A 2.5-V 45-Gb/s decision circuit using SiGe BiCMOS logic

A 45-Gb/s BiCMOS decision circuit operating from a 2.5-V supply is reported. The full-rate retiming flip-flop operates from the lowest supply voltage of any silicon-based flip-flop demonstrated to date at this speed. MOS and SiGe heterojunction-bipolar-transistor (HBT) current-mode logic families are compared. Capitalizing on the best features of both families, a true BiCMOS logic topology is presented that allows for operation from lower supply voltages than pure HBT implementations without compromising speed. The topology, based on a BiCMOS cascode, can also be applied to a number of millimeter-wave (mm-wave) circuits. In addition to the retiming flip-flop, the decision circuit includes a broadband transimpedance preamplifier to improve sensitivity, a tuned 45-GHz clock buffer, and a 50-/spl Omega/ output driver. The first mm-wave transformer is employed along the clock path to perform single-ended-to-differential conversion. The entire circuit, which is implemented in a production 130-nm BiCMOS process with 150-GHz f/sub T/ SiGe HBT, consumes 288 mW from a 2.5-V supply, including only 58 mW from the flip-flop.     

Low-power high-speed InP MISFET direct-coupled FET logic

High-dynamic-range n-channel InP MISFET direct-coupled FET logic ring oscillator and inverter integrated circuits with minimum observed propagation delay per staget_{pd} = 62ps with associated power delay product of 41 fJ and minimum observed power delay productPt_{pd} = 22fJ with associated delay of 84 ps have been fabricated on Fe-doped semi-insulating substrate material using ion implantation for contact and load channel regions and pyrolytic SiO2as the gate insulator. Accumulation-type enhancement-mode MISFET structures with source-drain separations of 1.5 µm and gate metallization lengths of 3.0 µm were employed as driver devices while both MESFET's and 1.5-µm-length ungated "velocity saturation" structures were used as loads. WithV_{DD} = 4.5V representative inverter structures exhibited logic swings of 3.58 V, noise margins of 1.00 and 0.92 V, and dc gain in the linear region of 2.2.     

A 80-gbit/s D-type flip-flop circuit using InP HEMT technology

80-Gbit/s operation of a static D-type flip-flop (D-FF) circuit was achieved using InP-based HEMT technology, which has a cut-off frequency of 245 GHz and a transconductance of 1500 mS/mm. The circuit was designed with differential operation based on source-coupled FET logic (SCFL). To overcome deterioration of the 80-GHz clock signals in a single-ended to differential signal converter in the input buffer, a rat-race circuit was used as a converter. Measurements showed that the circuit achieved a gain of over 2 dB higher than a conventional converter using a differential pair circuit, and power consumption was reduced from 380 to 260 mW. The power supply voltage was -5.7 V, and total power consumption was 1.2 W. Since there is no commercially available 80-Gbit/s-pulse pattern generator, we developed a selector module to measure the D-FF. These measurements showed that the D-FF successfully operated at 80 Gbit/s, which is almost twice the speed reported to date.     

An integrated direct-coupled 10-Gb/s driver for common-cathode VCSELs

An all- npn integrated driver for directly modulating common-cathode vertical-cavity surface-emitting lasers (VCSELs) at high speeds (such as 10 Gb/s) is proposed and experimentally demonstrated. Special biasing techniques allow the output transistors to operate with small collector-emitter voltages while maintaining their fast current-switching capabilities. A current-splitting technique in the output stage minimizes the transients through the bias source and reduces jitter and overshoot.     

An 80-Gb/s optoelectronic delayed flip-flop IC using resonanttunneling diodes and uni-traveling-carrier photodiode

This paper describes an 80-Gb/s optoelectronic delayed flip-flop (D-FF) IC that uses resonant tunneling diodes (RTDs) and a uni-traveling-carrier photodiode (UTC-PD). A circuit design that considers the AC currents passing through RTDs and UTC-PD is key to boosting circuit operation speed. A monolithically fabricated IC operated at 80 Gb/s with a low power dissipation of 7.68 mW. The operation speed of 80 Gb/s is the highest among all reported flip-flops. To clarify the maximum operation speed, we analyze the factors limiting circuit speed. Although the bandwidth of UTC-PD limits the maximum speed of operation to 80 Gb/s at present, the circuit has the potential to offer 100-Gb/s-class operation     

A 2.5-Gb/s 15-mW clock recovery circuit

This paper describes the design of a 2.5-Gb/s 15-mW clock recovery circuit based on the quadricorrelator architecture. Employing both phase and frequency detection, the circuit combines high-speed operations such as differentiation, full-wave rectification, and mixing in one stage to lower the power dissipation. In addition, a two-stage voltage-controlled oscillator is utilized that incorporates both phase shift elements to provide a wide tuning range and isolation techniques to suppress the feedthrough due to input data transitions. Fabricated in a 20-GHz 1-μm BiCMOS technology, the circuit exhibits an rms jitter of 9.5 ps and a capture range of 300 MHz     

A novel high-speed latching operation flip-flop (HLO-FF) circuitand its application to a 19-Gb/s decision circuit using a 0.2-μm GaAsMESFET

This paper describes a novel high-speed flip-flop circuit named the High-speed Latching Operation Flip-Flop (HLO-FF) for GaAs Low-power Source-Coupled FET Logic (LSCFL). We reveal the high-speed operation mechanism of the HLO-FF using newly proposed analytical propagation delay time expressions. A design methodology for series-gated master slave flip-flops and HLO-FF's based on these expressions is also proposed. A SPICE simulation and the fabrication of two decision IC's confirm the accuracy of our analytical method and the high-speed operation of a HLO-FF decision circuit at 19 Gb/s     

Application of Pt/a-Si:H gate GaAs FETs to wide noise margin direct-coupled FET logic

The use of a Pt/a-Si:H gate on GaAs MESFET structures is shown to produce a rectifying gate with lower currents in forward bias, this being applicable to increasing noise margins in direct coupled FET logic schemes. FETs show good DC transconductance, low hysteresis in the current voltage (I/V) characteristics, and the absence of severe drift. This confirms that the approach is not hampered by slow surface states.<>     

A CMOS clock recovery circuit for 2.5-Gb/s NRZ data

This paper describes a phase-locked clock recovery circuit that operates at 2.5 Gb/s in a 0.4-μm digital CMOS technology. To achieve a high speed with low power dissipation, a two-stage ring oscillator is introduced that employs an excess phase technique to operate reliably across a wide range. A sample-and-hold phase detector is also described that combines the advantages of linear and nonlinear phase detectors. The recovered clock exhibits an rms jitter of 10.8 ps for a PRBS sequence of length 2⁷-1 and a phase noise of -80 dBc/Hz at a 5-MHz offset. The core circuit dissipates a total power of 33.5 mW from a 3.3-V supply and occupies an area of 0.8×0.4 mm²     

A 4-Gb/s CMOS clock and data recovery circuit using 1/8-rate clock technique

A 4-Gb/s clock and data recovery (CDR) circuit is realized in a 0.25-/spl mu/m standard CMOS technology. The CDR circuit exploits 1/8-rate clock technique to facilitate the design of a voltage-controlled oscillator (VCO) and to eliminate the need of 1:4 demultiplexer, thereby achieving low power consumption. The VCO incorporates the ring oscillator configuration with active inductor loads, generating four half-quadrature clocks. The VCO control line comprises both a programmable 6-bit digital coarse control and a folded differential fine control through a charge-pump and a low pass filter. Duty-cycle correction of clock signals is obtained by exploiting a high common-mode rejection ratio differential amplifier at the ring oscillator output. A 1/8-rate linear phase detector accomplishes the phase error detection with no systematic phase offset and inherently performs the 1:4 demultiplexing. Test chips demonstrate the jitter of the recovered clock to be 5.2 ps rms and 47 ps pk-pk for 2/sup 31/-1 pseudorandom bit sequence (PRBS) input data. The phase noise is measured to be -112 dBc/Hz at 1-MHz offset. The measured bit error rate is less than 10/sup -6/ for 2/sup 31/-1 PRBS. The chip excluding output buffers dissipates 70 mW from a single 2.5-V supply.     

A redundant multivalued logic for a 10-Gb/s CMOS demultiplexer IC

A redundant multivalued logic is proposed for high-speed communication ICs. In this logic, serial binary data are received and converted into parallel redundant multivalued data. Then they are restored into parallel binary data. Because of the multivalued data conversion, this logic makes it possible to achieve higher operating speeds than that of a conventional binary logic. Using this logic, a 1:4 demultiplexer (DEMUX, serial-parallel converter) IC was fabricated using a 0.18-/spl mu/m CMOS process. The IC achieved an operating speed of 10 Gb/s with a supply voltage of only 1.3 V and with power consumption of 38 mW. This logic may achieve CMOS communication ICs with an operating speed several times greater than 10 Gb/s.     

Complementary Josephson-junction circuit: a fast flip-flop and logic gate

A logic circuit with Josephson junctions has been developed that operates as logic gate or as a flip-flop. Despite the latching-type characteristic of the Josephson tunnel junction, the complementary logic circuit is nonlatching. The test circuit has a power dissipation of 16.4 ?W and a signal risetime of approximately 60 ps has been measured.     

A 20-Gb/s silicon bipolar 1:4-demultiplexer IC

This paper presents a 20-Gb/s 1:4-demultiplexer for future fiber-optic transmission systems. It uses an 0.4-μm emitter double polysilicon 21-GHz f_T Si bipolar foundry process. This is the highest data rate of a 1:4-DEMUX reported so far in any technology. The 1:4-DEMUX features a tree-type architecture with one frequency divider and a channel switch circuit. The circuit design was carefully optimized to achieve high speed and moderate power dissipation. It consumes 1.4 W with a single -4.5-V supply     

A Josephson adder employing high-gain direct-coupled logic gate

A wide-margin adder with a simple configuration employing high-gain direct-coupled logic gates (HDCL's) was studied. A wide-margin half-adder circuit, consisting of a single junction and three HDCL buffer gates, is proposed. In order to obtain a wide-margin circuit, gates were designed to be protective against a noise signal. The experimental circuit fabricated by a conventional Pb alloy Josephson technology with 5-µm minimum line width has shown wide-margin (more than a ± 30-percent bias signal margin) characteristics, as predicted by a computer simulation. This paper also demonstrates that the adder can be simply modified into a wide-margin full adder with a simple configuration by connecting an additional single junction and a buffer gate for a carry signal.     

A 20-Gb/s Adaptive Equalizer in 0.13-$muhbox m$CMOS Technology

An adaptive equalizer incorporates spectrum-balancing technique to achieve high speed and low power dissipation. Obviating the need for slicers, this circuit compares the low and high frequency components of the data spectrum and adjusts the boosting accordingly. Fabricated in 0.13-$muhbox m$CMOS technology, this circuit achieves a data rate of 20 Gb/s while consuming 60mW from a 1.5-V supply.     

A high-speed CMOS circuit for 1.2-Gb/s 16*16 ATM switching

The authors describe a 0.7- mu m CMOS asynchronous transfer mode (ATM) switch circuit of 350 K transistors, the kernel of a fully autonomous 16*16 ATM switching matrix devoted to telecommunications. This matrix is able to switch ATM multiplexes with a throughput of up to 1.2 Gb/s per access line, and was implemented using 16 receiver/transmitter circuits and a control circuit. The architecture of the ATM switch circuit is based on a large embedded and shared dual-access memory. Each chip processes 4-b slices of each incoming multiplex. Seven such chips working in parallel are enough to achieve standard ATM cell switching. Up-to-date test features, such as boundary scan, built-in self-test, and redundancy were implemented in the circuit.<>     

Femto Joule logic circuit with enhancement-type Schottky barrier gate FET

As an approach to an advanced LSI logic, a high-speed and low-power femto-joule logic circuit has been developed by using an enhancement-type Schottky barrier gate FET (ESBT) with³¹P implanted channel layer. A direct coupled transistor logic (DCTL) was designed using ESBT and resistor as a basic logic circuit. To evaluate the dynamic performance of the logic circuit, a 15-stage ring oscillator with an output buffer was integrated on a chip. A power-delay product was found in the femto-joule range. The logic swing is about 0.4 V and typical noise margin is 30 percent of the logic swing. A high-speed (40 ns) and low-power (10 mW) 4 bit ALU has been developed by using DCTL, NOR gates. Furthermore, improving ESBT channel layer carrier profile to the higher carrier concentration and abruptly changing shallower carrier profile by³¹P and¹¹B double implantation resulted in advanced characteristics of ESBT and logic circuit using it as follows. ESBT transconductance was increased by a factor of two. Power-delay product reduced to 80 percent of that of logic circuit, using ESBT with³¹P single implanted channel layer, was satisfactorily confirmed, together with a circuit density as large as 300 gates/ mm².     

A fault detection sensor for circuit aging using double-edge-triggered flip-flop

In nanoscale technology, transistor aging is one of the most critical problems that impact on the reliability of circuits. Aging sensor is a good online way to detect the circuit aging, which performs during the operating time with no influence of the normal operation of circuits. In this paper, a Double-edge-triggered Detection Sensor for circuit Aging (DSDA) is proposed, which employs data signal of logic circuits as its clock to control the sampling process. The simulation is done by Hspice using 45 nm technology. The results show that this technique is not sensitive to the process variations. The worst case of the detection precision is more than 80% under the different process variations. It can detect aging fault effectively with the 8% power cost and 30% performance cost.     

A 25-Gb/s inductorless SiGe BiCMOS receiver for 100-Gb/s optical links

This paper presents the design and measurements of a 25-Gb/s inductorless optical receiver in a 0.25-μm SiGe BiCMOS process for 100-Gb/s (25-Gb/s × 4 lines) Ethernet. As the first stage of the proposed optical receiver, a transimpedance amplifier (TIA) employing a pseudo-differential structure with a feedback resistor incorporates DC offset cancellation (DOC) to enhance the input dynamic range. Cascaded by the improved two-stage limiting amplifiers and a 50-Ω output buffer, the receiver achieves high differential swings. For a bit-error rate (BER) of 10⁻¹² at 25 Gb/s, the measured transimpedance gain, bandwidth, sensitivity, and output swing are 63.17 dBΩ, 20.7 GHz, −10.3 dBm, and 352.7 mV, respectively. The power consumption of the entire receiver is 111.6 mW and the core area of the die is 640 μm × 135 μm.     

A 20-Gb/s Burst-Mode Clock and Data Recovery Circuit Using Injection-Locking Technique

A 20-Gb/s clock and data recovery circuit incorporates injection-locking technique to achieve high-speed operation with low power dissipation. The circuit creates spectral line at the frequency of data rate and injection-locks two cascaded LC oscillators. A frequency-monitoring mechanism is employed to ensure a close matching between the VCO natural frequency and data rate. Fabricated in 90-nm CMOS technology, this circuit achieves a bit error rate of less than 10^-9 in both continuous (PRBS of 2³¹-1) and burst modes while consuming 175 mW from a 1.5-V supply.