Accepted in ApJ
Preprint typeset using LATEX style emulateapj v. 11/10/09
THE ARCADE 2 INSTRUMENT
1
2,3
J. Singal , D.J. Fixsen
3
, A. Kogut , S. Levin4 , M. Limon5 , P. Lubin6 , P. Mirel7,3 , M. Seiffert4 , T. Villela8 ,
E. Wollack3 , and C.A. Wuensche8
arXiv:0901.0546v2 [astro-ph.IM] 3 Apr 2010
Accepted in ApJ
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 extra-Galactic emission at six frequencies from 3
to 90 GHz. ARCADE 2 utilizes a double-nulled design where emission from the sky is compared to
that from an external cryogenic full-aperture blackbody calibrator by cryogenic switching radiometers
containing internal blackbody reference loads. In order to further minimize sources of systematic error,
ARCADE 2 features a cold fully open aperture with all radiometrically active components maintained
at near 2.7 K without windows or other warm objects, achieved through a novel thermal design. We
discuss the design and performance of the ARCADE 2 instrument in its 2005 and 2006 flights.
Subject headings: instrumentation: detectors – cosmic microwave background – radio continuum:
galaxies
1. INTRODUCTION
ARCADE 2 is part of a long term effort to characterize the cosmic microwave background (CMB) and Galactic and extra-Galactic microwave emission at cm wavelengths. The background spectrum has been shown to be
a nearly ideal blackbody from ∼60 to ∼600 GHz with a
temperature of 2.725 ±.001 K (Fixsen & Mather 2002).
At lower frequencies, however, where deviations in the
spectrum are expected, existing measurements have uncertainties ranging from 10 mK at 10 GHz to 140 mK
at 2 GHz (see Fixsen et al. (2004) for a recent review).
These uncertainties are primarily the result of instrumentation systematics, with all previous measurement
programs below 60 GHz needing significant corrections
for instrument emission, atmospheric emission, or both.
Muehlner & Weiss (1970), Johnson & Wilkinson (1987),
and Staggs et al. (1996) are examples of notable CMB
absolute temperature measurements from high altitude
balloons.
To improve existing radiometric temperature measurements, an instrument must be fully cryogenic, so that
microwave emission from front end components of the
instrument is negligible. In any absolute radiometric
temperature measurement, radiation from the source being measured is compared by the radiometer to that
from a blackbody emitter of known temperature. Be1
Kavli Institute for Particle Astrophysics and Cosmology
SLAC National Accelerator Laboratory, Stanford University
Menlo Park, CA 94025
email: jsingal@stanford.edu
2 University of Maryland
3 Code 665, NASA Goddard Space Flight Center
Greenbelt, MD 20771
4 Jet Propulsion Laboratory
4800 Oak Drive, Pasadena, CA, 91109
5 Columbia University
1027 Pupin Hall, Box 47, New York , NY 10027
6 University of California, Santa Barbara
Santa Barbara, CA, 93106
7 Wyle Information Systems
8 Instituto Nacional de Pesquisas Espaciais, Divisão de Astrofı́sica, Caixa Postal 515, 12245-970 - São José dos Campos,
SP, Brazil
cause of drifts in the gain of the radiometer, a comparison within the radiometer to another blackbody emitter
of known temperature is necessary. Therefore, in order
to achieve significantly lower uncertainties than previous
measurements, ARCADE 2 is a high altitude balloonbased double-nulled instrument with open-aperture cryogenic optics mounted at the top of an open bucket liquid
helium dewar.
A previous generation instrument with observing channels at 10 and 30 GHz, built to test the cold open aperture design, observed in 2003 (Fixsen et al. 2004) and
is described by Kogut et al. (2004a). Following verification of the instrument concepts, the full six frequency
ARCADE 2 instrument, with observing channels ranging
from 3 to 90 GHz was designed and built, and observed
in 2005 and 2006. Scientific analysis of the 2005 and 2006
flights is presented by Singal et al. (2006), and Fixsen et
al. (2010), Kogut et al. (2010), and Seiffert et al. (2010)
respectively.
This paper describes the design and performance of
ARCADE 2. We present the radiometric and thermal
properties of the instrument necessary to achieve the results presented in the companion papers, and also describe elements of engineering design that may be useful.
2. INSTRUMENT DESIGN
ARCADE 2 reduces systematic errors through a combination of radiometer design and thermal engineering.
The instrument core is contained within a large (1.5 m
diameter, 2.4 m tall) open bucket liquid helium dewar.
Maintaining such a large volume and mass at cryogenic
temperatures in an open environment without significant
atmospheric condensation presents considerable instrumental challenges. The external calibrator, aperture, antennas, and radiometers are maintained at temperatures
near 2.7 K through the use of liquid helium tanks fed
from the helium bath at the bottom of the dewar by a
network of superfluid pumps. Boiloff helium gas is used
for the initial cool-down of components on ascent, and
directed in flight to discourage the condensation of ambient nitrogen on the aperture.
2
Singal et al.
Figure 1 shows an overview of the ARCADE 2 instrument. The corrugated horn antennas for each frequency
band hang from a flat horizontal aluminum aperture
plate at the top of the open dewar. There are seven observing channels, one each at 3, 5, 8, 10, 30, and 90 GHz,
and an additional channel at 30 GHz with a much narrower antenna beam to provide a cross-check on emission
in the antenna sidelobes. Cryogenic temperatures within
the external calibrator and internal reference loads, as
well as on components throughout the instrument core,
are read with ruthenium oxide resistance thermometers.
A carousel structure containing both a port for sky
viewing and the external calibrator sits atop the aperture plate and turns about a central axis to alternately
expose the horns to either the sky or the calibrator. The
open port hole and external calibrator are both ellipses
measuring 700 mm x 610 mm. Each radiometer measures the difference in emitted power between radiation
incident on the horn and that from an internal blackbody
reference load. The experiment performs a doubly nulled
measurement, with the radiometric temperature of the
sky compared to the physical temperature of the external
calibrator in order to eliminate systematic effects within
the radiometer to first order. The radiometer itself differences the internal reference load from the horn signal
allowing a determination of the coupling of radiometer
output to instrument temperatures and a near-nulling of
the radiometer output to reduce the effects of gain fluctuations.
The ARCADE 2 instrument is flown on a high altitude
790 Ml balloon to the upper atmosphere (37 km altitude)
in order to reduce atmospheric emission and contamination from terrestrial microwave sources to negligible levels. Balloon launch and recovery operations are handled
by the Columbia Scientific Balloon Facility (CSBF) in
Palestine, TX (31.8◦ lat, -95.7◦ long.).
Fig. 1.— ARCADE 2 instrument schematic, components not
to scale. Cryogenic radiometers compare the sky to an external
blackbody calibrator. The antennas and external calibrator are
maintained near 2.7 K at the mouth of an open bucket dewar;
there are no windows or other warm objects between the antenna
and the sky. Cold temperatures are maintained at the top of the
dewar via boil-off helium gas and tanks filled with liquid helium fed
by superfluid pumps in the bath. For observing the sky, everything
shown is suspended below a high altitude balloon.
2.1. Antenna aperture and carousel configuration
The corrugated horn antennas are arrayed on the aperture plate in three clusters, with the 3 GHz horn occupying one, the 5 and 8 GHz horns occupying a second, and
the remaining horns, the ”high bands”, occupying the
third. All horns except the 3 GHz point 30◦ from zenith
in one direction, while the 3 GHz points 30◦ from zenith
in the opposite direction. The sky port in the carousel is
surrounded by reflective stainless steel flares which shield
the edge of the antenna beams from instrument contamination and direct boiloff helium gas out of the port to
discourage nitrogen condensation in the horn aperture.
Figure 2 shows a photograph of the carousel being lowered onto the aperture plate, with the sky port and the
radiometric side of the external calibrator visible. Figure
3 shows the arrangement of the horn apertures.
The carousel is supported 1.5 mm above the aperture
plate by wheel bearings on the edge and a Kynar c plastic bearing in the center. It is turned with a motor and
chain drive, with the motor mounted outside of the dewar. The motor is commanded to run and stops, with the
engagement of a logic switch, when the proper alignment
of the carousel relative to the aperture plate is reached,
thus ensuring accurate and repeatable positioning of the
carousel. There are three carousel stopping positions for
sky and external calibrator viewing, which are, in counterclockwise order, 1) the ”5 and 8 sky” position, where
Fig. 2.— Photograph of the carousel being lowered on top of the
aperture plate by two of the authors, showing the port hole for
sky viewing and the external calibrator. The carousel turns atop
the aperture plate to expose groups of horns to the sky or to the
external calibrator. The external calibrator is a blackbody emitter
consisting of 298 cones of Steelcast absorber cast onto aluminum
cores, and the radiometric side is visible in the photograph. Horn
antennas in the core are visible hanging down from the aperture
plate.
ARCADE 2 Instrument
3
Fig. 4.— ARCADE 2 gondola in 2006 flight configuration. A
deployable lid protects the cold optics during launch and ascent.
The reflective shield screening the flight train and parachute from
view of the antennas, and the bar on which it sits, are the largest
single sources of systematic uncertainty and is measured in flight
by heating the shield.
Fig. 3.— Layout of horn apertures on the aperture plate and
the position of the sky port and external calibrator in the three
sky viewing positions of the carousel. The horns are arrayed in
three groups, with the 3 GHz occupying one, the 5 and 8 GHz
occupying another, and the high bands occupying the third. The
horns and are sliced at the apertures so that the beams point 30◦
from zenith in the directions indicated, with all horns pointing in
the same direction with the exception of the 3 GHz, which points
opposite. For each of the carousel positions, the open ellipse shows
the position of the sky port, and the filled ellipse shows the position
of the external calibrator. The arrows indicate the direction in
which the horn beams point 30◦ from vertical.
the 5 and 8 GHz horns view the sky and the 3 GHz horn
views the external calibrator, 2) the ”3 sky” position in
which the 3 GHz horn views the sky and the high band
horns view the calibrator, and 3) the ”high sky” position
where the high band horns view the sky and the 5 and 8
GHz horns view the external calibrator. The remaining
group of horns viewing neither the sky nor the external calibrator in any of these positions are viewing a flat
metal plate of the carousel. A fourth stopping position is
used on ascent to align a vent hole on the carousel with
one on the aperture plate, thereby channeling boil-off gas
to cool the back of the external calibrator.
Carousel alignment at each of the four positions is repeatable to within one millimeter on the outside circumference. The sky port and external calibrator each have
nearly the same area and shape as the aperture of the
largest horn aperture (3 GHz). Mis-alignment of more
than 1 cm would allow a section of the aperture plate
and flares to extend over the antenna aperture. The
actual alignment precision of 1 mm prevents such misalignment.
2.2. Gondola configuration
Figure 4 shows a schematic of the entire payload. The
dewar is mounted in an external frame supported 64 m
below the balloon, and boxes containing the read out
and control electronics and batteries are mounted on the
frame. The external frame is suspended by two vertical cables from a horizontal spreader bar 1.14 m above
the top of the dewar, which itself is suspended by two
cables from a rotator assembly. The rotator maintains
the rotation of the payload below the balloon at approximately 0.6 RPM. The rotator assembly is suspended from
a truck plate, above which is the flight train. Reflector
plates of metalized foam are mounted on the spreader bar
to shield the edge of the antenna beam from the flight
train. Figure 5 shows a photograph of the payload prior
to a launch. The total mass at launch, including liquid
helium, is 2400 kg.
The dewar can be tipped to angles of 2◦ from vertical,
changing the angle of the antennas with respect to the reflector plates and flight train, by moving the battery box
outward from the frame. A fiberglass lid mounted on the
frame is closed to cover the dewar on ascent and descent
and opened for observations. Thermometry, heater, and
other signals are interfaced between the dewar and the
exterior electronics box via cabling and a collar of insulated connectors at the top of the dewar.
Three-axis magnetometers and clinometers mounted
on the frame, along with GPS latitude, longitude, and
altitude data recorded by CSBF instruments, allow the
reconstruction of the pointing of the antenna beams during flight. During the 2006 flight, the magnetometers
failed, and the pointing was reconstructed with a combination of the clinometers and radiometric observations
of the Galactic plane crossings. The uncertainty in the
reconstructed pointing (a small fraction of a degree) is
small compared to the beam size.
2.3. Horn Antennas
The corrugated antennas for six of the channels have
11.6◦ full width at half power Gaussian beams, while
the antenna for the 30 GHz narrow beam channel has a
4◦ full width at half power Gaussian beam. The horns
4
Singal et al.
Fig. 5.— Photograph of the ARCADE 2 instrument just prior
to 2006 flight. The instrument core is contained within the large
(1.5 m diameter, 2.4 m tall) bucket dewar. The lid is shown closed
for launch. The electronics box containing the warm stages of the
radiometers and the payload electronics is mounted to the left of
the dewar, and the three-axis magnetometers to determine orientation relative to the Earths magnetic field are seen to the left of
the electronics box at the edge of the frame. The reflector plate
is visible at the top of the photo above the lid. Cardboard pads
at the four bottom corners absorb some of the impact when the
payload hits the ground upon termination of the flight.
are sliced at the aperture to point 30◦ from zenith when
hung from the flat aluminum aperture plate. This is so
the antenna beam boresights are directed away from the
flight train and so that they trace out a circle 60◦ on
the sky as the dewar rotates below the balloon. In order
to achieve the narrowest possible beam at 3 GHz given
the spatial constraints, a curved profiling of the horns
was employed. A full discussion of the horns is given by
(Singal et al. 2005).
The horns were designed using mode matching simulation software. The beam pattern for the fabricated 10
GHz horn was mapped over greater than 2π steradian
at the Goddard Electromagnetic Anechoic Chamber test
range at the NASA Goddard Space Flight Center. The
measured beam pattern in the plane containing the maximal effect of the 30◦ aperture slice is shown along with
the predicted one from design in Figure 6. The aperture
slice has a minimal effect on the symmetry of the beam,
and is seen primarily in the first sidelobe response, depressing the response on the long side and increasing the
response on the short side. This effect is a few dB, overlayed on a first sidelobe response that is 20 dB below the
main beam. Response in far sidelobes below the aperture
plane is surpressed by more than 50 dB. Because of the
large physical sizes involved, it is not practical to measure
the beam pattern for the lower frequency horns, or any of
the horns once installed in the instrument, but the correctness of the simulated beam patterns, the low sidelobe
response, and the negligable effect of the slice has been
demonstrated with the 10 GHz horn beam map. Fur-
Fig. 6.— Measured response for the 10 GHz horn, at the center
band frequency of 10.11 GHz (dots) as a function of angle, scanning
in the plane where the effect of the 30◦ aperture slice is maximal.
The predicted simulated beam pattern for an unsliced horn is also
shown (solid line). The sidelobe response is low, and the effect of
the slice is seen primarily in the first sidelobe. Far sidelobes are
further supressed. The beam pattern does not vary appreciably
over the frequency band.
thermore, in the ARCADE I instrument, we were able
to measure the beam of its 10 GHz horn both prior to
and after installation in the instrument, and the far sidelobe response was identical (Kogut et al. 2004a), further
demonstrating consistency.
2.4. Radiometers
Figure 7 shows a block diagram of the radiometers. Radiation incident from the horn goes through a compact
circular to rectangular waveguide transition (Wollack
1996). Typical reflections from the horn and transition
system are less than -30 dB across the entire radiometer band. A switch chops at 75 Hz between radiation
from the horn and that from the internal reference load.
At 3 and 5 GHz, a micro-electrical-mechanical system
(MEMS) switch is used, while at the other channels a
latching ferrite waveguide switch is used. At 3 and 5
GHz, the internal reference load is a simple coax termination stood off with a stainless steel coax section. At 8
and 10 GHZ, the reference load is a wedge termination
in waveguide with a layer of Steelcast, a microwave absorber consisting of stainless steel powder mixed into a
commercially available epoxy (Wollack et al. 2008) cast
onto an aluminum substrate. At 30 and 90 GHz, the reference load is a split block wedge configuration of Steelcast in waveguide. These wedge termination and splitblock reference loads have reflected power attenuation of
more than 30 dB across the entire frequency band and
are described by Wollack et al. (2007). We measure the
temperature of the absorber in each reference load with
ruthenium-oxide resistance thermometers (see §3.1).
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 amplified by the cold HEMT and propagates, via coaxial cable at the low frequencies and waveguide at 30 and 90 GHz, out of the dewar to the warm
stage contained in the electronics box. The warm stage
ARCADE 2 Instrument
Fig. 7.— Block diagram of ARCADE 2 radiometers. The demodulated output is proportional to the difference in temperature
between radiation from the antenna and the internal reference load.
The warm stage of the radiometers operate at ∼280 K.
Fig. 8.— Photograph of cold stage of ARCADE 2 8 GHz radiometer. The throat of the horn antenna bolts to the open circular waveguide end. The internal reference load of this radiometer
is a wedge termination in waveguide, which is surrounded by the
insulative cylinder visible in the photograph.
features a warm HEMT amplifier, an attenuator to eliminate reflections and tune the output power to match
downstream components, a band pass filter to select the
desired frequency band, a second warm HEMT amplifier, and a power divider to split the signal into a high
and a low frequency channel. Each of the high and low
channels has a bandpass filter, a detector diode, and an
audio frequency preamp, outputting a voltage level corresponding to the power of the radiation incident on the
diode. The voltage signal is then carried to an electronics board where a lockin amplifier demodulates the signal
in phase with the switch, integrates it for .533 seconds,
and digitizes it. In this way, the final output is proportional to the difference in temperature between what
the radiometer is viewing and the internal reference load,
averaged over the integration period. There is also a total power output signal in the data stream which is not
demodulated by the lockin amplifier.
Tables 1 and 2 show selected radiometer properties
5
and figures of merit. Figure 8 shows a photograph of
a radiometer cold stage. The custom designed radiometer components are the horns, the circular to rectangular
waveguide transitions (Wollack 1996), the ferrite waveguide switches, and the cold amplifiers at 3, 5, 8, and 10
GHz.
The cold stages of the radiometers are housed in pans
that can hold liquid helium, with a cascading pond system linking the pans. The components are stood off from
but thermally linked to the liquid, with the liquid providing cooling to maintain the components at temperatures
below 4 K. For the horn throats and internal loads, temperatures within this range are selected by SPID control,
as described in §3.1, while the switches and HEMT amplifiers run near the liquid temperature.
2.5. External calibrator
ARCADE 2 determines the radiometric temperature of
the sky by using a full-aperture blackbody external calibrator as an absolute temperature reference. To function
as a blackbody emitter, it should have very low power
reflection across the entire ARCADE 2 frequency range.
The emitting surfaces of the external calibrator need to
be close to isothermal, and, while located on the carousel
at the top of an open bucket dewar, have to be precisely
temperature controlled within the range of 2.5 K to 3.1
K.
The external calibrator is based on the successful
model from COBE/FIRAS of a full aperture calibrator
that is absorptive and isothermal, achieved through a
combination of material and geometry (Mather et al.
1999). The ARCADE 2 calibrator is as good as the
COBE/FIRAS one radiometrically, having a similar level
of reflections, although the ARCADE 2 frequency coverage extends to much longer wavelengths. The calibrator
has a reflected power attenuation of less than -55 dB in
the range from 5 to 90 GHz and of -42 dB at 3 GHz. Measured reflected power attenuation values are presented in
1 and the external calibrator’s radiometric performance
is described in detail by Fixsen et al. (2006).
The calibrator consists of 298 cones, each 88 mm long
and 35 mm in diameter at the base, of Steelcast absorber
(see §2.4) cast onto an aluminum core. The aluminum
provides for enhanced thermal conductivity and therefore reduced thermal gradients, which is augmented by a
copper wire epoxied onto the end of the aluminum core
and running almost to the tip. The radiometric side of
the external calibrator is visible in Figure 2. Figure 9
shows a cut away view of a cone.
The cones are mounted pointing downward on a horizontal aluminum plate, behind which are 50 alternating layers of alloy 1100 aluminum and fiberglass sheets,
to give a low vertical but high horizontal thermal conductivity, to the end of getting the different cones as
isothermal as possible. This aluminum plate is welded
to an aluminum shielding which surrounds the elliptical circumference of the area of cones so that the cones
are surrounded on the top and sides by an isothermal
metal surface maintained at temperatures near 2.7 K.
Behind the stack of alternating layers is another horizontal aluminum back plate, and behind that a half inch
gap to allow room for bolt heads, heating elements, and
wiring. In the 2006 configuration the entire calibrator is
surrounded, except on the radiometric side, by a tank of
6
Singal et al.
TABLE 1
Table of selected radiometer hardware specifications
Low Band (GHz)
High Band (GHz)
Switch Typeb
Cold Amplifier Mfg.
Cold Amp Model #
Cold Amp Serial #
Cold Amp Gain (dB)
Warm Amp 1 Gain (dB)
Attenuator (dB)
Warm Amp 2 Gain (dB)
Detector Power (dBm)
X-Cal reflectionf (dB)
3 GHz
3.09-3.30
3.30-3.52
MEMS
Berkshire
S-3.5-30H
105
40
35
-26
35
-21.3
42.4
5 GHz
5.16-5.50
5.50-5.83
MEMS
Berkshire
C-5.0-25H
106
29
33
-13
33
-20.6
55.5
8 GHz
7.80-8.15
8.15-8.50
Cir
Berkshire
X8.0-30H
108
40
33
-26
33
-28
58.6
10 GHz
9.2-10.15
10.15-10.83
Cir
Berkshire
X10.0-30H
101
40
30
-20
31
-24.6
62.7
30 GHz
28.5-30.5
30.5-31.5
Cir
Spacek
26-3WCc
5D12
26
46
-3
-25.9
55.6
30#a GHz
28.5-30.5
30.5-31.5
Cir
Spacek
26-3WCd
6C21
20
22
-10
23
-42.5
xxg
90 GHz
87.5-89.0
89.0-90.5
Cir
JPL
90 GHz Amps
W82&W105d
52
-7e
-6
40
-24.6
56.6
a
30# designates the 30 GHz channel with the narrower antenna beam.
Switches are either MEMS (Micro-Electro-Mechanical System) or ferrite latching waveguide circulator switches (Cir).
These model numbers have the prefix ”SL315-”.
d At 90 GHz there two are cold amplifiers in series.
e The 90 GHz channel warm stage uses a mixer, with a 79.5 GHz local oscillator, to translate it to 8-11 GHz, and all following components
operate in that frequency range.
f This is the measured attenuation of reflections from the external calibrator, when it is viewed with the horn antenna for the channel.
This values presented here were measured in ground testing, and the measurement is described by Fixsen et al. (2006) .
g This external calibrator reflections were not measured directly with the 30# channel antenna, but they are assumed to be even lower
than for the other 30 GHz antenna.
b
c
two thermometers each, for a total of 26 thermometers
within the absorber. Resistance heaters are mounted on
the rear aluminum plate, and 7 additional thermometers
are affixed at various points. Thermometry and heater
signals from the calibrator, as well as from elsewhere on
the carousel, are carried through a hollow tube in the
axis of the carousel to the rest of the core inside the
dewar. The true emission temperature of the external
calibrator is a volume integral over the physical temperature distribution weighted by the absorber emissivity
and the electric field distribution from the antenna aperture. We approximate the true emission temperature as
a linear combination of the temperatures as measured by
the thermometers, with weights derived from the actual
flight data as described by Fixsen et al. (2010).
Fig. 9.— Cut away view of external calibrator cone showing
internal structure and predicted linear temperature profile from the
2006 flight, with the base thermally fixed and the cone immersed in
a cold atmosphere. The cone consists of Steelcast absorber (black)
cast onto an aluminum core (light grey). There is also a copper
wire running from the tip of the aluminum core to nearly the tip
of the cone (darker grey). The hole in the base is threaded for a
mounting bolt to affix the cone to the aluminum back plate. The
measured thermal gradient is 600 mK from the base to the tip of
the cones. 98% of the total gradient is in a region near the tip
containing 3% of the absorber, and the mean depth for absorption
varies with frequency.
liquid helium, which is thermally coupled with stainless
steel standoffs to the rear aluminum plate. This layer of
liquid helium intercepts any external heat load incident
on the external calibrator from the back and sides.
Temperatures are monitored in 23 of the cones with
thermometers embedded in the Steelcast absorber at
varying depths and radii. Three of the cones contain
2.6. Payload electronics
The main payload electronics consists of the cryogenic
thermometer resistance readout and control for heaters,
as described in §3.1, as well as radiometer lockin and integration, voltage readout for the various payload analog
devices, including the magnetometers, clinometers, and
ambient temperature transducers, digital logic level readout, generation of the switch driving current, and generation and amplification for commandable signals to the
lid, rotator, tilt, and carousel movement motors. These
functions are performed by custom electronics boards.
Typical power required for the electronics is 220 W, with
peak capacity 1800 W.
The digital data stream is relayed via the RS-232 serial data standard to the Consolidated Instrument Package (CIP) provided by CSBF, which transmits it, along
with data from the CSBF instruments and the video signal, to the ground. Each 1.067 s record of data consists
of digitized counts corresponding to a voltage reading
across every cryogenic thermometer, the voltage reading across the reference resistors on each thermometer
readout board (see §3.1), the voltage output of the various analog payload devices, the SPID control parameters
for every controlled heater, digital logic levels, and two
ARCADE 2 Instrument
7
TABLE 2
Table of selected measured radiometer performance specifications from the 2006 flight
Channel
3 GHz Lo
3 GHz Hi
5 GHz Lo
5 GHz Hi
8 GHz Lo
8 GHz Hi
10 GHz Lo
10 GHz Hi
30 GHz Lo
30 GHz Hi
30# GHz Lo
30# GHz Hi
90 GHz Lo
90 GHz Hi
Cntr Freq
(GHz)
3.15
3.41
5.33
5.67
7.97
8.33
9.72
10.49
29.5
31.0
29.5
31.0
88.2
89.8
Bandwidth
(MHz)
210
220
340
330
350
350
860
680
2000
1000
2000
1000
1500
1500
Trcvr a
(K)
5.5
6.5
6
6
10
8
13
11
75
72
270
340
44
38
Offsetb
(mK)
180
35
-210
-200
6
11
180
35
-30
-15
32
38
-75
-95
c
Noise (pre-flt)
√
(mK s)
7.1
7.1
3.2
3.5
1.4
1.4
6.8
5.3
21.5
14.9
18
14
5.2
5.7
d
Noise (rsduls)
√
(mK s)
9.3
7.8
5.5
6.1
3.7
3.7
208.2
103.3
885.0
418.5
50.5
25.3
f
Noise (map)
√
(mK s)
11.8
10.1
5.5
5.2
3.0
3.0
206.9
103.4
880.4
406.1
42.0
27.1
a
Trcvr is the receiver noise temperature of the amplifier, a figure of merit that is equal to the temperature that would be observed by a
total power radiometer containing the amplifier and viewing a source at a temperature of absolute zero. This values presented here were
measured in ground testing prior to the 2006 flight.
b The constant offset is the radiometer output when the internal reference load and the object being viewed are at the same temperature,
multiplied here by the gain to be expressed as a temperature.
c This is the white noise as measured in ground testing prior to the 2006 flight.
d This is the white noise determined from the residuals of the data from the 2006 flight.
e This is the white noise as determined from the sky map variance in the 2006 flight.
readings of the demodulated and total voltage output of
every radiometer channel. Each 1.067 s also allows four
two byte commands to be transmitted via the CIP to
the instrument and executed. The commands include
the setting of SPID parameters, lockin amplifier gains,
and motor movement.
A video camera mounted on the spreader bar above the
dewar allows direct imaging of the cold optics in flight.
Two banks of light-emitting diodes provide the necessary
illumination. The camera and lights can be commanded
on and off, and we do not use data for science analysis
from times when they are on.
2.7. Flight operations
When the instrument is ready for flight, the lid is closed
and the dewar is cooled with nitrogen to around 100 K.
We then fill the dewar with around 1900 liters of liquid
helium, which takes several hours. We await a launch
opportunity, with the helium level topped off each day.
Approximately 200 litres of liquid helium are boiled off
per day. To avoid freezing the lid to the dewar with ice
accumulation while the payload awaits flight, we position
an outflow hose leading from a vent hole in the top of
the lid to a position close to level with the dewar rim.
This maintains some positive pressure and helium gas
outflow at the lid-dewar interface, which discourages the
condensation of ambient water vapor.
We launch the instrument with the carousel in the ”ascent” position that aligns the vent holes in the aperture
and carousel, which directs boil-off gas across the back
of the external calibrator, providing a powerful cooling
source for its large thermal mass. On ascent, about 1/3
of the helium is boiled off, with the remainder cooled continuously to 1.5 K when float altitude is reached. We turn
on the superfluid pumps once the helium bath is below
the superfluid transition temperature of 2.177 K. After
three hours of ascent, float altitude is reached and we
open the instrument lid for observing. During observing,
we move the carousel to a new position about once every
five minutes. We observe for around four hours, limited by the westward drift of the balloon out of range of
telemetry. The lid is closed for descent, and the payload
is severed from the balloon and returns to the ground
on a parachute, to be recovered by CSBF staff. At the
end of the 2006 flight, nearly 800 liters of liquid helium
remained in the dewar, or enough for 6 hours more of
observation.
In the 2005 flight, the carousel became stuck in one position soon after observing commenced. The cause was
traced to the output torque of the carousel motor exceeding the maximum torque of the attached gear box,
stripping the gears, and was remedied for the 2006 flight.
In the 2006 flight, the most significant instrumental failure was with the 5 GHz MEMS switch, rendering data
from that channel not useful for science analysis.
3. CRYOGENIC PERFORMANCE
The ability to measure and control the temperature
of cryogenic components while in the presence of a variable helium gas flow and potentially exposed to ambient sources of warming is the limiting factor in the precision of radiometric temperature measurements made
with ARCADE 2. Thermal measurement and control is
the most important and challenging facet of the experiment.
ARCADE 2 requires component temperatures to be
maintained near 2.7 K. Some components are controlled
passively by being thermally sunk to liquid helium and
exposed to cold helium gas. Other components are actively temperature controlled, with both a coupling to
liquid helium and controllable resistance heaters. The
gas cooling is an undesirable perturbative effect on actively controlled components. Liquid helium is moved to
needed areas outside of the liquid helium bath, such as
the aperture plate and carousel at the top of the dewar
and the cold stages of the radiometers. At ballooning
8
Singal et al.
altitudes liquid helium is well below the superfluid transition temperature and does not respond to mechanical
pumping, so superfluid pumps 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 several meters. ARCADE 2 features 13 superfluid pumps which
can move 55 liters per minute to a height of 2.5 m.
3.1. Thermometry and temperature control
We measure 104 cryogenic temperatures in the ARCADE 2 payload. 26 are within the external calibrator
cones, 7 are elsewhere on the calibrator, another 19 are
at various points on the carousel, 26 are on components
of the radiometers including the internal reference loads,
switches, horns, and the pans in which the radiometers
sit, and 26 are at various other points within the dewar.
We measure these temperatures using four-wire AC resistance measurements of ruthenium-oxide resistors, whose
resistance is a strong function of temperature below 4 K.
The thermometers are excited by a 1 µA 37.5 Hz square
wave current, and signals are carried through cold stages
via a sequence of brass, copper, and manganin cryo-wire.
The resistances of the thermometers are read by custom
electronics boards (Fixsen et al. 2002), which output a
digitized voltage level for every thermometer in every
data record. The boards also contain five on-board resistors of known resistance spanning the dynamic range of
the cryogenic thermometer resistances that are excited
with the same current as the cryogenic thermometers.
In this way, a reliable digital voltage counts to resistance
conversion is obtained in every data record to express the
measured resistance of the cryogenic thermometers. The
measured signals can then be converted in software to
temperatures using look-up tables containing the resistance versus temperature curves for each thermometer.
The resistance versus temperature curves for the
ruthenium-oxide thermometers are determined in ground
testing against a thermometer of identical design previously calibrated by NIST. This style of thermometers
has demonstrated calibration stability to within 1 mK
over four years (Kogut et al. 2004a), verified with the
observed lambda superfluid transition as an absolute reference. Thermometer self-heating is negligible (less than
1 nW) and is included in the resistance-to-temperature
calibration process.
Desired temperatures are maintained in key places,
such as the radiometer internal reference loads and
switches, horn throats, and external calibrator, through
resistance heaters under SPID (set point, proportional,
integral, and differential) control. The values S,P,I, and
D are set in real-time by the user for each of 32 SPID
channels, and the desired set point temperature as well as
the current actual temperature are expressed in counts,
with the user performing the conversion between counts
and temperature as necessary. The output voltage level
for each SPID channel is recalculated once per record
(1.067 s) in firmware. Figure 10 shows the accuracy and
precision of the SPID temperature control of the 5 GHz
internal reference load. This internal reference load has
a relatively small mass (~10 g) and is located in a relatively stable thermal environment. Figure 11 shows the
SPID temperature control of the external calibrator, a
much larger system that is subject to changeable ther-
Fig. 10.— Temperature of the 5 GHz internal reference load vs.
time for a section of data from the 2006 flight showing SPID temperature control of the load. The load has a relatively small mass
and is located in a steady thermal environment. The temperature
was commanded to be set to 2.730 K at 51 seconds, and back to
2.724 K at 138 seconds. In general, it is important that the temperature of components under SPID control be steady and known, but
not that specific commanded temperatures be obtained exactly.
Fig. 11.— Temperature vs. time of a cone in the external calibrator for a section of data from the 2006 flight showing SPID
temperature control of the external calibrator. The external calirator is a relatively large system exposed to changeable thermal
conditions. The ringing is beneficial as it is slow compared to the
radiometer response time and allows the sampling of a wider range
of calibrator temperatures.
mal dynamics.
3.2. External calibrator thermometer calibration
Special care is taken in determining the resistance versus temperature curves for the ruthenium-oxide thermometers embedded in the cones of the external calibrator, as errors and uncertainties in these curves lead
directly to errors and uncertainties in the measured radiometric temperature of the sky. A specific calibration
setup is employed for these thermometers to minimize
any possible thermal gradients and wiring differences between calibration conditions and flight conditions. In this
configuration, the cones and a NIST calibrated standard
ARCADE 2 Instrument
Fig. 12.— The upper plot shows a resistance versus temperature curve for a typical external calibrator cone thermometer. The
lower plot shows the fractional error in temperature from removing
each temperature point in sequence to form a reduced information
curve and feeding that reduced curve the resistance of the removed
temperature point. We take the magnitude of this error as an estimate of the uncertainty in the resistance versus temperature curve
arising from sources other than statistical uncertainty and the capacitance correction. The different shaped points represent data
from the three different calibration runs.
thermometer are mounted on an 1100 series aluminum
plate inside of an evacuated stainless steel pressure vessel,
which is itself submerged within the liquid helium bath
of a large (.3 m diameter, 1.5 m tall) test dewar. The
aluminum plate is thermally linked to the bath through a
copper rod which passes through a superfluid-tight hole
in the pressure vessel. The liquid helium bath of the
test dewar is pumped in stages, with the pumping valve
opened some amount and then the bath left alone to
eventually equilibrate to some steady temperature. In
this way, once a steady bath temperature is reached,
with the evacuated vessel submerged in an isothermal
liquid helium bath, there are no obvious sources of thermal gradients in the system. The procedure leads to a
steady thermal situation such that thermometer outputs
show no coherent movement over timescales of ten minutes. The aluminum plate, and therefore the cones and
NIST standard thermometer, should be quite isothermal
when the bath has equilibrated at a given temperature,
as the aluminum plate is thermally lined to the bath and
within an evacuated vessel with the walls at nearly the
same temperature as the plate.
This process leads to discrete temperature points
where a definite temperature, read by the NIST standard
thermometer, can be associated with a measured resistance for each of the cone thermometers. The fourteen
total calibration points from three separate runs can be
combined to form a reasonably dense calibration curve
with values between 2.5 and 4.2 K.
As we measure the resistance of the thermometers using a 37.5 Hz square wave excitation, this measured resistance includes the effect of shunt capacitance in the
harnessing connecting the wires to the board. Our thermometer calibration procedure assigns a temperature to
each measured resistance using identical harnessing and
electionics boards as in flight., so the effect of the shunt
capacitance is automatically acounted for in the resis-
9
tance vs. temperature curve obtained for each thermometer. However, the NIST calibrated thermometer,
which is read by the flight thermometer boards and flight
harnessing in our calibration setup, was itself calibrated
with DC readout techniques featuring no shunt capacitance. To correct for the small effect on the measured
resistance of the NIST calibrated thermometer, an estimate of the effect of the shunt capacitance of the calibration setup is formed using data taken with 20 resistors
of known value placed at the end of the harnessing. This
yields a linear relationship between the actual resistance
and the error in the measured resistance. The shunt capacitance is around 300 pF which results in an offset in
the measured resistance of 8 Ω when measuring a 20000
Ω resistor, which is the resistance of a thermometer at
around 2.7 K. An 8 Ω offset results in a 2 mK offset
in the thermometer reading. This small estimated effect
is then subtracted from the raw NIST calibrated thermometer data, with the standard deviation of the linear
fit providing the uncertainty in this applied correction.
The total uncertainty in the determined resistance versus temperature curves of the cone thermometers consists
of contributions from raw statistical uncertainty, from
the uncertainty in the shunt capacitance correction for
the NIST calibrated thermometer, and that from any
other sources. The first two are straightforward to determine, with statistical uncertainty well below 1 mK except at above 3.4 K where it is 1 mK, and the capacitance
correction uncertainty well below 1 mK at temperatures
below 3 K and rising to 3 mK at 3.4 K. We estimate the
remaining uncertainty from other sources using the figure of merit of how well data from the three separate test
runs agree with each other. Each run contains only four
or five discrete temperature points. We combine all 13
points to form a single resistance vs. temperature curve
for each thermometer. To estimate the remaining uncertainty, we remove each temperature point in sequence
and compare the point to the curve determined from the
12 remaining points. This is a test of the consistency
of the combined curve and of the three runs. Figure 12
shows a resistance versus temperature curve for a typical
external calibrator cone thermometer, and the fractional
error in temperature at each temperature point determined by removing the point.
We take the magnitude of the error in reproducing
the temperature of the removed point given the reduced
information curve as an estimate of the uncertainty of
that point in the curve, neglecting statistical and capacitance correction uncertainty. Figure 13 shows an average
over all cones of the uncertainty due to all three sources
and the total uncertainty, taken to be the quadrature
sum. The uncertainties in temperature generally increase
steeply above 3 K because at higher temperatures the resistance versus temperature curves flatten. At 2.7 K, the
average total uncertainty in the resistance versus temperature curves is 1.3 mK.
3.3. Heat flow
ARCADE 2 successfully maintains critical components
near 2.7 K at and near the top of an open bucket dewar at
37 km altitude. This is achieved by moving large quantities of superfluid liquid helium so that each component
is either passively controlled at the bath temperature, or
has a thermal path to the liquid to allow active temper-
10
Singal et al.
Fig. 13.— Uncertainty in the resistance versus temperature curve
determination for the cones of the external calibrator, as a function of temperature. The result presented is averaged over all of
the cones. The dashed line is statistical uncertainty, the dotted line
is the uncertainty due to the applied correction to the NIST calibrated thermometer for shunt capacitance in the read-out wiring,
and the dash-dot line is the magnitude of the error in reproducing
the temperature of a calibration point given the resistance of that
point and a reduced information curve not containing the point.
The solid line is the quadrature sum of the three sources, which is
taken to be the total uncertainty.
ature control. Other components of the instrument are
not thermally controlled, but generally come to temperatures between 1.5 K and 20 K depending on their location
relative to external heat sources. We monitor temperatures on these non-critical components to demonstrate
that their fluctuations do not affect the radiometer output. The overall thermal performance of the entire instrument is well measured and is understood.
The thermal behavior of the instrument core in flight
results from the balance of warming and cooling power.
Cooling is provided by both pumped liquid helium and
boil-off helium gas. The boil-off gas is channeled out of
the dewar through the 38 mm perimeter gap between the
aperture plate and the inner dewar wall. There is also
one 100 mm diameter gas vent port each on the aperture
plate and carousel, which, when aligned, channel boil-off
gas up across the back plate of the external calibrator and
out through a path in the foam insulation covering the
carousel. This configuration is used to direct maximum
helium gas flow to the calibrator back upon ascent.
The sources of warming in flight are A) infrared radiation and conduction from the warmer ambient atmosphere, B) radiation and conduction from warm components of the instrument such as the dewar walls or the
carousel rotation drive chain, and C) SPID heaters as described above. There are three distinct paths by which
ambient warmth can heat the core: 1) warming of the
carousel from above, 2) warming of the aperture plate
from above through the sky port of the carousel, and 3)
warming via radiation from the warm wall of the dewar
near the top. To reduce the heat load from the dewar
wall, there are three concentric sheet aluminum baffles
inside the dewar wall extending .6 m down from the top
of the dewar. In addition, at the very top there is a five
layer Vacuum Deposited Aluminum with Dacron concentric veil extending 200 mm down from the top outside
of the outermost baffle. To reduce the heat load to the
carousel from above, it is covered in 1 m3 of foam insulation (Fomo Handi-Foam SR, a two-component slow-rise
polyurethane foam), which is topped with a .25 mm reflective aluminum sheet.
In the 2006 configuration, the aperture plate was
cooled directly by six pump-fed liquid helium tanks
mounted on the underside in various places. Being relatively thin and covering a large area, the aperture plate
is subject to thermal gradients even with liquid helium
tanks in contact at discrete places. The large horns at
3,5, and 8 GHz have sheet metal ’collars’ on the exteriors, providing a sheath for liquid helium to maintain
each horn aperture at the bath temperature. Aperture
plate temperatures in flight vary between 1.5 K and 10 K
depending on location and time, while temperatures on
the carousel vary between 4 K and 20 K. In general, the
aperture plate is colder than, and cools, the carousel, as
the aperture plate is less warmed by heating from above
and is cooled by helium tanks.
The major temporal temperature variations in flight
result from differing gas flow dynamics in the different
positions of the carousel. In the position in which the
5 and 8 GHz horns are viewing the sky, the vent holes
of the aperture plate and carousel are almost aligned,
causing increased flow through the holes and therefore
decreased flow to the perimeter and a resulting warming
of perimeter areas of the aperture plate and carousel. In
the other two carousel positions this effect is not present
and a sub-dominant effect is observed. The ”high sky”
position exposes the 10 through 90 GHz horns to the
sky. Because these horns have relatively small apertures
there is more metal of the aperture plate exposed to the
sky port and therefore more infrared radiation incident
on it from above. This effect is not present when the
3 GHz horn views the sky, as the 3 GHz horn aperture
completely fills the sky port and is maintained at the
bath temperature by direct contact with the superfluid
liquid helium collar on the exterior of the horn.
There are also transient effects when the carousel
moves from one position to another. Portions of the
carousel perimeter that pass near where the carousel
drive chain enters through the connector collar are momentarily warmed, and the back plate of the external
calibrator is momentarily cooled as the calibrator passes
over the aperture plate vent hole, briefly channeling gas
between the calibrator’s aluminum shielding and the surrounding liquid helium tank and then across the back
plate. These effects on radiometrically active parts of
the instrument used for science analysis are small, and
data from times in which the carousel is moving are not
used for science analysis.
The observed temperature gradient among the concentric baffles inside the dewar wall, from 9 K at the inner
baffle to around 30 K at the outer one, indicates that
the heat leak to the aperture plate from the warm dewar
walls is reduced, via the baffling and blow-by of boil-off
gas, to the level of a few mW. The observed gradient
through the foam topping the carousel indicates that the
heat load from above to the carousel is on the order of
30 W, and this is roughly consistent with the source being infrared radiation from a 200 K body incident on the
ARCADE 2 Instrument
11
aluminum sheet topping the foam.
In the 2006 flight, we experienced problems with temperature controlling the 3 and 8 GHz internal reference
loads, both becoming fixed at near the liquid helium bath
temperature for significant periods of the flight. In the
case of the 3 GHz load, this was likely caused by the
insulative housing falling off, which was observed upon
completion of the flight. The 5 GHz load, which was of
the same construction as the 3 GHz load, did not experience this problem, most likely because the 5 GHz load
was located above the liquid Helium level during ascent.
3.4. External calibrator thermal performance
In both the 2005 and 2006 flights, all radiometrically
active parts of the external calibrator were maintained
within 300 mK of 2.7 K. However, the external calibrator
cones displayed significant gradients from base to tip. In
the 2005 flight, the aperture plate, which was at temperatures significantly warmer than the calibrator, warmed
the tips relative to the base. The situation was reversed
in the 2006 flight, where the presence of the underside
liquid helium tanks caused the aperture plate to run
significantly colder and the tips of the cones were cold
relative to the bases. The wrap-around tank of liquid
helium surrounding the external calibrator successfully
intercepts all heat loads from the top or sides. The only
un-controlled thermal link is between the cones on the
radiometric side of the calibrator and the horn apertures
and aperture plate, which is responsible for the thermal
gradients observed in the calibrator. In the 2006 flight,
the maximum base to tip gradient was 600 mK. Figure 9
shows a predicted temperature profile based on this gradient. In light of the observed effects, a more uniform
temperature could be achieved through active temperature control of the aperture plate with heaters to maintain it at near 2.7 K.
3.5. Atmospheric condensation
The potential problem with a cold open aperture is
condensation from the atmosphere. Condensation on the
optics will reflect microwave radiation adding to the radiometric temperature observed by the instrument in an
unknown way. In the course of an ARCADE 2 observing flight, the aperture plate and external calibrator are
maintained at cryogenic temperatures and exposed open
to the sky for over four hours. Figure 14 shows time averaged video camera images of the dewar aperture taken
two hours apart during the 2006 flight. No condensation
is visible in the 3 GHz horn aperture despite the absence
of any window between the horn and the atmosphere.
It is seen that the efflux of cold boiloff helium gas from
the dewar is sufficient to reduce condensation in the horn
aperture to below visibly detectable levels.
4. INSTRUMENT PERIPHERALS CONTRIBUTION TO
MEASURED RADIOMETRIC TEMPERATURE
The instrument core is in the far sidelobes of the antenna beams so its thermal emission to the radiometers
is negligible. However the flight train, consisting of the
parachute, ladder, truck plate, FAA transmitter, and balloon are directly above the radiometer only 30◦ from the
center of the beams. Since the flight train is complicated and moves with the balloon rather than the gondola, a reflector constructed of aluminum foil covered
Fig. 14.— Time averaged video camera images of the dewar
aperture during two periods of the 2006 flight where the 3 GHz
horn antenna was viewing the sky. The time and date stamps in
the video image do not correspond to the actual time and date.
The bottom image was taken two hours after the top image. As
can be seen by comparing the faintly visible grooves, there is no
visible condensation in the horn antenna aperture over a two hour
period.
foam board was attached to the gondola to hide these
components from the antenna beams and instead reflect
the sky into the antennas. The V-shaped reflector shield
and the spreader bar on which it is mounted can be seen
in Figure 5. The total expected emission does not change
much because of the presence of the reflector, but it is
much easier to compute and more stable.
We convolve the measured antenna pattern with the
positions and emissivity estimates of the reflector plate,
spreader bar, balloon, and several other components to
estimate the radiometric temperature contribution from
each in each band. We consider both thermal emission
from the components themselves and reflection of thermal emission from the 300 K ground. These results are
presented in Table 3. We conservatively estimate an uncertainty of 30% for these values.
Near the end of the flight the reflector was heated from
240 K to 300 K to look for the emission signal from the
reflector shield in the radiometer outputs. The predicted
change due to the heating is smaller than the uncertainty.
No measurable signal was detected, putting a limit of the
contribution from the reflector shield to ∼3 times the
estimated signal. The geometric factor is highest for the
8 GHz radiometer as its beam boresight is closest to the
reflector shield and spreader bar.
5. DISCUSSION
The 2005 and 2006 flights have demonstrated the viability and utility of open-aperture cryogenic optics for
12
Singal et al.
TABLE 3
Estimates of the radiometric temperature contribution from the balloon and other components visible in the antenna
beams, for channels used for science analysis in the 2006 flight. See §4. All estimates are in mK.
Component
Balloon
Reflector shield
Lightsa
Spreader bar
Upper Suspensionb
Lower Suspensionb
Cold Flarec
Total
3L
0.1
1.5
0.3
8.5
0.1
0.0
0.3
10.8
3H
0.1
0.9
0.2
4.3
0.0
0.0
0.3
5.8
8L
0.2
4.2
0.0
31.4
0.2
0.2
0.3
36.6
8H
0.2
4.5
0.0
36.6
0.2
0.2
0.5
42.2
10L
0.3
0.5
0.0
1.7
0.0
0.0
0.3
2.9
10H
0.3
0.5
0.0
1.1
0.0
0.0
0.3
2.3
30L
1.9
0.8
0.0
1.3
0.0
0.0
0.3
4.4
30H
2.0
0.8
0.0
1.7
0.0
0.0
0.3
4.8
90L
14.8
0.8
0.0
1.3
0.0
0.0
0.3
17.2
90H
14.8
0.7
0.0
1.1
0.0
0.0
0.3
16.9
a A bank of lights can be commanded on for when the video camera views the aperture. Data when the lights are on are not used for
science analysis. This emission estimate is for the lights off.
b The lower suspension cables suspend the dewar from the spreader bar, while the upper suspension cabes support the spreader bar from
the truck plate, which is hidden from view of the antenna beams by the reflector plate. Both are visible in Figure 5.
c Stainless steel flares surround the sky port. See §2.1.
absolute temperature microwave astrophysical measurements. ARCADE 2 is able to maintain the external calibrator, antennas, and radiometers at temperatures near
2.7 K for many hours at 37 km altitude. Cold boil-off
gas reduces atmospheric condensation to negligible levels. In the future, temperature gradients in the external
calibrator can be greatly reduced by thermally standing
off the aperture plate from its liquid helium tanks and
controlling its temperature with SPID heaters.
We thank the staff at CSBF for launch support. We
thank Victor Kulesh for contributions to the ground software, and Adam Bushmaker, Jane Cornett, Paul Cursey,
Sarah Fixsen, Luke Lowe, and Alexandre Rischard for
their work on the project. We thank the Cryogenics
Branch at GSFC for supporting the ARCADE 2 design and thermometer calibration, Todd Gaier for the
90 GHz amplifiers, and Custom Microwave, Bechdon,
Inc., Flight Fab, and JMD for fabrication of components. This research has been supported by NASAs
Science Mission Directorate under the Astronomy and
Physics Research and Analysis suborbital program. The
research described in this paper was performed in part
at the Jet Propulsion Laboratory, Californai Institute of
Technology, under a contract with the National Aeronautics and Space Administration. T.V. acknowledges
support from CNPq grants 466184/00-0, 305219-2004-9
and 303637/2007-2-FA, and the technical support from
Luiz Reitano. C.A.W. acknowledges support from CNPq
grant 307433/2004-8-FA.
REFERENCES
Fixsen, D.J., & Mather, J. 2002, ApJ, 581, 817
Fixsen, D.J, Mirel, P., Kogut, A., & Seiffert, M. 2002, Rev Sci
Instrum, 73, 3659
Fixsen, D.J., Kogut, A., Levin, S., Limon, M., Lubin, P., Mirel,
P., Seiffert, M., & Wollack, E. 2004, ApJ, 612, 86
Fixsen, D.J., Wollack, E., Kogut, A., Limon, M., Mirel, P., Singal,
J., & Fixsen, S. M. 2006, Rev. Sci. Instrum, 77, No. 064905
Fixsen, D.J. et al. 2010, submitted, arXiv:0901.0559
Johnson D & Wilkinson D, 1987, ApJ, 313, L1
Kogut, A. et al. 2010, submitted, arXiv:0901.0562
Kogut, A., Fixsen, D.J., Levin, S., Limon, M., Lubin, P., Mirel,
P., Seiffert, M., & Wollack, E. 2004a, ApJS, 154, 293
Kogut, A., Wollack, E., Fixsen, D.J., Limon, M., Mirel, P., Levin,
S., Seiffert, M., & Lubin, P. 2004b, Rev. Sci. Instrum, 75, No.
12
Mather, J., Fixsen, D., Shafer, R.A., Mosier, C., & Wilkinson,
D.T. 1999 ApJ, 512, 511
Muehlner, D., & Weiss, R. 1973, Phys. Rev. Lett., 24, 742
Seiffert, M. et al. 2010, submitted, arXiv:0901.0559
Singal, J., Fixsen, D. J., Kogut, A., Limon, M., Lubin, P., Mirel,
P., Seiffert, M., & Wollack, E. 2006, ApJ, 653, 835
Singal, J., Wollack, E., Kogut, A., Limon, M., Lubin, P., Mirel,
P., & Seiffert, M., 2005, Rev. Sci. Instrum, 76, No. 124703
Staggs, S., Jarosic, N.C., Meyer, S., & Wilkinson, D.T. 1996,
ApJ, 473, L1
Wollack, E. 1996, ”TCHEB x: Homogeneous Stepped Waveguide
Transformers,” May 1996, NRAO, EDTN Memo Series # 176.
Wollack, E. J, Fixsen, D. J., Henry, R., Kogut, A., Limon, M., &
Mirel, P. 2008, Int Journal Infrared and Millimeter Waves,
29-1, 51
Wollack, E. J, Fixsen, D. J., Henry, R., Limon, M., Mirel, P. &
Singal, J. 2007, IEEE Trans. Instrum. Meas., 56, 5