Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Log inSign up
Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Blog
Job Board
Q3AS

© 2026 Aqora Quantum S.A.S.

TermsPrivacyLegal Notice

IEEE Transactions on Applied Superconductivity

Improvement of a bit error rate measuring system for high-temperature superconducting circuits

Masahiro Horibe, Kazuhiro Shimaoka, Y. Tarutani, K. Tanabe

Published Jun 1, 2003• 3 citations
Publisher

Abstract

We have evaluated and tried to improve the measurement precision of a bit error rate (BER) measuring system for superconducting circuits equipped with a magnetic shield and a cryocooler. A noise model has suggested that the most effective way of improving BER is to reduce the noise figure of the wideband pulse amplifier (WPA). BER measurement using a 3-Gbps, 2/sup 15/-1 pseudo random binary signal showed that by using low-noise WPA (noisefigure=4 dB) a 12.0-mV output voltage is required to obtain a minimum BER of 1.19/spl times/10/sup -15/ (a 10/sup -7/ improvement over the previous system) for a 50 /spl Omega/ standard microstrip. A BER of a La/sub 0.2/Yb/sub 0.9/Ba/sub 1.9/Cu/sub 3/O/sub x//MgO/Au microstrip-line to coplanar-waveguide transition-line converter (MCC) was also measured. It was found that the minimum BER was 4.87/spl times/10/sup -13/ (a 10/sup -5/ improvement over the previous system) when the output voltage was 23.8-mV. However these measured BER values were larger than those expected by the noise model. The cause of the larger measured BER is discussed. A minimum BER of 3.64/spl times/10/sup -55/ in 14.8-mV output voltage for a 50 /spl Omega/ standard microstrip was obtained when the cryo-cooler was not operated (without the noise from a cryo-cooler).

DOI
10.1109/tasc.2003.813887

Related papers

High-speed bit-error-rate measurement system for high-temperature superconducting digital circuits

Masahiro Horibe, Y. Tarutani, K. Tanabe

We have investigated and tried to improve the packaging technologies for a bit-error-rate (BER) measurement system for high-T/sub c/ superconducting circuits. Signal cables, test fixture, and magnetic shields were specially designed and assembled for the BER measurement system, taking into account electrical losses, signal reflection, heat inflow, and power consumption of a cryocooler. BER was further improved by changing the way of connection between the semiconductor amplifiers outside the sample vacuum chamber. These improvements led to a BER less than 10/sup -12/ at an output voltage (V/sub out/) of 1.7 and 2.3 mV for a 50-/spl Omega/ standard microstrip transmission line and a superconducting microstrip-line to coplanar-waveguide transmission-line converter, respectively. The temperature rise of the test fixture was as low as about 2 K. These values are approximately one order of magnitude smaller than those in the previous measurement system and small enough to be achieved by using superconducting interface circuits. Furthermore, 1-Gb/s operation of a superconducting quantum interference device array-type interface circuit was demonstrated in the BER measurement system at 40 K.

IEEE Transactions on Applied Superconductivity•Dec 1, 2003•2 citations

Bit error rate measurement of a high-speed small-voltage signal using a superconducting transmission line

Kazuhiro Shimaoka, Seiichi Tokunaga, Masaaki Nemoto +1 more

We developed a measuring system that has a magnetically shielded wide-band test fixture mounted on a closed-cycle cryo-cooler, and used this system to investigate the feasibility of cryo-packaging technology for high-Tc superconducting digital electronics. In this work, we evaluated the bit-error-rate (BER) performance of a system in which was installed a 50-/spl Omega/ standard microstrip line (MSTL) and a YBa/sub 2/Cu/sub 3/Ox microstrip line to a coplanar waveguide (CPW) transmission line converter (MCC). In the experiment, we used 3-Gbps, 2/sup 15/-1 pseudo-random binary sequence (PRBS) signals. The temperature of the test fixture was 20 K. Results showed that the BER was 3.74/spl times/10/sup -9/ for the 50-/spl Omega/ MSTL when the signal amplitude at the sample was 13.3 mV, and 1.52/spl times/10/sup -9/ for the MCC when the amplitude was 30.8 mV. The effect of the signal loss and impedance mismatch on the BER is also discussed.

IEEE Transactions on Applied Superconductivity•Mar 1, 2001•2 citations

Bit error rate measurements for GHz code generator circuits

E.J. Dean, P.D. Dresseihaus, John X. Przybysz +3 more

A new method of analyzing bit error rates (BERs) for SFQ circuits and their outputs at GHz speeds will be presented. This method was used to test four- and six-bit pseudorandom code generators. The code generators were operated in free-run mode, i.e., without a synchronization (Reset) pulse. This enabled us to conduct BER testing to much lower levels than when the code generator is operated in Reset mode. The SFQ circuit output was amplified with a superconducting output latch to give voltage levels suitable for display on a sampling scope. The GHz output data was analyzed by external computer software; this permitted extended tests without user supervision. With the code generators operating in free-run mode, they routinely gave bit error rates (BERs) in the 10/sup -11/ range; the best BER recorded was 1.1*10/sup -13/ (2 errors in /spl sim/2.5 hours) at 2 GHz. We have successfully tested these code generators at frequencies in excess of 3.5 GHz with BERs in the 10/sup -10/ range as well.

IEEE Transactions on Applied Superconductivity•Jun 1, 1999•3 citations

Determination of the Bit Error Rate Due to Thermal Noise Using JoSIM Superconducting Circuit Simulator and the Monte Carlo Method

Tessa Hall, Johannes A. Delport, Coenrad Johann Fourie

Thermal noise requires careful consideration during the design process of RSFQ circuits since it reduces circuit operating margins and can cause switching errors. Bit error rate (BER) provides a useful means of analysing the effect of this noise on a circuit. With tools developed under the IARPA SuperTools project, we present an implementation of the Monte-Carlo technique for determining BER via simulation. The Monte-Carlo method has been considered previously but not widely implemented due to slow simulation times preventing the determination of small BER values. We use JoSIM – an efficient simulator with SPICE syntax that allows for the simulation of superconducting components and noise sources with improved speed over older superconductor simulators. This speed improvement makes the Monte-Carlo method viable. The simulated BER is plotted against bias current with values in the low-BER region interpolated from the simulated values using MATLAB. We show how the results need to be used as a further qualifier of circuit performance for the characterization of a logic cell library.

IEEE Transactions on Applied Superconductivity•Mar 2, 2023•7 citations

Temperature-dependent bit-error rate of a clocked superconducting digital circuit

Quentin Herr, M.W. Johnson, M. J. Feldman

We measured the bit-error rate (BER) of an RS latch, a clocked SFQ circuit. A digital error-detection circuit was used to detect BER in the range unity to 10/sup -13/; below 10/sup -7/, the circuit was operated with a 12 GHz on-chip clock. BER was measured as a function of control current; both positive and negative control current was applied, leading to two distinct modes of error incidence. The error function curves extrapolate to 10/sup -80/ for optimal control current at a temperature of 5.5 K. Measurements were repeated over the range 3-7 K. Comparison to theoretical error-function estimates of BER indicate that the noise is strictly thermal.

IEEE Transactions on Applied Superconductivity•Jun 1, 1999•28 citations

Measurement of low bit-error-rates of adiabatic quantum-flux-parametron logic using a superconductor voltage driver

Naoki Takeuchi, Hideo Suzuki, Nobuyuki Yoshikawa

Adiabatic quantum-flux-parametron (AQFP) is an energy-efficient superconductor logic. The advantage of AQFP is that the switching energy can be reduced by lowering operation frequencies or by increasing the quality factors of Josephson junctions, while keeping the energy barrier height much larger than thermal energy. In other words, both low energy dissipation and low bit error rates (BERs) can be achieved. In this paper, we report the first measurement results of the low BERs of AQFP logic. We used a superconductor voltage driver with a stack of dc superconducting-quantum-interference-devices to amplify the logic signals of AQFP gates into mV-range voltage signals for the BER measurement. Our measurement results showed 3.3 dB and 2.6 dB operation margins, in which BERs were less than 10−20, for 1 Gbps and 2 Gbps data rates, respectively. While the observed BERs were very low, the estimated switching energy for the 1-Gbps operation was only 2 zJ or 30kBT, where kB is the Boltzmann's constant and T is the temperature. Unlike conventional non-adiabatic logic, BERs are not directly associated with switching energy in AQFP.

Applied Physics Letters•May 15, 2017•27 citations
We have evaluated and tried to improve the measurement precision of a bit error rate (BER) measuring system for superconducting circuits equipped with a magnetic shield and a cryocooler. A noise model has suggested that the most effective way of improving BER is to reduce the noise figure of the wideband pulse amplifier (WPA). BER measurement using a 3-Gbps, 2/sup 15/-1 pseudo random binary signal showed that by using low-noise WPA (noisefigure=4 dB) a 12.0-mV output voltage is required to obtain a minimum BER of 1.19/spl times/10/sup -15/ (a 10/sup -7/ improvement over the previous system) for a 50 /spl Omega/ standard microstrip. A BER of a La/sub 0.2/Yb/sub 0.9/Ba/sub 1.9/Cu/sub 3/O/sub x//MgO/Au microstrip-line to coplanar-waveguide transition-line converter (MCC) was also measured. It was found that the minimum BER was 4.87/spl times/10/sup -13/ (a 10/sup -5/ improvement over the previous system) when the output voltage was 23.8-mV. However these measured BER values were larger than those expected by the noise model. The cause of the larger measured BER is discussed. A minimum BER of 3.64/spl times/10/sup -55/ in 14.8-mV output voltage for a 50 /spl Omega/ standard microstrip was obtained when the cryo-cooler was not operated (without the noise from a cryo-cooler).
We have investigated and tried to improve the packaging technologies for a bit-error-rate (BER) measurement system for high-T/sub c/ superconducting circuits. Signal cables, test fixture, and magnetic shields were specially designed and assembled for the BER measurement system, taking into account electrical losses, signal reflection, heat inflow, and power consumption of a cryocooler. BER was further improved by changing the way of connection between the semiconductor amplifiers outside the sample vacuum chamber. These improvements led to a BER less than 10/sup -12/ at an output voltage (V/sub out/) of 1.7 and 2.3 mV for a 50-/spl Omega/ standard microstrip transmission line and a superconducting microstrip-line to coplanar-waveguide transmission-line converter, respectively. The temperature rise of the test fixture was as low as about 2 K. These values are approximately one order of magnitude smaller than those in the previous measurement system and small enough to be achieved by using superconducting interface circuits. Furthermore, 1-Gb/s operation of a superconducting quantum interference device array-type interface circuit was demonstrated in the BER measurement system at 40 K.
We developed a measuring system that has a magnetically shielded wide-band test fixture mounted on a closed-cycle cryo-cooler, and used this system to investigate the feasibility of cryo-packaging technology for high-Tc superconducting digital electronics. In this work, we evaluated the bit-error-rate (BER) performance of a system in which was installed a 50-/spl Omega/ standard microstrip line (MSTL) and a YBa/sub 2/Cu/sub 3/Ox microstrip line to a coplanar waveguide (CPW) transmission line converter (MCC). In the experiment, we used 3-Gbps, 2/sup 15/-1 pseudo-random binary sequence (PRBS) signals. The temperature of the test fixture was 20 K. Results showed that the BER was 3.74/spl times/10/sup -9/ for the 50-/spl Omega/ MSTL when the signal amplitude at the sample was 13.3 mV, and 1.52/spl times/10/sup -9/ for the MCC when the amplitude was 30.8 mV. The effect of the signal loss and impedance mismatch on the BER is also discussed.
A new method of analyzing bit error rates (BERs) for SFQ circuits and their outputs at GHz speeds will be presented. This method was used to test four- and six-bit pseudorandom code generators. The code generators were operated in free-run mode, i.e., without a synchronization (Reset) pulse. This enabled us to conduct BER testing to much lower levels than when the code generator is operated in Reset mode. The SFQ circuit output was amplified with a superconducting output latch to give voltage levels suitable for display on a sampling scope. The GHz output data was analyzed by external computer software; this permitted extended tests without user supervision. With the code generators operating in free-run mode, they routinely gave bit error rates (BERs) in the 10/sup -11/ range; the best BER recorded was 1.1*10/sup -13/ (2 errors in /spl sim/2.5 hours) at 2 GHz. We have successfully tested these code generators at frequencies in excess of 3.5 GHz with BERs in the 10/sup -10/ range as well.
Thermal noise requires careful consideration during the design process of RSFQ circuits since it reduces circuit operating margins and can cause switching errors. Bit error rate (BER) provides a useful means of analysing the effect of this noise on a circuit. With tools developed under the IARPA SuperTools project, we present an implementation of the Monte-Carlo technique for determining BER via simulation. The Monte-Carlo method has been considered previously but not widely implemented due to slow simulation times preventing the determination of small BER values. We use JoSIM – an efficient simulator with SPICE syntax that allows for the simulation of superconducting components and noise sources with improved speed over older superconductor simulators. This speed improvement makes the Monte-Carlo method viable. The simulated BER is plotted against bias current with values in the low-BER region interpolated from the simulated values using MATLAB. We show how the results need to be used as a further qualifier of circuit performance for the characterization of a logic cell library.
We measured the bit-error rate (BER) of an RS latch, a clocked SFQ circuit. A digital error-detection circuit was used to detect BER in the range unity to 10/sup -13/; below 10/sup -7/, the circuit was operated with a 12 GHz on-chip clock. BER was measured as a function of control current; both positive and negative control current was applied, leading to two distinct modes of error incidence. The error function curves extrapolate to 10/sup -80/ for optimal control current at a temperature of 5.5 K. Measurements were repeated over the range 3-7 K. Comparison to theoretical error-function estimates of BER indicate that the noise is strictly thermal.
Adiabatic quantum-flux-parametron (AQFP) is an energy-efficient superconductor logic. The advantage of AQFP is that the switching energy can be reduced by lowering operation frequencies or by increasing the quality factors of Josephson junctions, while keeping the energy barrier height much larger than thermal energy. In other words, both low energy dissipation and low bit error rates (BERs) can be achieved. In this paper, we report the first measurement results of the low BERs of AQFP logic. We used a superconductor voltage driver with a stack of dc superconducting-quantum-interference-devices to amplify the logic signals of AQFP gates into mV-range voltage signals for the BER measurement. Our measurement results showed 3.3 dB and 2.6 dB operation margins, in which BERs were less than 10−20, for 1 Gbps and 2 Gbps data rates, respectively. While the observed BERs were very low, the estimated switching energy for the 1-Gbps operation was only 2 zJ or 30kBT, where kB is the Boltzmann's constant and T is the temperature. Unlike conventional non-adiabatic logic, BERs are not directly associated with switching energy in AQFP.