the arcade 2 instrument

12 Pages·2011·1.51 MB·English

TheAstrophysicalJournal,730:138(12pp),2011April1 doi:10.1088/0004-637X/730/2/138
(cid:2)C2011.TheAmericanAstronomicalSociety.Allrightsreserved.PrintedintheU.S.A.
THEARCADE2INSTRUMENT
J.Singal1,D.J.Fixsen2,3,A.Kogut3,S.Levin4,M.Limon5,P.Lubin6,P.Mirel3,7,M.Seiffert4,T.Villela8,
E.Wollack3,andC.A.Wuensche8
1KavliInstituteforParticleAstrophysicsandCosmology,SLACNationalAcceleratorLaboratoryandStanfordUniversity,
MenloPark,CA94025,USA;[email protected]
2DepartmentofPhysics,UniversityofMaryland,Code665,NASAGoddardSpaceFlightCenter,Greenbelt,MD20771,USA
3Code665,NASAGoddardSpaceFlightCenter,Greenbelt,MD20771,USA
4JetPropulsionLaboratory,4800OakDrive,Pasadena,CA91109,USA
5DepartmentofPhysics,ColumbiaUniversity,1027PupinHall,Box47,NewYork,NY10027,USA
6DepartmentofPhysics,UniversityofCalifornia,SantaBarbara,CA93106,USA
7WyleInformationSystems,Code665,NASAGoddardSpaceFlightCenter,Greenbelt,MD20771,USA
8InstitutoNacionaldePesquisasEspaciais,Divisa˜odeAstrof´ısica,CaixaPostal515,12245-970-Sa˜oJose´dosCampos,SP,Brazil
Received2008December19;accepted2011January25;published2011March15
ABSTRACT
The second generation Absolute Radiometer for Cosmology, Astrophysics, and Diffuse Emission (ARCADE 2)
instrument is a balloon-borne experiment to measure the radiometric temperature of the cosmic microwave
background and Galactic and extragalactic emission at six frequencies from 3 to 90 GHz. ARCADE 2 utilizes
adouble-nulleddesignwhereemissionfromtheskyiscomparedtothatfromanexternalcryogenicfull-aperture
blackbodycalibratorbycryogenicswitchingradiometerscontaininginternalblackbodyreferenceloads.Inorderto
furtherminimizesourcesofsystematicerror,ARCADE2featuresacoldfullyopenaperturewithallradiometrically
active components maintained at near 2.7 K without windows or other warm objects, achieved through a novel
thermaldesign.WediscussthedesignandperformanceoftheARCADE2instrumentinits2005and2006flights.
Keywords: cosmicbackgroundradiation–instrumentation:detectors–radiocontinuum:galaxies
Online-onlymaterial:colorfigures
observedin2003(Fixsenetal.2004)andisdescribedbyKogut
1.INTRODUCTION
etal.(2004a).Followingverificationoftheinstrumentconcepts,
Absolute Radiometer for Cosmology, Astrophysics, and the full six frequency ARCADE 2 instrument, with observing
Diffuse Emission (ARCADE 2) is part of a long-term effort channelsrangingfrom3to90GHzwasdesignedandbuilt,and
to characterize the cosmic microwave background (CMB) and observedin2005and2006.Scientificanalysisofthe2005flight
Galactic and extragalactic microwave emission at centimeter is presented by Singal et al. (2006), and scientific analysis of
wavelengths. The background spectrum has been shown to be the2006flightwillbepresentedinforthcomingpapers.
a nearly ideal blackbody from ∼60 to ∼600 GHz with a tem- This paper describes the design and performance of
peratureof2.725±0.001K(Fixsen&Mather2002).Atlower ARCADE2.Wepresenttheradiometricandthermalproperties
frequencies,however,wheredeviationsinthespectrumareex- of the instrument, and also describe elements of engineering
pected,existingmeasurementshaveuncertaintiesrangingfrom designthatmaybeuseful.
10mKat10GHzto140mKat2GHz(seeFixsenetal.2004
forarecentreview).Theseuncertaintiesareprimarilytheresult
2.INSTRUMENTDESIGN
ofinstrumentationsystematics,withallpreviousmeasurement
programsbelow60GHzneedingsignificantcorrectionsforin- ARCADE2reducessystematicerrorsthroughacombination
strumentemission,atmosphericemission,orboth.Muehlner& of radiometer design and thermal engineering. The instrument
Weiss (1970), Johnson & Wilkinson (1987), and Staggs et al. coreiscontainedwithinalarge(1.5mdiameter,2.4mtall)open
(1996)areexamplesofnotableCMBabsolutetemperaturemea- bucket liquid helium dewar. Maintaining such a large volume
surementsfromhigh-altitudeballoons. and mass at cryogenic temperatures in an open environment
Toimproveexistingradiometrictemperaturemeasurements, without significant atmospheric condensation presents consid-
aninstrumentmustbefullycryogenic,sothatmicrowaveemis- erableinstrumentalchallenges.Theexternalcalibrator,aperture,
sionfromfront-endcomponentsoftheinstrumentisnegligible. antennas,andradiometersaremaintainedattemperaturesnear
Inanyabsoluteradiometrictemperaturemeasurement,radiation 2.7 K through the use of liquid helium tanks fed from the he-
fromthesourcebeingmeasurediscomparedbytheradiometer liumbathatthebottomofthedewarbyanetworkofsuperfluid
tothatfromablackbodyemitterofknowntemperature.Because pumps.Boil-offheliumgasisusedfortheinitialcool-downof
ofdriftsinthegainoftheradiometer,acomparisonwithinthe components on ascent and directed in flight to discourage the
radiometer to another blackbody emitter of known tempera- condensationofambientnitrogenontheaperture.
ture is necessary. Therefore, in order to achieve significantly Figure 1 shows an overview of the ARCADE 2 instrument.
lower uncertainties than previous measurements, ARCADE 2 The corrugated horn antennas for each frequency band hang
isahigh-altitudeballoon-baseddouble-nulledinstrumentwith fromaflathorizontalaluminumapertureplateatthetopofthe
open-aperture cryogenic optics mounted at the top of an open opendewar.Therearesevenobservingchannels,oneeachat3,
bucketliquidheliumdewar. 5,8,10,30,and90GHz,andanadditionalchannelat30GHz
A previous generation instrument with observing channels with a much narrower antenna beam to provide a cross-check
at 10 and 30 GHz, built to test the cold open-aperture design, on emission in the antenna sidelobes. Cryogenic temperatures
1
TheAstrophysicalJournal,730:138(12pp),2011April1 Singaletal.
Figure2.Photographofthecarouselbeingloweredontopoftheapertureplate
bytwooftheauthors,showingtheportholeforskyviewingandtheexternal
calibrator.Thecarouselturnsatoptheapertureplatetoexposegroupsofhorns
totheskyortotheexternalcalibrator.Theexternalcalibratorisablackbody
emitterconsistingof298conesofSteelcastabsorbercastontoaluminumcores,
andtheradiometricsideisvisibleinthephotograph.Hornantennasinthecore
arevisiblehangingdownfromtheapertureplate.
(Acolorversionofthisfigureisavailableintheonlinejournal.)
Figure 1. ARCADE 2 instrument schematic, components not to scale.
Cryogenic radiometers compare the sky to an external blackbody calibrator.
Theantennasandexternalcalibratoraremaintainednear2.7Katthemouthof
anopenbucketdewar;therearenowindowsorotherwarmobjectsbetweenthe the “high bands,” occupying the third. All horns except the
antennaandthesky.Coldtemperaturesaremaintainedatthetopofthedewar 3GHzpoint30◦ fromzenithinonedirection,whilethe3GHz
viaboil-offheliumgasandtanksfilledwithliquidheliumfedbysuperfluid points 30◦ from zenith in the opposite direction. The sky port
pumpsinthebath.Forobservingthesky,everythingshownissuspendedbelow
ahighaltitudeballoon. inthecarouselissurroundedbyreflectivestainlesssteelflares
which shield the edge of the antenna beams from instrument
contaminationanddirectboil-offheliumgasoutoftheportto
within the external calibrator and internal reference loads, as
discouragenitrogencondensationinthehornaperture.Figure2
wellasoncomponentsthroughouttheinstrumentcore,areread
shows a photograph of the carousel being lowered onto the
withruthenium-oxideresistancethermometers.
apertureplate,withtheskyportandtheradiometricsideofthe
A carousel structure containing both a port for sky viewing
external calibrator visible. Figure 3 shows the arrangement of
and the external calibrator sits atop the aperture plate and
thehornapertures.
turns about a central axis to alternately expose the horns to
The carousel is supported 1.5 mm above the aperture plate
eithertheskyorthecalibrator.Theopenportholeandexternal
calibrator are both ellipses measuring 700 mm × 610 mm. bywheelbearingsontheedgeandaKynar(cid:2)c plasticbearingin
the center. It is turned with a motor and chain drive, with the
Each radiometer measures the difference in emitted power
motormountedoutsideofthedewar.Themotoriscommanded
betweenradiationincidentonthehornandthatfromaninternal
torunandstops,withtheengagementofalogicswitch,when
blackbody reference load. The experiment performs a doubly
theproperalignmentofthecarouselrelativetotheapertureplate
nulledmeasurement,withtheradiometrictemperatureofthesky
isreached,thusensuringaccurateandrepeatablepositioningof
comparedtothephysicaltemperatureoftheexternalcalibrator
thecarousel.Therearethreecarouselstoppingpositionsforsky
in order to eliminate systematic effects within the radiometer
andexternalcalibratorviewing,whichare,incounterclockwise
to first order. The radiometer itself differences the internal
order, (1) the “5 and 8 sky” position, where the 5 and 8 GHz
reference load from the horn signal allowing a determination
horns view the sky and the 3 GHz horn views the external
ofthecouplingofradiometeroutputtoinstrumenttemperatures
calibrator, (2) the “3 sky” position in which the 3 GHz horn
andanear-nullingoftheradiometeroutputtoreducetheeffects
viewstheskyandthehighbandhornsviewthecalibrator,and
ofgainfluctuations.
(3)the“highsky”positionwherethehighbandhornsviewthe
The ARCADE 2 instrument is flown on a high-altitude
sky and the 5 and 8 GHz horns view the external calibrator.
790 Ml balloon to the upper atmosphere (37 km altitude)
The remaining group of horns viewing neither the sky nor the
in order to reduce atmospheric emission and contamination
external calibrator in any of these positions are viewing a flat
fromterrestrialmicrowavesourcestonegligiblelevels.Balloon
metal plate of the carousel. A fourth stopping position is used
launch and recovery operations are handled by the Columbia
Scientific Balloon Facility (CSBF) in Palestine, TX (31◦.8 lat, on ascent to align a vent hole on the carousel with one on the
−95◦.7long). apertureplate,therebychannelingboil-offgastocooltheback
oftheexternalcalibrator.
Carouselalignmentateachofthefourpositionsisrepeatable
2.1.AntennaApertureandCarouselConfiguration
towithin1mmontheoutsidecircumference.Theskyportand
Thecorrugatedhornantennasarearrayedontheapertureplate externalcalibratoreachhavenearlythesameareaandshapeas
inthreeclusters,withthe3GHzhornoccupyingone,the5and theapertureofthelargesthornaperture(3GHz).Misalignment
8 GHz horns occupying a second, and the remaining horns, ofmorethan1cmwouldallowasectionoftheapertureplateand
2
TheAstrophysicalJournal,730:138(12pp),2011April1 Singaletal.
Figure4.ARCADE2gondolain2006flightconfiguration.Adeployablelid
protectsthecoldopticsduringlaunchandascent.Thereflectiveshieldscreening
theflighttrainandparachutefromviewoftheantennas,andthebaronwhich
itsits,arethelargestsinglesourcesofsystematicuncertaintyandismeasured
inflightbyheatingtheshield.
Figure3.Layoutofhornaperturesontheapertureplateandthepositionof
theskyportandexternalcalibratorinthethreeskyviewingpositionsofthe
carousel.Thehornsarearrayedinthreegroups,withthe3GHzoccupyingone,
the5and8GHzoccupyinganother,andthehighbandsoccupyingthethird.
Thehornsareslicedattheaperturessothatthebeamspoint30◦ fromzenith
inthedirectionsindicated,withallhornspointinginthesamedirectionwith
theexceptionofthe3GHz,whichpointsopposite.Foreachofthecarousel
positions, the open ellipse shows the position of the sky port, and the filled
ellipseshowsthepositionoftheexternalcalibrator.Thearrowsindicatethe
directioninwhichthehornbeamspoint30◦fromvertical.
flarestoextendovertheantennaaperture.Theactualalignment
precisionof1mmpreventssuchmisalignment.
2.2.GondolaConfiguration
Figure5.PhotographoftheARCADE2instrumentjustpriorto2006flight.
Figure4showsaschematicoftheentirepayload.Thedewaris Theinstrumentcoreiscontainedwithinthelarge(1.5mdiameter,2.4mtall)
bucketdewar.Thelidisshownclosedforlaunch.Theelectronicsboxcontaining
mountedinanexternalframesupported64mbelowtheballoon,
thewarmstagesoftheradiometersandthepayloadelectronicsismountedto
and boxes containing the readout and control electronics and theleftofthedewar,andthethree-axismagnetometerstodetermineorientation
batteries are mounted on the frame. The external frame is relativetotheEarth’smagneticfieldareseentotheleftoftheelectronicsboxat
suspended by two vertical cables from a horizontal spreader theedgeoftheframe.Thereflectorplateisvisibleatthetopofthephotoabove
thelid.Cardboardpadsatthefourbottomcornersabsorbsomeoftheimpact
bar1.14mabovethetopofthedewar,whichitselfissuspended
whenthepayloadhitsthegrounduponterminationoftheflight.
by two cables from a rotator assembly. The rotator maintains
(Acolorversionofthisfigureisavailableintheonlinejournal.)
therotationofthepayloadbelowtheballoonatapproximately
0.6rpm.Therotatorassemblyissuspendedfromatruckplate,
abovewhichistheflighttrain.Reflectorplatesofmetalizedfoam the dewar and the exterior electronics box via cabling and a
aremountedonthespreaderbartoshieldtheedgeoftheantenna collarofinsulatedconnectorsatthetopofthedewar.
beamfromtheflighttrain.Figure5showsaphotographofthe Three-axis magnetometers and clinometers mounted on the
payload prior to a launch. The total mass at launch, including frame, along with GPS latitude, longitude, and altitude data
liquidhelium,is2400kg. recordedbyCSBFinstruments,allowthereconstructionofthe
Thedewarcanbetippedtoanglesof2◦fromvertical,chang- pointing of the antenna beams during flight. During the 2006
ingtheangleoftheantennaswithrespecttothereflectorplates flight, the magnetometers failed, and the pointing was recon-
and flight train, by moving the battery box outward from the structedwithacombinationoftheclinometersandradiometric
frame.Afiberglasslidmountedontheframeisclosedtocover observationsoftheGalacticplanecrossings.Theuncertaintyin
the dewar on ascent and descent and opened for observations. thereconstructedpointing(asmallfractionofadegree)issmall
Thermometry,heater,andothersignalsareinterfacedbetween comparedwiththebeamsize.
3
TheAstrophysicalJournal,730:138(12pp),2011April1 Singaletal.
0 2.4.Radiometers
20 Measured horn Figure 7 shows a block diagram of the radiometers. Radi-
with 30? slice
CCHORN predicted -10 ation incident from the horn goes through a compact circular
10 unsliced horn to rectangular waveguide transition (Wollack 1996). Typical
-20 R reflections from the horn and transition system are less than
Bi) 0 elativ −753H0zdbBetawceroesnsrathdeiaetinotnirferormaditohmeheoterrnbananddth.aAtfsrwomitcthhecihnoteprsnaalt
d -30 e
Gain ( -10 Gain rseyfsetreemnc(eMloEaMd.SA)tsw3iatnchd5isGuHsezd,,awmhiicleroa-teltehcetroicthael–rmchecahnanneilcsaal
-40 (d latchingferritewaveguideswitchisused.At3and5GHz,the
-20 B) internal reference load is a simple coax termination stood off
-50 with a stainless steel coax section. At 8 and 10 GHz, the ref-
-30 erence load is a wedge termination in waveguide with a layer
-60 ofSteelcast,amicrowaveabsorberconsistingofstainlesssteel
powder mixed into a commercially available epoxy (Wollack
-90 -75 -60 -45 -30 -15 0 15 30 45 60 75 90 etal.2008)castontoanaluminumsubstrate.At30and90GHz,
Angle (deg)
thereferenceloadisasplit-blockwedgeconfigurationofSteel-
Figure6.Measuredresponseforthe10GHzhorn,atthecenterbandfrequency cast in waveguide. These wedge termination and split-block
of10.11GHz(dots)asafunctionofangle,scanningintheplanewheretheeffect
reference loads have reflected power attenuation of more than
ofthe30◦aperturesliceismaximal.Thepredictedsimulatedbeampatternfor
30 dB across the entire frequency band and are described by
anunslicedhornisalsoshown(solidline).Thesideloberesponseislow,andthe
effectofthesliceisseenprimarilyinthefirstsidelobe.Farsidelobesarefurther Wollack et al. (2007). We measure the temperature of the ab-
suppressed. The beam pattern does not vary appreciably over the frequency sorber in each reference load with ruthenium-oxide resistance
band. thermometers(seeSection3.1).
(Acolorversionofthisfigureisavailableintheonlinejournal.) Each radiometer uses a cryogenic high electron mobility
transistor (HEMT) based front-end amplifier which sets the
system noise temperature. The radiation exiting the switch is
amplifiedbythecoldHEMTandpropagates,viacoaxialcable
2.3.HornAntennas at the low frequencies and waveguide at 30 and 90 GHz, out
of the dewar to the warm stage contained in the electronics
Thecorrugatedantennasforsixofthechannelshave11◦.6full box. The warm stage features a warm HEMT amplifier, an
widthathalf-powerGaussianbeams,whiletheantennaforthe attenuator to eliminate reflections and tune the output power
30GHznarrowbeamchannelhasa4◦ fullwidthathalf-power to match downstream components, a bandpass filter to select
Gaussianbeam.Thehornsareslicedattheaperturetopoint30◦ the desired frequency band, a second warm HEMT amplifier,
fromzenithwhenhungfromtheflataluminumapertureplate. and a power divider to split the signal into a high- and a low-
Thisissotheantennabeamboresightsaredirectedawayfrom frequency channel. Each of the high and low channels has a
theflighttrainandsothattheytraceoutacircle60◦ onthesky bandpassfilter,adetectordiode,andanaudiofrequencypreamp,
asthedewarrotatesbelowtheballoon.Inordertoachievethe outputting a voltage level corresponding to the power of the
narrowestpossiblebeamat3GHzgiventhespatialconstraints, radiationincidentonthediode.Thevoltagesignalisthencarried
acurvedprofilingofthehornswasemployed.Afulldiscussion to an electronics board where a lock-in amplifier demodulates
ofthehornsisgivenbySingaletal.(2005). the signal in phase with the switch, integrates it for 0.533 s,
The horns were designed using mode matching simulation and digitizes it. In this way, the final output is proportional
software. The beam pattern for the fabricated 10 GHz horn to the difference in temperature between what the radiometer
was mapped over greater than 2π steradian at the Goddard is viewing and the internal reference load, averaged over the
Electromagnetic Anechoic Chamber test range at the NASA integrationperiod.Thereisalsoatotalpoweroutputsignalinthe
GoddardSpaceFlightCenter.Themeasuredbeampatterninthe datastreamwhichisnotdemodulatedbythelock-inamplifier.
planecontainingthemaximaleffectofthe30◦ aperturesliceis Tables1and2showselectedradiometerpropertiesandfigures
shownalongwiththepredictedonefromdesigninFigure6.The ofmerit.Figure8showsaphotographofaradiometercoldstage.
apertureslicehasaminimaleffectonthesymmetryofthebeam, Thecustom-designedradiometercomponentsarethehorns,the
andisseenprimarilyinthefirstsideloberesponse,depressing circulartorectangularwaveguidetransitions,theferritewaveg-
theresponseonthelongsideandincreasingtheresponseonthe uideswitches,andthecoldamplifiersat3,5,8,and10GHz.
short side. This effect is a few dB, overlaid on a first sidelobe Thecoldstagesoftheradiometersarehousedinpansthatcan
response that is 20 dB below the main beam. Response in far hold liquid helium, with a cascading pond system linking the
sidelobesbelowtheapertureplaneissuppressedbymorethan pans.Thecomponentsarestoodofffrombutthermallylinked
50 dB. Because of the large physical sizes involved, it is not totheliquid,withtheliquidprovidingcoolingtomaintainthe
practical to measure the beam pattern for the lower frequency components at temperatures below 4 K. For the horn throats
horns, or any of the horns once installed in the instrument, and internal loads, temperatures within this range are selected
but the correctness of the simulated beam patterns, the low bySPIDcontrol,asdescribedinSection3.1,whiletheswitches
sideloberesponse,andthenegligibleeffectoftheslicehasbeen andHEMTamplifiersrunneartheliquidtemperature.
demonstratedwiththe10GHzhornbeammap.Furthermore,in
theARCADE1instrument,wewereabletomeasurethebeam
2.5.ExternalCalibrator
of its 10 GHz horn both prior to and after installation in the
instrument,andthefarsideloberesponsewasidentical(Kogut ARCADE 2 determines the radiometric temperature of the
etal.2004a),furtherdemonstratingconsistency. sky by using a full-aperture blackbody external calibrator as
4
TheAstrophysicalJournal,730:138(12pp),2011April1 Singaletal.
Figure7.BlockdiagramofARCADE2radiometers.Thedemodulatedoutputisproportionaltothedifferenceintemperaturebetweenradiationfromtheantennaand
theinternalreferenceload.Thewarmstagesoftheradiometersoperateat∼280K.
Table1
TableofSelectedRadiometerHardwareSpecifications
Specification 3GHz 5GHz 8GHz 10GHz 30GHz 30#aGHz 90GHz
Lowband(GHz) 3.09–3.30 5.16–5.50 7.80–8.15 9.2–10.15 28.5–30.5 28.5–30.5 87.5–89.0
Highband(GHz) 3.30–3.52 5.50–5.83 8.15–8.50 10.15–10.83 30.5–31.5 30.5–31.5 89.0–90.5
Switchtypeb MEMS MEMS Cir Cir Cir Cir Cir
Coldamplifiermfg. Berkshire Berkshire Berkshire Berkshire Spacek Spacek JPL
Coldampmodel S-3.5-30H C-5.0-25H X8.0-30H X10.0-30H 26-3WCc 26-3WCd 90GHzAmps
Coldampserial 105 106 108 101 5D12 6C21 W82&W105d
Coldampgain(dB) 40 29 40 40 26 20 52
Warmamp1gain(dB) 35 33 33 30 46 22 −7e
Attenuator(dB) −26 −13 −26 −20 −3 −10 −6
Warmamp2gain(dB) 35 33 33 31 ··· 23 40
Detectorpower(dBm) −21.3 −20.6 −28 −24.6 −25.9 −42.5 −24.6
X-Calreflectionf(dB) 42.4 55.5 58.6 62.7 55.6 xxg 56.6
Notes.
a30#designatesthe30GHzchannelwiththenarrowerantennabeam.
bSwitchesareeitherMEMS(micro-electro-mechanicalsystem)orferritelatchingwaveguidecirculatorswitches(Cir).
cThesemodelnumbershavetheprefix“SL315-.”
dAt90GHztherearetwocoldamplifiersinseries.
eThe90GHzchannelwarmstageusesamixer,witha79.5GHzlocaloscillator,totranslateitto8–11GHz,andallfollowingcomponentsoperatein
thatfrequencyrange.
f Thisisthemeasuredattenuationofreflectionsfromtheexternalcalibrator,whenitisviewedwiththehornantennaforthechannel.Thesevalues
presentedhereweremeasuredingroundtesting,andthemeasurementisdescribedbyFixsenetal.(2006).
gThisexternalcalibratorreflectionswerenotmeasureddirectlywiththe30#channelantenna,buttheyareassumedtobeevenlowerthanfortheother
30GHzantenna.
anabsolutetemperaturereference.Tofunctionasablackbody radiometricperformanceisdescribedindetailbyFixsenetal.
emitter, it should have very low power reflection across the (2006).
entire ARCADE 2 frequency range. The emitting surfaces of The calibrator consists of 298 cones, each 88 mm long and
theexternalcalibratorneedtobeclosetoisothermal,and,while 35 mm in diameter at the base, of Steelcast absorber (see
located on the carousel at the top of an open bucket dewar, Section 2.4) cast onto an aluminum core. The aluminum pro-
havetobepreciselytemperaturecontrolledwithintherangeof videsforenhancedthermalconductivityandthereforereduced
2.5–3.1K. thermalgradients,whichisaugmentedbyacopperwireepoxied
The external calibrator is based on the successful model, onto the end of the aluminum core and running almost to the
from COBE/FIRAS and the first generation ARCADE in- tip.Theradiometricsideoftheexternalcalibratorisvisiblein
strument, of a full aperture calibrator that is absorptive and Figure2.Figure9showsacutawayviewofacone.
isothermal, achieved through a combination of material and The cones are mounted pointing downward on a horizontal
geometry (Mather et al. 1999; Kogut et al. 2004b). The AR- aluminumplate,behindwhichare50alternatinglayersofalloy
CADE 2 calibrator is as good as the COBE/FIRAS one ra- 1100 aluminum and fiberglass sheets, to give a low vertical
diometrically, having a similar level of reflections, although but high horizontal thermal conductivity, to the end of getting
the ARCADE 2 frequency coverage extends to much longer the different cones as isothermal as possible. This aluminum
wavelengths. The calibrator has a reflected power attenuation plateisweldedtoanaluminumshieldingwhichsurroundsthe
of less than −55 dB in the range from 5 to 90 GHz and elliptical circumference of the area of cones so that the cones
of −42 dB at 3 GHz. Measured reflected power attenuation are surrounded on the top and sides by an isothermal metal
values are presented in Table 1 and the external calibrator’s surfacemaintainedattemperaturesnear2.7K.Behindthestack
5
TheAstrophysicalJournal,730:138(12pp),2011April1 Singaletal.
Table2
TableofSelectedMeasuredRadiometerPerformanceSpecificationsfromthe2006Flight
CntrFreq Bandwidth Trcvra Offsetb Noise(p√re-flt)c Noise(r√sduls)d Noise(√map)e
Channel (GHz) (MHz) (K) (mK) (mK s) (mK s) (mK s)
3GHzLo 3.15 210 5.5 180 7.1 9.3 11.8
3GHzHi 3.41 220 6.5 35 7.1 7.8 10.1
5GHzLo 5.33 340 6 −210 3.2 ... ...
5GHzHi 5.67 330 6 −200 3.5 ... ...
8GHzLo 7.97 350 10 6 1.4 5.5 5.5
8GHzHi 8.33 350 8 11 1.4 6.1 5.2
10GHzLo 9.72 860 13 180 6.8 3.7 3.0
10GHzHi 10.49 680 11 35 5.3 3.7 3.0
30GHzLo 29.5 2000 75 −30 21.5 208.2 206.9
30GHzHi 31.0 1000 72 −15 14.9 103.3 103.4
30#GHzLo 29.5 2000 270 32 18 885.0 880.4
30#GHzHi 31.0 1000 340 38 14 418.5 406.1
90GHzLo 88.2 1500 44 −75 5.2 50.5 42.0
90GHzHi 89.8 1500 38 −95 5.7 25.3 27.1
aTrcvristhereceivernoisetemperatureoftheamplifier,afigureofmeritthatisequaltothetemperaturethatwouldbeobservedbya
totalpowerradiometercontainingtheamplifierandviewingasourceatatemperatureofabsolutezero.Thesevaluespresentedhere
weremeasuredingroundtestingpriortothe2006flight.
bTheconstantoffsetistheradiometeroutputwhentheinternalreferenceloadandtheobjectbeingviewedareatthesametemperature,
multipliedherebythegaintobeexpressedasatemperature.
cThisisthewhitenoiseasmeasuredingroundtestingpriortothe2006flight.
dThisisthewhitenoisedeterminedfromtheresidualsofthedatafromthe2006flight.
eThisisthewhitenoiseasdeterminedfromtheskymapvarianceinthe2006flight.
Figure9.Cutawayviewofexternalcalibratorconeshowinginternalstructure
and predicted linear temperature profile from the 2006 flight, with the base
thermallyfixedandtheconeimmersedinacoldatmosphere.Theconeconsists
ofSteelcastabsorber(black)castontoanaluminumcore(lightgray).Thereis
alsoacopperwirerunningfromthetipofthealuminumcoretonearlythetip
Figure 8. Photograph of cold stage of ARCADE 2 8 GHz radiometer. The ofthecone(darkergray).Theholeinthebaseisthreadedforamountingbolt
throatofthehornantennaboltstotheopencircularwaveguideend.Theinternal toaffixtheconetothealuminumbackplate.Themeasuredthermalgradientis
referenceloadofthisradiometerisawedgeterminationinwaveguide,whichis 600mKfromthebasetothetipofthecones.Ninety-eightpercentofthetotal
surroundedbytheinsulativecylindervisibleinthephotograph. gradientisinaregionnearthetipcontaining3%oftheabsorber,andthemean
(Acolorversionofthisfigureisavailableintheonlinejournal.) depthforabsorptionvarieswithfrequency.
(Acolorversionofthisfigureisavailableintheonlinejournal.)
ofalternatinglayersisanotherhorizontalaluminumbackplate, heating elements, and wiring. In the 2006 configuration the
and behind that a half inch gap to allow room for bolt heads, entirecalibratorissurrounded,exceptontheradiometricside,
6
TheAstrophysicalJournal,730:138(12pp),2011April1 Singaletal.
by a tank of liquid helium, which is thermally coupled with we position an outflow hose leading from a vent hole in the
stainless steel standoffs to the rear aluminum plate. This layer top of the lid to a position close to level with the dewar rim.
of liquid helium intercepts any external heat load incident on Thismaintainssomepositivepressureandheliumgasoutflow
theexternalcalibratorfromthebackandsides. atthelid–dewarinterface,whichdiscouragesthecondensation
Temperatures are monitored in 23 of the cones with ther- ofambientwatervapor.
mometersembeddedintheSteelcastabsorberatvaryingdepths We launch the instrument with the carousel in the “ascent”
and radii. Three of the cones contain two thermometers each, positionthatalignstheventholesintheapertureandcarousel,
for a total of 26 thermometers within the absorber. Resistance which directs boil-off gas across the back of the external
heatersaremountedontherearaluminumplate,andsevenaddi- calibrator, providing a powerful cooling source for its large
tionalthermometersareaffixedatvariouspoints.Thermometry thermalmass.Onascent,aboutone-thirdoftheheliumisboiled
andheatersignalsfromthecalibrator,aswellasfromelsewhere off, with the remainder cooled continuously to 1.5 K when
onthecarousel,arecarriedthroughahollowtubeintheaxisof floataltitudeisreached.Weturnonthesuperfluidpumpsonce
the carousel to the rest of the core inside the dewar. The true the helium bath is below the superfluid transition temperature
emissiontemperatureoftheexternalcalibratorisavolumeinte- of 2.177 K. After 3 hr of ascent, float altitude is reached and
graloverthephysicaltemperaturedistributionweightedbythe we open the instrument lid for observing. During observing,
absorber emissivity and the electric field distribution from the we move the carousel to a new position about once every
antenna aperture. We approximate the true emission tempera- 5minutes.Weobserveforaround4hr,limitedbythewestward
tureasalinearcombinationofthetemperaturesasmeasuredby driftoftheballoonoutofrangeoftelemetry.Thelidisclosed
the thermometers, with weights derived from the actual flight for descent, and the payload is severed from the balloon and
data. returns to the ground on a parachute, to be recovered by
CSBF staff. At the end of the 2006 flight, nearly 800 liters
2.6.PayloadElectronics ofliquidheliumremainedinthedewar,orenoughfor6hrmore
ofobservation.
The main payload electronics consists of the cryogenic Inthe2005flight,thecarouselbecamestuckinoneposition
thermometer resistance readout and control for heaters, as soonafterobservingcommenced. Thecausewastracedtothe
described in Section 3.1, as well as radiometer lock-in and output torque of the carousel motor exceeding the maximum
integration, voltage readout for the various payload analog torque of the attached gearbox, stripping the gears, and was
devices,includingthemagnetometers,clinometers,andambient remedied for the 2006 flight. In the 2006 flight, the most
temperaturetransducers,digitallogiclevelreadout,generation significant instrumental failure was with the 5 GHz MEMS
oftheswitchdrivingcurrent,andgenerationandamplification switch,renderingdatafromthatchannelnotusefulforscience
for commandable signals to the lid, rotator, tilt, and carousel analysis.
movement motors. These functions are performed by custom
electronicsboards.Typicalpowerrequiredfortheelectronicsis
3.CRYOGENICPERFORMANCE
220Wwithpeakcapacity1800W.
ThedigitaldatastreamisrelayedviatheRS-232serialdata The ability to measure and control the temperature of
standardtotheConsolidatedInstrumentPackage(CIP)provided cryogeniccomponentswhileinthepresenceofavariablehelium
by CSBF, which transmits it, along with data from the CSBF gasflowandpotentiallyexposedtoambientsourcesofwarming
instruments and the video signal, to the ground. Each 1.067 s isthelimitingfactorintheprecisionofradiometrictemperature
record of data consists of digitized counts corresponding to a measurements made with ARCADE 2. Thermal measurement
voltagereadingacrosseverycryogenicthermometer,thevoltage and control is the most important and challenging facet of the
reading across the reference resistors on each thermometer experiment.
readout board (see Section 3.1), the voltage output of the ARCADE 2 requires component temperatures to be main-
various analog payload devices, the SPID control parameters tained near 2.7 K. Some components are controlled passively
foreverycontrolledheater,digitallogiclevels,andtworeadings by being thermally sunk to liquid helium and exposed to cold
ofthedemodulatedandtotalvoltageoutputofeveryradiometer helium gas. Other components are actively temperature con-
channel. Each 1.067 s also allows four 2 byte commands to trolled, with both a coupling to liquid helium and controllable
be transmitted via the CIP to the instrument and executed. resistance heaters.The gas cooling isan undesirable perturba-
ThecommandsincludethesettingofSPIDparameters,lock-in tiveeffectonactivelycontrolledcomponents.Liquidheliumis
amplifiergains,andmotormovement. movedtoneeded areasoutsideoftheliquidheliumbath,such
Avideocameramountedonthespreaderbarabovethedewar as the aperture plate and carousel at the top of the dewar and
allowsdirectimagingofthecoldopticsinflight.Twobanksof thecoldstagesoftheradiometers.Atballooningaltitudesliquid
light-emitting diodes provide the necessary illumination. The heliumiswellbelowthesuperfluidtransitiontemperatureand
cameraandlightscanbecommandedonandoff,andwedonot doesnotrespondtomechanicalpumping,sosuperfluidpumps
usedataforscienceanalysisfromtimeswhentheyareon. with no moving parts are used. In such pumps, with heating
power of less than 1 W, liquid can be pumped to a height of
2.7.FlightOperations severalmeters.ARCADE2features13superfluidpumpswhich
canmove55litersminute−1 toaheightof2.5m.
Whentheinstrumentisreadyforflight,thelidisclosedand
thedewariscooledwithnitrogentoaround100K.Wethenfill
3.1.ThermometryandTemperatureControl
thedewarwitharound1900litersofliquidhelium,whichtakes
several hours. We await a launch opportunity, with the helium Wemeasure104cryogenictemperaturesintheARCADE2
level topped off each day. Approximately 200 liters of liquid payload. Twenty-six are within the external calibrator cones,
helium are boiled off per day. To avoid freezing the lid to the 7 are elsewhere on the calibrator, another 19 are at various
dewar with ice accumulation while the payload awaits flight, pointsonthecarousel,26areoncomponentsoftheradiometers
7
TheAstrophysicalJournal,730:138(12pp),2011April1 Singaletal.
Figure 10. Temperature of the 5 GHz internal reference load vs. time for a
sectionofdatafromthe2006flightshowingSPIDtemperaturecontrolofthe
load.Theloadhasarelativelysmallmassandislocatedinasteadythermal Figure 11. Temperature vs. time of a cone in the external calibrator for a
environment.Thetemperaturewascommandedtobesetto2.730Kat51s sectionofdatafromthe2006flightshowingSPIDtemperaturecontrolofthe
andbackto2.724Kat138s.Ingeneral,itisimportantthatthetemperature externalcalibrator.Theexternalcalibratorisarelativelylargesystemexposedto
ofcomponentsunderSPIDcontrolbesteadyandknown,butnotthatspecific changeablethermalconditions.Theringingisbeneficialasitisslowcompared
commandedtemperaturesbeobtainedexactly. totheradiometerresponsetimeandallowsthesamplingofawiderrangeof
calibratortemperatures.
includingtheinternalreferenceloads,switches,horns,andthe in firmware. Figure 10 shows the accuracy and precision of
pans in which the radiometers sit, and 26 are at various other the SPID temperature control of the 5 GHz internal reference
points within the dewar. We measure these temperatures us- load. This internal reference load has a relatively small mass
ingfour-wireACresistancemeasurementsofruthenium-oxide (∼10g)andislocatedinarelativelystablethermalenvironment.
resistors, whose resistance is a strong function of temperature Figure 11 shows the SPID temperature control of the external
below 4 K. The thermometers are excited by a 1 μA 37.5 Hz calibrator, a much larger system that is subject to changeable
squarewavecurrent,andsignalsarecarriedthroughcoldstages thermaldynamics.
via a sequence of brass, copper, and manganin cryo-wire. The
resistances of the thermometers are read by custom electron-
3.2.ExternalCalibratorThermometerCalibration
icsboards(Fixsenetal.2002),whichoutputadigitizedvoltage
levelforeverythermometerineverydatarecord.Theboardsalso Special care is taken in determining the resistance versus
containfiveon-boardresistorsofknownresistancespanningthe temperaturecurvesfortheruthenium-oxidethermometersem-
dynamicrangeofthecryogenicthermometerresistancesthatare bedded in the cones of the external calibrator, as errors and
excited with the same current as the cryogenic thermometers. uncertainties in these curves lead directly to errors and uncer-
In this way, a reliable digital voltage counts to resistance con- tainties in the measured radiometric temperature of the sky. A
versionisobtainedineverydatarecordtoexpressthemeasured specific calibration setup is employed for these thermometers
resistanceofthecryogenicthermometers.Themeasuredsignals to minimize any possible thermal gradients and wiring differ-
canthenbeconvertedinsoftwaretotemperaturesusinglook-up ences between calibration conditions and flight conditions. In
tables containing the resistance versus temperature curves for this configuration, the cones and an NIST calibrated standard
eachthermometer. thermometer are mounted on an 1100 series aluminum plate
Theresistanceversustemperaturecurvesfortheruthenium- inside of an evacuated stainless steel pressure vessel, which
oxide thermometers are determined in ground testing against is itself submerged within the liquid helium bath of a large
a thermometer of identical design previously calibrated by (0.3mdiameter,1.5mtall)testdewar.Thealuminumplateis
NIST.Thisstyleofthermometershasdemonstratedcalibration thermallylinkedtothebaththroughacopperrodwhichpasses
stability to within 1 mK over four years (Kogut et al. 2004a), through a superfluid-tight hole in the pressure vessel. The liq-
verified with the observed lambda superfluid transition as an uid helium bath of the test dewar is pumped in stages, with
absolutereference.Thermometerself-heatingisnegligible(less thepumpingvalveopenedsomeamountandthenthebathleft
than 1 nW) and is included in the resistance-to-temperature alone to eventually equilibrate to some steady temperature. In
calibrationprocess. this way, once a steady bath temperature is reached, with the
Desiredtemperaturesaremaintainedinkeyplaces,suchasthe evacuatedvesselsubmergedinanisothermalliquidheliumbath,
radiometerinternalreferenceloadsandswitches,hornthroats, therearenoobvioussourcesofthermalgradientsinthesystem.
and external calibrator, through resistance heaters under SPID Theprocedureleadstoasteadythermalsituationsuchthatther-
(set point, proportional, integral, and differential) control. The mometeroutputsshownocoherentmovementovertimescales
values S, P, I, and D are set in real time by the user for each of 10 minutes. The aluminum plate, and therefore the cones
of 32 SPID channels, and the desired set point temperature and NIST standard thermometer, should be quite isothermal
as well as the current actual temperature are expressed in when the bath has equilibrated at a given temperature, as the
counts,withtheuserperformingtheconversionbetweencounts aluminum plate is thermally lined to the bath and within an
and temperature as necessary. The output voltage level for evacuatedvesselwiththewallsatnearlythesametemperature
each SPID channel is recalculated once per record (1.067 s) astheplate.
8
TheAstrophysicalJournal,730:138(12pp),2011April1 Singaletal.
Figure12.Theupperplotshowsaresistancevs.temperaturecurveforatypical
externalcalibratorconethermometer.Thelowerplotshowsthefractionalerror
intemperaturefromremovingeachtemperaturepointinsequencetoforma
reducedinformationcurveandfeedingthatreducedcurvetheresistanceofthe Figure13.Uncertaintyintheresistancevs.temperaturecurvedetermination
removedtemperaturepoint.Wetakethemagnitudeofthiserrorasanestimate for the cones of the external calibrator, as a function of temperature. The
oftheuncertaintyintheresistancevs.temperaturecurvearisingfromsources resultpresentedisaveragedoverallofthecones.Thedashedlineisstatistical
otherthanstatisticaluncertaintyandthecapacitancecorrection.Thedifferent uncertainty,thedottedlineistheuncertaintyduetotheappliedcorrectiontothe
shapedpointsrepresentdatafromthethreedifferentcalibrationruns. NISTcalibratedthermometerforshuntcapacitanceinthereadoutwiring,and
thedash-dottedlineisthemagnitudeoftheerrorinreproducingthetemperature
ofacalibrationpointgiventheresistanceofthatpointandareducedinformation
curvenotcontainingthepoint.Thesolidlineisthequadraturesumofthethree
This process leads to discrete temperature points where a
sources,whichistakentobethetotaluncertainty.
definite temperature, read by the NIST standard thermometer,
can be associated with a measured resistance for each of the
capacitancecorrectionuncertaintywellbelow1mKattemper-
cone thermometers. The 14 total calibration points from three
aturesbelow3Kandrisingto3mKat3.4K.Weestimatethe
separate runs can be combined to form a reasonably dense
remaining uncertainty from other sources using the figure of
calibrationcurvewithvaluesbetween2.5and4.2K.
merit of how well data from the three separate test runs agree
As we measure the resistance of the thermometers using
with each other. Each run contains only four or five discrete
a 37.5 Hz square wave excitation, this measured resistance
temperaturepoints.Wecombineall13pointstoformasingle
includes the effect of shunt capacitance in the harnessing
resistance versus temperature curve for each thermometer. To
connectingthewirestotheboard.Ourthermometercalibration
estimatetheremaininguncertainty,weremoveeachtemperature
procedure assigns a temperature to each measured resistance
pointinsequenceandcomparethepointtothecurvedetermined
usingidenticalharnessingandelectronicsboardsasinflight,so
fromthe12remaining points.Thisisatestoftheconsistency
the effect of the shunt capacitance is automatically accounted
ofthecombinedcurveandofthethreeruns.Figure12showsa
forintheresistanceversustemperaturecurveobtainedforeach
resistance versus temperature curve for a typical external cali-
thermometer.However,theNISTcalibratedthermometer,which
bratorconethermometerandthefractionalerrorintemperature
is read by the flight thermometer boards and flight harnessing
ateachtemperaturepointdeterminedbyremovingthepoint.
in our calibration setup, was itself calibrated with DC readout
We take the magnitude of the error in reproducing the
techniques featuring no shunt capacitance. To correct for the
temperatureoftheremovedpointgiventhereducedinformation
smalleffectonthemeasuredresistanceoftheNISTcalibrated
curveasanestimateoftheuncertaintyofthatpointinthecurve,
thermometer,anestimateoftheeffectoftheshuntcapacitanceof
neglecting statistical and capacitance correction uncertainty.
thecalibrationsetupisformedusingdatatakenwith20resistors
Figure 13 shows an average over all cones of the uncertainty
ofknownvalueplacedattheendoftheharnessing.Thisyields
due to all three sources and the total uncertainty, taken to be
alinearrelationshipbetweentheactualresistanceandtheerror
thequadraturesum.Theuncertaintiesintemperaturegenerally
in the measured resistance. The shunt capacitance is around
increase steeply above 3 K because at higher temperatures
300pFwhichresultsinanoffsetinthemeasuredresistanceof
8Ωwhenmeasuringa20,000Ωresistor,whichistheresistance the resistance versus temperature curves flatten. At 2.7 K, the
ofathermometerataround2.7K.An8Ωoffsetresultsina2mK average total uncertainty in the resistance versus temperature
curvesis1.3mK.
offset in the thermometer reading. This small estimated effect
is then subtracted from the raw NIST calibrated thermometer
3.3.HeatFlow
data,withthestandarddeviationofthelinearfitprovidingthe
uncertaintyinthisappliedcorrection. ARCADE2successfullymaintainscriticalcomponentsnear
Thetotaluncertaintyinthedeterminedresistanceversustem- 2.7 K at and near the top of an open bucket dewar at 37 km
peraturecurvesoftheconethermometersconsistsofcontribu- altitude. This is achieved by moving large quantities of super-
tions from raw statistical uncertainty, from the uncertainty in fluid liquid helium so that each component is either passively
the shunt capacitance correction for the NIST calibrated ther- controlled at the bath temperature or has a thermal path to the
mometer, and that from any other sources. The first two are liquidtoallowactivetemperaturecontrol.Othercomponentsof
straightforward to determine, with statistical uncertainty well theinstrumentarenotthermallycontrolled,butgenerallycome
below 1 mK except at above 3.4 K where it is 1 mK, and the to temperatures between 1.5 K and 20 K depending on their
9
TheAstrophysicalJournal,730:138(12pp),2011April1 Singaletal.
locationrelativetoexternalheatsources.Wemonitortempera- bath temperature by direct contact with the superfluid liquid
turesonthesenon-criticalcomponentstodemonstratethattheir heliumcollarontheexteriorofthehorn.
fluctuations do not affect the radiometer output. The overall There are also transient effects when the carousel moves
thermalperformanceoftheentireinstrumentiswellmeasured fromonepositiontoanother.Portionsofthecarouselperimeter
andunderstood. that pass near where the carousel drive chain enters through
Thethermalbehavioroftheinstrumentcoreinflightresults the connector collar are momentarily warmed, and the back
from the balance of warming and cooling power. Cooling is plate of the external calibrator is momentarily cooled as the
providedbybothpumpedliquidheliumandboil-offheliumgas. calibrator passes over the aperture plate vent hole, briefly
Theboil-offgasischanneledoutofthedewarthroughthe38mm channeling gas between the calibrator’s aluminum shielding
perimeter gap between the aperture plate and the inner dewar and the surrounding liquid helium tank and then across the
wall.Thereisalsoone100mmdiametergasventporteachon backplate.Theseeffectsonradiometricallyactivepartsofthe
the aperture plate and carousel, which, when aligned, channel instrument used for science analysis are small, and data from
boil-offgasupacrossthebackplateoftheexternalcalibratorand timesinwhichthecarouselismovingarenotusedforscience
outthroughapathinthefoaminsulationcoveringthecarousel. analysis.
Thisconfigurationisusedtodirectmaximumheliumgasflow The observed temperature gradient among the concentric
tothecalibratorbackuponascent. baffles inside the dewar wall, from 9 K at the inner baffle to
Thesourcesofwarminginflightare(A)infraredradiationand around30Kattheouterone,indicatesthattheheatleaktothe
conductionfromthewarmerambientatmosphere,(B)radiation aperture plate from the warm dewar walls is reduced, via the
andconductionfromwarmcomponentsoftheinstrumentsuch bafflingandblow-byofboil-offgas,tothelevelofafewmW.
asthedewarwallsorthecarouselrotationdrivechain,and(C) The observed gradient through the foam topping the carousel
SPIDheatersasdescribedabove.Therearethreedistinctpaths indicatesthattheheatloadfromabovetothecarouselisonthe
by which ambient warmth can heat the core: (1) warming of order of 30 W, and this is roughly consistent with the source
the carousel from above, (2) warming of the aperture plate being infrared radiation from a 200 K body incident on the
from above through the sky port of the carousel, and (3) aluminumsheettoppingthefoam.
warming via radiation from the warm wall of the dewar near Inthe2006flight,weexperiencedproblemswithtemperature
thetop.Toreducetheheatloadfromthedewarwall,thereare controlling the 3 and 8 GHz internal reference loads, both
three concentric sheet aluminum baffles inside the dewar wall becoming fixed at near the liquid helium bath temperature for
extending 0.6 m down from the top of the dewar. In addition, significantperiodsoftheflight.Inthecaseofthe3GHzload,this
attheverytopthereisafivelayervacuumdepositedaluminum was likely caused by the insulative housing falling off, which
withDacronconcentricveilextending200mmdownfromthe was observed upon completion of the flight. The 5 GHz load,
top outside of the outermost baffle. To reduce the heat load whichwasofthesameconstructionasthe3GHzload,didnot
to the carousel from above, it is covered in 1 m3 of foam experience this problem, most likely because the 5 GHz load
insulation(FomoHandi-FoamSR,atwo-componentslow-rise waslocatedabovetheliquidheliumlevelduringascent.
polyurethanefoam),whichistoppedwitha0.25mmreflective
aluminumsheet. 3.4.ExternalCalibratorThermalPerformance
In the 2006 configuration, the aperture plate was cooled
directly by six pump-fed liquid helium tanks mounted on the In both the 2005 and 2006 flights, all radiometrically active
undersideinvariousplaces.Beingrelativelythinandcovering partsoftheexternalcalibratorweremaintainedwithin300mK
a large area, the aperture plate is subject to thermal gradients of 2.7 K. However, the external calibrator cones displayed
evenwithliquidheliumtanksincontactatdiscreteplaces.The significant gradients from base to tip. In the 2005 flight, the
large horns at 3, 5, and 8 GHz have sheet metal “collars” on apertureplate,whichwasattemperaturessignificantlywarmer
the exteriors, providing a sheath for liquid helium to maintain than the calibrator, warmed the tips relative to the base. The
each horn aperture at the bath temperature. Aperture plate situation was reversed in the 2006 flight, where the presence
temperaturesinflightvarybetween1.5Kand10Kdepending of the underside liquid helium tanks caused the aperture plate
on location and time, while temperatures on the carousel vary to run significantly colder and the tips of the cones were cold
between 4Kand20K.Ingeneral, theapertureplateiscolder relative to the bases. The wrap-around tank of liquid helium
than,andcools,thecarousel,astheapertureplateislesswarmed surrounding the external calibrator successfully intercepts all
byheatingfromaboveandiscooledbyheliumtanks. heatloadsfromthetoporsides.Theonlyuncontrolledthermal
The major temporal temperature variations in flight result linkisbetweentheconesontheradiometricsideofthecalibrator
from differing gas flow dynamics in the different positions of andthehornaperturesandapertureplate,whichisresponsible
the carousel. In the position in which the 5 and 8 GHz horns forthethermalgradientsobservedinthecalibrator.Inthe2006
are viewing the sky, the vent holes of the aperture plate and flight,themaximumbasetotipgradientwas600mK.Figure9
carousel are almost aligned, causing increased flow through showsapredictedtemperatureprofilebasedonthisgradient.In
the holes and therefore decreased flow to the perimeter and a lightoftheobservedeffects,amoreuniformtemperaturecould
resulting warming of perimeter areas of the aperture plate and beachievedthroughactivetemperaturecontroloftheaperture
carousel. In the other two carousel positions this effect is not platewithheaterstomaintainitatnear2.7K.
presentandasub-dominanteffectisobserved.The“highsky”
positionexposesthe10–90GHzhornstothesky.Becausethese 3.5.AtmosphericCondensation
horns have relatively small apertures there is more metal of
The potential problem with a cold open aperture is conden-
the aperture plate exposed to the sky port and therefore more
sation from the atmosphere. Condensation on the optics will
infrared radiation incident on it from above. This effect is not
reflect microwave radiation adding to the radiometric temper-
presentwhenthe3GHzhornviewsthesky,asthe3GHzhorn
ature observed by the instrument in an unknown way. In the
aperture completely fills the sky port and is maintained at the
courseofanARCADE2observingflight,theapertureplateand
10