Search

In this work we describe upgrades to the Spider balloon-borne telescope in preparation for its second flight, currently planned for December 2021. The Spider instrument is optimized to search for a primordial B-mode polarization signature in the cosmic microwave background at degree angular scales. During its first flight in 2015, Spider mapped ~10% of the sky at 95 and 150 GHz. The payload for the second Antarctic flight will incorporate three new 280 GHz receivers alongside three refurbished 95- and 150 GHz receivers from Spider's first flight. In this work we discuss the design and characterization of these new receivers, which employ over 1500 feedhorn-coupled transition-edge sensors. We describe pre-flight laboratory measurements of detector properties, and the optical performance of completed receivers. These receivers will map a wide area of the sky at 280 GHz, providing new information on polarized Galactic dust emission that will help to separate it from the cosmological signal., Comment: 13 pages, 8 figures; as published in the conference proceedings for SPIE Millimeter, Submillimeter, and Far-Infrared Detectors and Instrumentation for Astronomy X (2020)

The CCAT-prime project's first light array will be deployed in Mod-Cam, a single-module testbed and first light cryostat, on the Fred Young Submillimeter Telescope (FYST) in Chile's high Atacama desert in late 2022. FYST is a six-meter aperture telescope being built on Cerro Chajnantor at an elevation of 5600 meters to observe at millimeter and submillimeter wavelengths.1 Mod-Cam will pave the way for Prime-Cam, the primary first generation instrument, which will house up to seven instrument modules to simultaneously observe the sky and study a diverse set of science goals from monitoring protostars to probing distant galaxy clusters and characterizing the cosmic microwave background (CMB). At least one feedhorn-coupled array of microwave kinetic inductance detectors (MKIDs) centered on 280 GHz will be included in Mod-Cam at first light, with additional instrument modules to be deployed along with Prime-Cam in stages. The first 280 GHz detector array was fabricated by the Quantum Sensors Group at NIST in Boulder, CO and includes 3,456 polarization-sensitive MKIDs. Current mechanical designs allow for up to three hexagonal arrays to be placed in each single instrument module. We present details on this first light detector array, including mechanical designs and cold readout plans, as well as introducing Mod-Cam as both a testbed and predecessor to Prime-Cam., Comment: 9 pages, 6 figures, SPIE Astronomical Telescopes + Instrumentation 2020, Paper Number: 11453-58

The Simons Observatory (SO) will observe the cosmic microwave background (CMB) from Cerro Toco in the Atacama Desert of Chile. The observatory consists of three 0.5 m Small Aperture Telescopes (SATs) and one 6 m Large Aperture Telescope (LAT), covering six frequency bands centering around 30, 40, 90, 150, 230, and 280 GHz. The SO observations will transform the understanding of our universe by characterizing the properties of the early universe, measuring the number of relativistic species and the mass of neutrinos, improving our understanding of galaxy evolution, and constraining the properties of cosmic reionization. As a critical instrument, the Large Aperture Telescope Receiver (LATR) is designed to cool $\sim$ 60,000 transition-edge sensors (TES) to $<$ 100 mK on a 1.7 m diameter focal plane. The unprecedented scale of the LATR drives a complex design. In this paper, we will first provide an overview of the LATR design. Integration and validation of the LATR design are discussed in detail, including mechanical strength, optical alignment, and cryogenic performance of the five cryogenic stages (80 K, 40 K, 4 K, 1 K, and 100 mK). We will also discuss the microwave-multiplexing ($\mu$Mux) readout system implemented in the LATR and demonstrate the operation of dark prototype TES bolometers. The $\mu$Mux readout technology enables one coaxial loop to read out $\mathcal{O}(10^3)$ TES detectors. Its implementation within the LATR serves as a critical validation for the complex RF chain design. The successful validation of the LATR performance is not only a critical milestone within the Simons Observatory, it also provides a valuable reference for other experiments, e.g. CCAT-prime and CMB-S4., Comment: 20 pages, 12 figures, submitted to the 2020 SPIE Astronomical Telescopes + Instrumentation

The Simons Observatory (SO) will be a cosmic microwave background (CMB) survey experiment with three small-aperture telescopes and one large-aperture telescope, which will observe from the Atacama Desert in Chile. In total, SO will field $\sim$70,000 transition-edge sensor (TES) bolometers in six spectral bands centered between 27 and 280 GHz in order to achieve the sensitivity necessary to measure or constrain numerous cosmological quantities. The SO Universal Focal Plane Modules (UFMs) each contain a 150 mm diameter TES detector array, horn or lenslet optical coupling, cold readout components, and magnetic shielding. SO will use a microwave SQUID multiplexing ($\mu$MUX) readout at an initial multiplexing factor of $\sim$1000; the cold (100 mK) readout components are packaged in a $\mu$MUX readout module, which is part of the UFM, and can also be characterized independently. The 100 mK stage TES bolometer arrays and microwave SQUIDs are sensitive to magnetic fields, and their measured response will vary with the degree to which they are magnetically shielded. We present measurements of the magnetic pickup of test microwave SQUID multiplexers as a study of various shielding configurations for the Simons Observatory. We discuss how these measurements motivated the material choice and design of the UFM magnetic shielding., Comment: 5 pages, 6 figures, conference proceedings submitted to IEEE Transactions on Applied Superconductivity

The BLAST Observatory is a proposed superpressure balloon-borne polarimeter designed for a future ultra-long duration balloon campaign from Wanaka, New Zealand. To maximize scientific output while staying within the stringent superpressure weight envelope, BLAST will feature new 1.8m off-axis optical system contained within a lightweight monocoque structure gondola. The payload will incorporate a 300L $^4$He cryogenic receiver which will cool 8,274 microwave kinetic inductance detectors (MKIDs) to 100mK through the use of an adiabatic demagnetization refrigerator (ADR) in combination with a $^3$He sorption refrigerator all backed by a liquid helium pumped pot operating at 2K. The detector readout utilizes a new Xilinx RFSOC-based system which will run the next-generation of the BLAST-TNG KIDPy software. With this instrument we aim to answer outstanding questions about dust dynamics as well as provide community access to the polarized submillimeter sky made possible by high-altitude observing unrestricted by atmospheric transmission. The BLAST Observatory is designed for a minimum 31-day flight of which 70$\%$ will be dedicated to observations for BLAST scientific goals and the remaining 30$\%$ will be open to proposals from the wider astronomical community through a shared-risk proposals program., Comment: Presented at SPIE Ground-based and Airborne Telescopes VIII, December 13-18, 2020

The Next Generation Balloon-borne Large Aperture Submillimeter Telescope (BLAST-TNG) is a submillimeter polarimeter designed to map interstellar dust and galactic foregrounds at 250, 350, and 500 microns during a 24-day Antarctic flight. The BLAST-TNG detector arrays are comprised of 918, 469, and 272 MKID pixels, respectively. The pixels are formed from two orthogonally oriented, crossed, linear-polarization sensitive MKID antennae. The arrays are cooled to sub 300mK temperatures and stabilized via a closed cycle $^3$He sorption fridge in combination with a $^4$He vacuum pot. The detectors are read out through a combination of the second-generation Reconfigurable Open Architecture Computing Hardware (ROACH2) and custom RF electronics designed for BLAST-TNG. The firmware and software designed to readout and characterize these detectors was built from scratch by the BLAST team around these detectors, and has been adapted for use by other MKID instruments such as TolTEC and OLIMPO. We present an overview of these systems as well as in-depth methodology of the ground-based characterization and the measured in-flight performance., Comment: Presented at SPIE Millimeter, Submillimeter, and Far-Infrared Detectors and Instrumentation for Astronomy X, December 13-18, 2020

The Next Generation Balloon-Borne Large Aperture Submillimeter Telescope (BLAST-TNG) was a unique instrument for characterizing the polarized submillimeter sky at high-angular resolution. BLAST-TNG flew from the Long Duration Balloon Facility in Antarctica in January 2020. Despite the short flight duration, the instrument worked very well and is providing significant information about each subsystem that will be invaluable for future balloon missions. In this contribution, we discuss the performance of telescope and gondola., Comment: Submitted to SPIE Astronomical Telescopes + Instrumentation, Ground-based and Airborne Telescopes VIII

We present a superconducting microresonator thermometer based on two-level systems (TLS) that is drop-in compatible with cryogenic microwave systems. The operational temperature range is 50-1000~mK (which may be extended to 5~mK), and the sensitivity (50-75~$\mu$K/$\sqrt{\mathrm{Hz}}$) is relatively uniform across this range. The miniature footprint that conveniently attaches to the feedline of a cryogenic microwave device facilitates the measurement of on-chip device temperature and requires no additional thermometry wiring or readout electronics. We demonstrate the practical use of these TLS thermometers to investigate static and transient chip heating in a kinetic inductance traveling-wave parametric amplifier operated with a strong pump tone. TLS thermometry may find broad application in cryogenic microwave devices such as superconducting qubits and detectors.

Deep neural networks (DNNs) are state-of-the-art algorithms for multiple applications, spanning from image classification to speech recognition. While providing excellent accuracy, they often have enormous compute and memory requirements. As a result of this, quantized neural networks (QNNs) are increasingly being adopted and deployed especially on embedded devices, thanks to their high accuracy, but also since they have significantly lower compute and memory requirements compared to their floating point equivalents. QNN deployment is also being evaluated for safety-critical applications, such as automotive, avionics, medical or industrial. These systems require functional safety, guaranteeing failure-free behaviour even in the presence of hardware faults. In general fault tolerance can be achieved by adding redundancy to the system, which further exacerbates the overall computational demands and makes it difficult to meet the power and performance requirements. In order to decrease the hardware cost for achieving functional safety, it is vital to explore domain-specific solutions which can exploit the inherent features of DNNs. In this work we present a novel methodology called fault-aware training (FAT), which includes error modeling during neural network (NN) training, to make QNNs resilient to specific fault models on the device. Our experiments show that by injecting faults in the convolutional layers during training, highly accurate convolutional neural networks (CNNs) can be trained which exhibits much better error tolerance compared to the original. Furthermore, we show that redundant systems which are built from QNNs trained with FAT achieve higher worse-case accuracy at lower hardware cost. This has been validated for numerous classification tasks including CIFAR10, GTSRB, SVHN and ImageNet.

Obtaining the magnetic field vector accurately in the solar atmosphere is essential for studying changes in field topology during flares and to reliably model space weather. We tackle this problem by applying various inversion methods to a confined X2.2 flare in NOAA AR 12673 on September 6, 2017, comparing the photospheric and chromospheric magnetic field vector with those from two numerical models of this event. We obtain the photospheric field from Milne-Eddington (ME) and (non-)local thermal equilibrium (non-LTE) inversions of Hinode SOT/SP Fe I 6301.5{\AA} and 6302.5{\AA}. The chromospheric field is obtained from a spatially-regularised weak field approximation (WFA) and non-LTE inversions of Ca II 8542{\AA} observed with CRISP at the Swedish 1-m Solar Telescope. The LTE- and non-LTE-inferred photospheric field components are strongly correlated throughout the atmosphere, with stronger field and higher temperatures in the non-LTE inversions. For the chromospheric field, the non-LTE inversions correlate well with the spatially-regularised WFA. We find strong-field patches of over 4.5 kG in the photosphere, co-located with similar concentrations exceeding 3 kG in the chromosphere. The obtained field strengths are up to 2-3 times higher than in the numerical models, with more concentrated and structured photosphere-to-chromosphere shear close to the polarity inversion line. The LTE and non-LTE Fe I inversions yield essentially the same photospheric field, while ME inversions fail to reproduce the field vector orientation where Fe I is in emission. Our inversions confirm the locations of flux rope footpoints that are predicted by numerical models. However, pre-processing and lower spatial resolution lead to weaker and smoother field in the models than what the data indicate. This emphasises the need for higher spatial resolution in the models to better constrain pre-eruptive flux ropes., Comment: Submitted to Astronomy & Astrophysics. 16 pages, 8 figures

We present a theoretical model and experimental characterization of a microwave kinetic inductance traveling-wave amplifier (KIT), whose noise performance, measured by a shot-noise tunnel junction (SNTJ), approaches the quantum limit. Biased with a dc current, the KIT operates in a three-wave mixing fashion, thereby reducing by several orders of magnitude the power of the microwave pump tone and associated parasitic heating compared to conventional four-wave mixing KIT devices. It consists of a 50 Ohms artificial transmission line whose dispersion allows for a controlled amplification bandwidth. We measure $16.5^{+1}_{-1.3}$ dB of gain across a 2 GHz bandwidth with an input 1 dB compression power of -63 dBm, in qualitative agreement with theory. Using a theoretical framework that accounts for the SNTJ-generated noise entering both the signal and idler ports of the KIT, we measure the system-added noise of an amplification chain that integrates the KIT as the first amplifier. This system-added noise, $3.1\pm0.6$ quanta (equivalent to $0.66\pm0.15$ K) between 3.5 and 5.5 GHz, is the one that a device replacing the SNTJ in that chain would see. This KIT is therefore suitable to read large arrays of microwave kinetic inductance detectors and promising for multiplexed superconducting qubit readout.

The role of magnetic fields in the chromospheric heating problem remains greatly unconstrained. Most theoretical predictions from numerical models rely on a magnetic configuration, field strength and connectivity whose details have not been well established with observational studies. High-resolution studies of chromospheric magnetic fields in plage are very scarce or non-existent in general. Our aim is to study the stratification of the magnetic field vector in plage regions. We use high-spatial resolution full-Stokes observations acquired with CRISP instrument at the Swedish 1-m Solar Telescope in the Mg I $\lambda$5173, Na I $\lambda$5896 and Ca II $\lambda$8542 lines. We have developed a spatially-regularized weak-field approximation (WFA) method based on the idea of spatial regularization. This method allows for a fast computation of magnetic field maps for an extended field of view. The fidelity of this new technique has been assessed using a snapshot from a realistic 3D magnetohydrodynamics simulation. We have derived the depth-stratification of the line-of-sight component of the magnetic field from the photosphere to the chromosphere in a plage region. The magnetic fields are concentrated in the intergranular lanes in the photosphere and expand horizontally toward the chromosphere, filling all the space and forming a canopy. Our results suggest that the lower boundary of this canopy must be located around 400-600 km from the photosphere. The mean canopy total magnetic field strength in the lower chromosphere ($z\approx760$ km) is 658 G. At $z=1160$ km we estimate $\approx 417$ G. We propose a modification to the WFA that improves its applicability to data with worse signal-to-noise ratio. These methods provide a quick and reliable way of studying multi-layer magnetic field observations without the many difficulties inherent to other inversion methods., Comment: Accepted for publication on 2020-08-27

We aim to investigate the temperature enhancements and formation heights of impulsive heating phenomena in solar active-regions such as Ellerman bombs (EBs), ultraviolet bursts (UVBs), and flaring active-region fibrils (FAFs) using interferometric observations in the millimeter (mm) continuum provided by the Atacama Large Millimeter/submillimeter Array (ALMA). We examined 3 mm signatures of heating events identified in Solar Dynamics Observatory (SDO) observations of an active region and compared the results with synthetic spectra from a 3D radiative magnetohydrodynamic simulation. We estimated the contribution from the corona to the mm brightness using differential emission measure analysis. We report the null detection of EBs in the 3 mm continuum at $\sim1.2$" spatial resolution, which is evidence that they are sub-canopy events that do not significantly contribute to heating the upper chromosphere. In contrast, we find the active region to be populated with multiple compact, bright, flickering mm bursts -- reminiscent of UVBs. The high brightness temperatures of up to $\sim14200$ K in some events have a significant contribution (up to $\sim$7%) from the corona. We also detect FAF-like events in the 3 mm continuum that show rapid motions of $>10000\,$K plasma launched with high plane-of-sky velocities ($37-340\rm\,km\,s^{-1}$) from bright kernels. The mm FAFs are the brightest class of warm canopy fibrils that connect magnetic regions of opposite polarities. The simulation confirms that ALMA should be able to detect the mm counterparts of UVBs and small flares and thus provide a complementary diagnostic for localized heating in the solar chromosphere., Comment: revised; accepted in Astronomy & Astrophysics

The performance of superconducting circuits for quantum computing is limited by materials losses. In particular, coherence times are typically bounded by two-level system (TLS) losses at single photon powers and millikelvin temperatures. The identification of low loss fabrication techniques, materials, and thin film dielectrics is critical to achieving scalable architectures for superconducting quantum computing. Superconducting microwave resonators provide a convenient qubit proxy for assessing performance and studying TLS loss and other mechanisms relevant to superconducting circuits such as non-equilibrium quasiparticles and magnetic flux vortices. In this review article, we provide an overview of considerations for designing accurate resonator experiments to characterize loss, including applicable types of loss, cryogenic setup, device design, and methods for extracting material and interface losses, summarizing techniques that have been evolving for over two decades. Results from measurements of a wide variety of materials and processes are also summarized. Lastly, we present recommendations for the reporting of loss data from superconducting microwave resonators to facilitate materials comparisons across the field., Comment: Review Article

Xilinx's AI Engine is a recent industry example of energy-efficient vector processing that includes novel support for 2D SIMD datapaths and shuffle interconnection network. The current approach to programming the AI Engine relies on a C/C++ API for vector intrinsics. While an advance over assembly-level programming, it requires the programmer to specify a number of low-level operations based on detailed knowledge of the hardware. To address these challenges, we introduce Vyasa, a new programming system that extends the Halide DSL compiler to automatically generate code for the AI Engine. We evaluated Vyasa on 36 CONV2D and 6 CONV3D workloads, and achieved geometric means of 7.6 and 23.3 MACs/cycle for 32-bit and 16-bit operands (which represent 95.9% and 72.8% of the peak performance respectively). For 4 of these workloads for which expert-written codes were available to us, Vyasa demonstrated a geometric mean performance improvement of 1.10x with 50x smaller code relative to the expert-written codes.

NASA's Interface Region Imaging Spectrograph (IRIS) provides high resolution observations of the solar atmosphere through UV spectroscopy and imaging. Since the launch of IRIS in June 2013, we have conducted systematic observation campaigns in coordination with the Swedish 1-m Solar Telescope (SST) on La Palma. The SST provides complementary high-resolution observations of the photosphere and chromosphere. The SST observations include spectro-polarimetric imaging in photospheric Fe I lines and spectrally-resolved imaging in the chromospheric Ca II 8542 A, H-alpha, and Ca II K lines. We present a database of co-aligned IRIS and SST datasets that is open for analysis to the scientific community. The database covers a variety of targets including active regions, sunspots, plage, quiet Sun, and coronal holes., Comment: Accepted for publication in Astronomy & Astrophysics (July 21, 2020)

Recent developments of transition-edge sensors (TESs), based on extensive experience in ground-based experiments, have been making the sensor techniques mature enough for their application on future satellite CMB polarization experiments. LiteBIRD is in the most advanced phase among such future satellites, targeting its launch in Japanese Fiscal Year 2027 (2027FY) with JAXA's H3 rocket. It will accommodate more than 4000 TESs in focal planes of reflective low-frequency and refractive medium-and-high-frequency telescopes in order to detect a signature imprinted on the cosmic microwave background (CMB) by the primordial gravitational waves predicted in cosmic inflation. The total wide frequency coverage between 34GHz and 448GHz enables us to extract such weak spiral polarization patterns through the precise subtraction of our Galaxy's foreground emission by using spectral differences among CMB and foreground signals. Telescopes are cooled down to 5Kelvin for suppressing thermal noise and contain polarization modulators with transmissive half-wave plates at individual apertures for separating sky polarization signals from artificial polarization and for mitigating from instrumental 1/f noise. Passive cooling by using V-grooves supports active cooling with mechanical coolers as well as adiabatic demagnetization refrigerators. Sky observations from the second Sun-Earth Lagrangian point, L2, are planned for three years. An international collaboration between Japan, USA, Canada, and Europe is sharing various roles. In May 2019, the Institute of Space and Astronautical Science (ISAS), JAXA selected LiteBIRD as the strategic large mission No. 2., Comment: Journal of Low Temperature Physics, in press

Neural Networks are currently one of the most widely deployed machine learning algorithms. In particular, Convolutional Neural Networks (CNNs), are gaining popularity and are evaluated for deployment in safety critical applications such as self driving vehicles. Modern CNNs feature enormous memory bandwidth and high computational needs, challenging existing hardware platforms to meet throughput, latency and power requirements. Functional safety and error tolerance need to be considered as additional requirement in safety critical systems. In general, fault tolerant operation can be achieved by adding redundancy to the system, which is further exacerbating the computational demands. Furthermore, the question arises whether pruning and quantization methods for performance scaling turn out to be counterproductive with regards to fail safety requirements. In this work we present a methodology to evaluate the impact of permanent faults affecting Quantized Neural Networks (QNNs) and how to effectively decrease their effects in hardware accelerators. We use FPGA-based hardware accelerated error injection, in order to enable the fast evaluation. A detailed analysis is presented showing that QNNs containing convolutional layers are by far not as robust to faults as commonly believed and can lead to accuracy drops of up to 10%. To circumvent that, we propose two different methods to increase their robustness: 1) selective channel replication which adds significantly less redundancy than used by the common triple modular redundancy and 2) a fault-aware scheduling of processing elements for folded implementations, Comment: 6 pages, 5 figures

Advanced ACTPol is the second generation polarization-sensitive upgrade to the $6\, \rm m$ aperture Atacama Cosmology Telescope (ACT), which increased detector count and frequency coverage compared to the previous ACTPol receiver. Advanced ACTPol utilizes a new two-stage time-division multiplexing readout architecture based on superconducting quantum interference devices (SQUIDs) to achieve a multiplexing factor as high as 64 (rows), fielding a 2,012 detector camera at 150/220 GHz and two 90/150 GHz cameras containing 1,716 detectors each. In a time domain system, aliasing introduces noise to the readout. In this work we present a figure of merit to measure this noise contribution and present measurements of the aliased noise fraction of the Advanced ACTPol receiver as deployed.