet-up nium, chosen for its low coefficient of thermal expan-
sion. Since the aluminum plate is actively cooled with
The OTR mechanical systems are shown schemati- helium but the arm is not, relative movement is thus
callyinFig.5(a)and 5(b). Theyweredesignedtosat- minimized. Thearmtemperatureismeasuredbyather-
isfy several requirements arising from the inaccessible mocouple. At the highest beam power reached so far,
high-radiationenvironmentnearthebeam: 145kW,thetemperatureriseis8◦C.
Anyofthefoilscanbepositionedinthebeambyro-
• theabilitytocontinuallycalibratetheoptics tating the disk. It is important that the foil is centered
• stabilitywithtemperature onthehornaxissothatthebeamdoesnotpassthrough
• long-termrobustness the thicker disk material between foils. The position
• easeofremotemaintenanceforpartreplacement. of the calibration foil is particularly important, since
5
Mirror 2
Mirror 3
Front plate of
TargetCooling horn support
Pipes module
Mirror1 Beam
Mirror 1
Target
Legs Target
Beam OTRDisk
withFoils Arm
(a) TheOTRcomponentsnearthebeam. (b) AviewoftheOTRsystemfromtherear,showingcomponents
mountedonthefrontplateofthehornsupportmodule.
Figure5:TheOTRsystemcomponents
Table1:FoilsusedintheOTRsystem
Material(numberoffoils) Thickness(µm) Operation
AF995R(1) 100 <1kWbeampower
Al1100(1) 1−40kWbeampower
Ti15-3-3-3(4) 50 >8kWbeampower
Ti15-3-3-3(1) calibrationwithnobeam
images of the calibration foil, taken periodically with
back-lighting,providethepositionofthenominalbeam
lineonthecamerapixelmatrix. Afterfinalinstallation
and alignment of the arm and disk, the calibration foil Titanium Calibration foil
wasrotatedintothebeampositionandsurveyedwitha foils
theodolite.Thecentralcalibrationholepositionwithre-
specttothenominalbeamlinealongthehornaxiswas
measuredwith0.3mmprecision. Helium
nozzle
The disk-rotation motor system is positioned 1.5 m
underneath the helium vessel lid, above the iron and
concrete shielding shown in Fig. 3. A long rigid steel
shaft couples to a flexible steel shaft that follows a 90
degreebendtothearmsupportingthefoildisk.Aspline Ceramic wafer Aluminum foil
couplingismadeattheendofthearmtoanotherflexi-
bleshaftwhichrunsalongthearmandconnectstothe
disk. The motor is connected to the shaft through a Figure6:Thefoildisk(upstreamside).
100:1 gearbox, so that the disk rotates slowly. Access
formaintenanceorreplacementofthemotorsystemis
possiblethroughthewindowinthelid.
6
Alignment Dowel
Helium Line Filament and
Thermocouple
Connectors
Plunger Helium Line
Arm Reflector
Motor Shaft
Pivot Point Connector
Microswitch
Shaft Coupling
Figure7:Thefoildisk(downstreamside).
Figure8:Theoutsideendofthearm.
The foil position can be determined in principle by
custom nozzle (see Fig. 6) to blow across the foil sur-
counting the number of motor steps using its encoder.
face.
However, due to backlash in the flexible shaft system,
In the event that a horn or the target require mainte-
alternate methods of ensuring precise foil positioning
nance,thediskand/orarmcanberemovedorreplaced
are necessary. The primary method to change foil po-
sitions is to run the motor until it is turned off by a byremotemanipulators. Toremovethearm,thespline
couplingofthemotorshaft,theheliumlineandthece-
micro-switch engaged by a machined titanium button
ramic connectors for the micro-switch and thermocou-
on the disk. This type of micro-switch (model 6302-
plemustfirstbedisconnectedusingthemanipulators.A
16 from Haydon), designed for extreme conditions, is
pictureoftheendofthearmisshowninFig.8. Forthe
used in high radiation environments at TRIUMF. The
T2Ktargetreplacementthearmcanberotateddownby
switchpositionwasadjustedsothatitengagesthebut-
about45◦ onapivotbearingontheoutsidelegwithout
ton just before a foil is in the correct position. A steel
disconnectingtheelectrical,heliumandsplineconnec-
ball-bearingplungermechanism, spring-loadedagainst
tions. Usingamechanismontheinnerlegthearmcan
the surface of the disk, then falls into a matching ma-
belockedbackintothecorrecthorizontalposition.
chineddepressioninthedisk,whichlocksitfirmlyinto
the correct position. Each foil position has a corre-
4.2. TheMechanicalDesignoftheOpticalSystem
sponding machined depression. Most of these features
can be seen in Fig. 7. When the motor is turned on Mirrors1and2alongtheOTRlightpath(seeFig.3
again to move to the next foil position, the torque of and also Fig. 5(b)) are mounted at either end of a long
themotorandshaftsystemissufficienttostartrotating steel tube and mirror 3 is at the bottom end of a sec-
thediskandbringtheballbearingoutofthedepression ond shorter steel tube. Some of the mirror tube de-
againstthespringforce. Thefoilpositionrepeatswith tails are shown schematically in Fig. 9. Each mirror
thismethodto0.1mmprecision. canberotationallyadjustedabouttwoaxes. Thetubes
are mounted precisely on ball mounts protruding from
Abackupsystemusesapressurizedheliumgasline,
flangesattachedtothebackofthefrontplateofthesup-
which ends in a brass tube with an end face parallel to
port module. They can be lifted out and replaced by
thebackofthediskabout0.1mmfromthesurface(see
crane from the lid of the helium vessel through ports
Fig. 7). As the disk rotates the pressure is maintained
directlyabovethetubes.
untilthecorrectfoilpositionisreached,atwhichpoint
thetubeendencountersaholethroughthedisk,reduc-
4.3. TheCalibrationLightingSystems
ing the pressure and causing a pressure switch located
outsidetheshieldingtoturnoffthemotor. Thismethod Therearethreelightingsystemsusedforcalibration
is precise only to ∼2 mm, but sufficient to position a by lighting the calibration foil from behind. Two of
foilinthebeaminthecaseoffailureofboththemicro- them use small red LED lasers (Sanyo DL3147-060,
switchandplunger. Thissameheliumlinecanbeused 650nm,7mW),onemountedjustabovethefusedsilica
toremoveanyaccumulateddustfromthefoil,sincethe window(outsidelaser)inthelidandtheothermounted
gas passingthrough thehole inthe diskis guided bya near the motor which drives the disk rotation (inside
7
Anchor point for
lifting crane
Mirror 2
Helium Line
Guiding sphere
for mirror 1
and indentation Filaments
cleaning
Inside laser
reflector
Front Plate
Bracket
Figure9:Mirrortubesystem.
laser). There are two internally electro-polished steel
tubes (10 and 13 mm diameter) which guide the laser Outside laser
light down to the region at the bottom of the tube ex- reflector
tendingfromthefrontplateofthehornsupportmodule, Motor
seen on the right side of Fig. 5(a). Two small steel re- Flexible Shaft
flectors guide the laser light to another reflector on the
arm, which can be seen in Fig. 8. The third system is
asetoffilamentlights(3forredundancy),custombuilt Figure10: Apicturefrombelowshowingthefilamentlights, laser
reflectors and coupling of the vertical to flexible shafts of the disk
withAlchromewirecoilsinparabolicreflectors. They
rotationsystem.
are installed in the same region as the laser reflectors
and also point at the arm reflector. Fig. 10 shows a
picture of this region. The filament coil is operated at the empty foil position onto the reflector and up to the
12 A current, causing the wire coil to glow with suffi- filamentlampregion(seeFig.11(b)).
cientlightoutput.
Thefilamentlightsarenotaccessibleafterthebeam 4.5. SpareOTRSystem
has been running, and they cannot be replaced by the In order to be prepared for possible future problems
remotemanipulators. Theoutsidelasercanbereplaced with the target region of T2K, a spare horn, support
wheneverthebeamisoffforabriefperiod, butthein- module, target and OTR system have been built. The
sidelasernearthemotorcanonlybereplacedinalong spareOTRsystemhasbeeninstalled,aligned,andcali-
shutdownwhentheheliumvessellidhasbeenremoved. bratedonthesparehornandsupportmodule.
4.4. Alignment
5. DataAcquisitionandSlowControl
Theopticalsystemwasalignedusingalaser. Aspe-
cial mount was made to attach the laser in front of the 5.1. OTRDAQsystem
empty foil position in the disk. The laser was aligned The main functions of the data acquisition system
topointalongthebeamlinethroughthehorn. Aplane (DAQ)aretotrigger,collectandprocesstheimagedata.
mirror was then attached parallel to the disk to reflect The components of the DAQ system are illustrated
the laser light at 90 degrees along the OTR light path in Fig. 12. The trigger signals and the image acquisi-
(see Fig. 11(a)). The mirrors were then adjusted one tion controls are handled by a FPGA chip located on
by one so that the laser light followed the correct cen- a frame-grabber board which interfaces to a host DAQ
tralpaththroughtheopticalsystemtothecamera. This computer via a PCI bus. Also located on the board
was greatly aided by small marks made the center of is a TriMedia TM1302 digital signal processor (DSP)
each parabolic mirror during their manufacture. The whichisresponsibleforthetransferofthedigitizedim-
arm reflectorwas aligned byshining a laserback from age frames to the host computer. Configuration of the
thecamerapositionthroughtheopticalsystem,through FPGAregistersisdoneviaTriMediasoftware.
8
Camera Mirror 4
Mirror 4
Laser Inside and
outside
laser guides
Mirror 2 Mirror 3
Mirror 3 Mirror 2
Filament
lights
Flat mirror mounted
on the foil disk
Mirror 1 Mirror 1 Reflector on
the arm
Laser
(a) Thisdiagramshowsthealigmentprocedureforthemirrors.(b) Thisdiagramshowsthealigmentprocedureforthearmre-
flector.
Figure11:FiguresillustratingthealigmentoftheOTRsystemcomponents.
Figure12:OTRmonitorslowcontrolandDAQsystem.
9
MR trig -100 msec the upstream proton beam monitors are used to deter-
mine whether the beam position and width are within
frame_reset_delay tolerance. Iftheyareoutsidetolerance, anabortsignal
frame reset issenttopreventfurtherextractionfromthemainaccel-
eratorringtotheneutrinobeam-line.
capture_delay
TMinterrupt 5.2. SlowControl
The OTR monitor has a number of remotely con-
nutrig-20usec 100msec-20usec trolledmotorstomovedifferentcomponentsofthede-
tector. In addition, the lighting system forthe periodic
calibrationoftheopticsrequiresanumberofpowersup-
plies.Theseelementsarepartoftheslowcontrolsystem
Figure13:Imageacquisitionsignals.
(Fig.12)whichismanagedusingMIDAS.
The hardware control of the disk, filter wheel and
camera stage motors is done using a Galil motor con-
Theanalogvideosignalfromthecameraisacquired trol module. The unit also collects the status from the
anddigitizedbytheframe-grabber. Fig.13showstim- pressure sensor and the micro-switches on the motors.
ingsignalsimportantfortheDAQoperation. Theread- The state of the disk micro-switch is constantly moni-
out cycle is initiated with a pre-trigger which arrives tored. Toavoidpossibleprotonbeamextractionduring
100 msec before a possible proton beam extraction to the foil disk rotation or when the disk is not properly
theneutrinobeam-line. Followingitsarrivalafterapro- positioned,abeaminterlocksignalisgeneratedwhenit
grammabledelay(frame_reset_delay)aframereset isnotengaged.
signalisissuedtothecamerawhichsynchronizesitand
the readout circuitry to the expected spill arrival time.
6. ImageCorrectionandAnalysis
The trigger signaling the beam extraction to the neu-
trino beam-line arrives 20 µsec before the spill. Af- 6.1. EfficiencyCorrection
ter it is received and while the camera data are ac-
Ray tracing simulations of the optical system re-
quired and digitized by the frame-grabber an interrupt
vealed that within ±15 mm of the foil center, the rela-
signal(TMinterrupt)issenttotheDSP.Thetimingof tivelightcollectionefficiencyvariesbymorethan50%
the interrupt relative to the frame reset is configurable
(Fig. 15(a)). This introduces a bias into the recon-
(capture_delay). The interrupt signal informs the
structedbeampositionandwidth,sincethevariationof
DSPthatthenextavailableframewillcontainthespill thelightcollectionefficiencyissignificantoverthesize
data and should be moved from the internal memory
ofthebeamspot.
bufferstoadedicatedmemoryaddressonthehostcom-
Anintegratingspherewasusedtoprovideauniform
puter. Aftermovingthespillimage,theDSPalsotrans-
lightsourceinordertomeasurethelightcollectionef-
fers the image data from the two subsequent frames.
ficiency of the system. This hollow cardboard sphere,
Theseimagesarelaterusedforthepedestalsubtraction
shown in Fig. 16, is 30.5 cm in diameter and painted
(seeSection6.3).
whiteontheinside. A12cmopeningatonepoleofthe
Once the images are copied to the memory of the sphereislitinternallybyanoff-equatorialringof8laser
hostcomputer,aMIDAS-based[14],front-endapplica- diodes. Thereisnodirectpathbetweenthediodesand
tioncompressesandsendstheimagedatatoadedicated theopening,soallofthelightexitingthesphereresults
eventserverthathandlesthedistributionofthemonitor from diffuse reflections off the rough inner surface of
dataforonlineanalysisandarchiving. thesphere. Thisinexpensivedevicegivesexcellentper-
The OTR online monitoring program displaying an formance, with light output measured to be uniform in
event from one of the first proton beam extractions to intensity within 5% across the entire opening. Prior to
the neutrino beam-line at J-PARC is shown in Fig. 14. theinstallationofthehornmoduleintotheheliumves-
The center of the T2K target is marked by a yellow sel, the sphere was positioned at the foil location with
cross-hair in the middle panel. The two panels on the theopeningfacingmirror1.Imagesofthelightthrough
rightshowinthehorizontalandverticalprojectionsthe the optical system were taken with the camera system,
results of a 2D fit that extracts the beam position and asshowninFig.15(b),andareusedtocorrectforeffi-
width. Thesevaluesalongwiththemeasurementsfrom ciency.
10