An Open Source RFSoC-based Readout Electronics System for Arrays of Kinetic Inductance Detectors and Superconducting Resonators
Cody Roberson, Nia McNichols, Jack Sayers, Sean Bryan, Daniel Cunnane, Peter K. Day, Fabien Defrance, Simon Hempel-Costello, Tracee Jamison-Hooks, Bradley R. Johnson, Maya Lee, Philip D. Mauskopf, Adrian Sinclair, Liam Walters, Eric Weeks
We present a novel readout electronics system for large arrays of superconducting electromagnetic resonators, such as kinetic inductance detectors (KIDs), based on a Radio Frequency System on Chip (RFSoC) architecture. Each channel in the readout system is designed for frequency division multiplexing of up to 1024 high quality factor resonances placed on a single microwave transmission line at unique frequencies within a 512 MHz bandwidth. We describe the design of the digital and analog signal processing chains for the implementation of a two-channel 2048-resonator system on the Xilinx ZCU111 RFSoC evaluation board in combination with a custom intermediate frequency (IF) system to convert the RFSoC band to higher frequencies up to 4 GHz. We also detail an associated software interface that provides a range of tools commonly utilized for characterizing KID resonators and for operating them as part of a photometric millimeter-wave imager. We additionally provide noise characterizations of the individual readout components, along with the complete readout system in isolation and in operation to readout KID resonators.
We present a novel readout electronics system for large arrays of superconducting electromagnetic resonators, such as kinetic inductance detectors (KIDs), based on a Radio Frequency System on Chip (RFSoC) architecture. Each channel in the readout system is designed for frequency division multiplexing of up to 1024 high quality factor resonances placed on a single microwave transmission line at unique frequencies within a 512 MHz bandwidth. We describe the design of the digital and analog signal processing chains for the implementation of a two-channel 2048-resonator system on the Xilinx ZCU111 RFSoC evaluation board in combination with a custom intermediate frequency (IF) system to convert the RFSoC band to higher frequencies up to 4 GHz. We also detail an associated software interface that provides a range of tools commonly utilized for characterizing KID resonators and for operating them as part of a photometric millimeter-wave imager. We additionally provide noise characterizations of the individual readout components, along with the complete readout system in isolation and in operation to readout KID resonators.
We have developed a flexible radio-frequency readout system suitable for a variety of superconducting detectors commonly used in millimeter and submillimeter astrophysics, including kinetic inductance detectors (KIDs), Thermal KID bolometers, and quantum capacitance detectors. Our system avoids custom FPGA-based readouts and instead uses commercially available software radio hardware for analog to digital converter chip/digital to analog converter chip and a GPU to handle real-time signal processing. Because this system is written in common C++/CUDA, the range of different algorithms that can be quickly implemented make it suitable for the readout of many others cryogenic detectors and for the testing of different and possibly more effective data acquisition schemes.
We present the development of a frequency-domain multiplexing readout of kinetic inductance detectors (KIDs) for pulse signals with a self-trigger system. The KIDs consist of an array of superconducting resonators that have different resonant frequencies individually, allowing us to read out multiple channels in the frequency domain with a single wire using a microwave-frequency comb. The energy deposited to the resonators break Cooper pairs, changing the kinetic inductance and, hence, the amplitude and the phase of the probing microwaves. For some applications such as X-ray detections, the deposited energy is detected as a pulse signal shaped by the time constants of the quasiparticle lifetime, the resonator quality factor, and the ballistic phonon lifetime in the substrate, ranging from microseconds to milliseconds. A readout system commonly used converts the frequency-domain data to the time-domain data. For the short pulse signals, the data rate may exceed the data transfer bandwidth, as the short time constant pulses require us to have a high sampling rate. In order to overcome this circumstance, we have developed a KID readout system that contains a self-trigger system to extract relevant signal data and reduces the total data rate with a commercial off-the-shelf FPGA board. We have demonstrated that the system can read out pulse signals of 15 resonators simultaneously with about 10 Hz event rate by irradiating α particles from 241 Am to the silicon substrate on whose surface aluminum KID resonators are formed.
The frequency comb generator is an essential component in the readout of superconducting detectors. This component is responsible for generating the resonant frequencies necessary to detect photon interaction with the detectors. This component however has not received the much needed attention in terms of readout systems optimization. Conventional implementations rely on either inverse fast Fourier transform (IFFT) engines, which scale poorly in power with increasing frequency resolution requirement but allow flexible tone updates, or lookup tables (LUTs), which demand large memory for high frequency resolution. Microwave Kinetic Inductance Detectors (MKIDs) are superconducting detectors which have gained significant interest for cryogenic astrophysical instrumentation due to their high sensitivity, intrinsic multiplexing capability, and scalable fabrication. However, their simplified detector architecture shifts complexity to the readout electronics, making the development of compact, low-power frequency comb generators especially valuable. The design of the compact frequency comb generator targets applications involving MKIDs which typically detect photons in the submilliter bands. In this thesis, I present the design and hardware realization of a compact frequency comb generator. The design has been verified on an Field Programmable Gate Array (FPGA) using the Xilinx's RFSoC ZCU111 evaluation board. The architecture consumes less than 5% of FPGA resources and less than 5 Watts of power with the capability of handling multiple independent channels of the compact frequency comb generator on the evaluation board. These results demonstrate the potential of this approach to enable scalable, energy-efficient MKID readout systems for large-format detector arrays in next-generation cryogenic experiments.
This report covers work done at Data Processing and Readout System Development Laboratory at Nazarbayev University implementing data processing algorithm for readout electronics. Microwave Kinetic Inductance Detectors (MKIDs) are photon detectors based on superconducting resonators. MKIDs are capable of detecting single photons with very small energy and micro second time resolutions. With proper readout electronic system it is possible to read arrays of thousands MKIDs using single signal line using frequency domain multiplexing (FDM). This report details the work done developing a laboratory setup for measuring MKIDs and developing code for frequency measurements. All the work was done using readout system based on a Re-configurable Open Archite Computing Hardware (ROACH) FPGA board and The Collaboration for Astronomy Signal Processing and Electronics Research (CASPER) software.
This work reports the performance evaluation of an SDR readout system based on the latest generation (Gen3) of the AMD's Radio Frequency System-on-Chip (RFSoC) processing platform, which integrates a full-stack processing system and a powerful FPGA with up to 32 high-speed and high-resolution 14-bit Digital-to-Analog Converters (DACs) and Analog-to-Digital Converters (ADCs). The proposed readout system uses a previously developed multi-band, double-conversion IQ RF-mixing board targeting a multiplexing factor of approximately 1,000 bolometers in a bandwidth between 4 and 8 GHz, in line with state-of-the-art microwave SQUID multiplexers (μMUX). The characterization of the system was performed in two stages, under the conditions typically imposed by the multiplexer and the cold readout circuit. First, in transmission, showing that noise and spurious levels of the generated tones are close to the values imposed by the cold readout. Second, in RF loopback, presenting noise values better than -100 dBc/Hz totally in agreement with the state-of-the-art readout systems. It was demonstrated that the RFSoC Gen3 device is a suitable enabling technology for the next generation of superconducting detector readout systems, reducing system complexity, increasing system integration, and achieving these goals without performance degradation.
The next generation of cryogenic CMB and submillimeter cameras under development require densely instrumented sensor arrays to meet their science goals. The readout of large numbers (~10,000-100,000 per camera) of sub-Kelvin sensors, for instance as proposed for the CMB-S4 experiment, will require substantial improvements in cold and warm readout techniques. To reduce the readout cost per sensor and integration complexity, efforts are presently focused on achieving higher multiplexing density while maintaining readout noise subdominant to intrinsic detector noise and presenting manageable thermal loads. Highly-multiplexed cold readout technologies in active development include Microwave Kinetic Inductance Sensors (MKIDs) and microwave rf-SQUIDs. Both exploit the high quality factors of superconducting microwave resonators to densely channelize sub-Kelvin sensors into the bandwidth of a microwave transmission line. In the case of microwave SQUID multiplexing, arrays of transition-edge sensors (TES) are multiplexed by coupling each TES to its own superconducting microwave resonator through an rf-SQUID. We present advancements in the development of a new warm readout system for microwave SQUID multiplexing, the SLAC Superconducting Microresonator RF electronics, or SMuRF, by adapting SLAC National Accelerator Laboratory's Advanced Telecommunications Computing Architecture (ATCA) FPGA Common Platform. SMuRF aims to read out 4000 microwave SQUID channels between 4 and 8 GHz per RF line. Each compact SMuRF system is built onto a single ATCA carrier blade. Daughter boards on the blade implement RF frequency-division multiplexing using FPGAs, fast DACs and ADCs, and an analog up- and down-conversion chain. The system reads out changes in flux in each resonator-coupled rf-SQUID by monitoring the change in the transmitted amplitude and frequency of RF tones produced at each resonator's fundamental frequency. The SMuRF system is unique in its ability to track each tone, minimizing the total RF power required to readout each resonator, thereby significantly reducing the linearity requirements on the cold and warm readout. Here, we present measurements of the readout noise and linearity of the first full SMuRF system, including a demonstration of closed-loop tone tracking on a 528 channel cryogenic microwave SQUID multiplexer. SMuRF is being explored as a potential readout solution for a number of future CMB projects including Simons Observatory, BICEP Array, CCAT-prime, Ali-CPT, and CMB-S4. In addition, parallel development of the platform is underway to adapt SMuRF to read out both MKID and fast X-ray TES calorimeter arrays.