Two-port Device Research Articles

HBT's RF noise parameter determination by means of an efficient method based on noise analysis of linear amplifier networks

A method for the evaluation of the RF noise figure of heterojunction bipolar transistors (HBTs) is presented. The noise analysis is based on the use of the correlation matrices. The two-port device is described as an interconnection of basic two-port devices whose noise behaviour is known. The circuit theory of linear noisy networks allows that any two-port device can be replaced by a noise equivalent circuit which consists of the original two-port assumed to be noiseless and possess two additional noise sources. The purpose of this paper is to obtain the four noise parameters of the device, taking into account the excess resistances and inductances. The calculations presented show good agreement with measurements, and as a consequence, they permit a good estimation of the noise performance of the structure without neglecting any parasitic elements of the equivalent circuit.

Stripline dual-mode ring resonators and their application to microwave devices

This paper describes the fundamental properties of two orthogonal resonant modes within stripline ring resonators and their application to microwave devices. There are two principal methods for application of ring resonators, the first is by using two-port devices and the second four-port devices, in the case of two-port configurations, dual-mode filters using coupling between two degenerate modes have been studied as typical applications. In this paper, new methods for the coupling of two modes and their applications to practical devices, are proposed, and then experimental results for proof of the new structures are presented. Four-port configurations, which make use of two resonant modes as being independent or having no coupling between them, have a wider range of application to microwave devices. Microwave circuits such as tuned amplifiers and oscillators with novel structures are proposed, and their excellent characteristics are demonstrated. The experimental results obtained in this study on two and four-port devices make it clear that dual-mode ring resonators have great potential for application to various microwave devices.

Novel nine-port network analyser for measuring scattering coefficients of two-port devices

It is known that the dual six-port network analyser is able to derive the unknown scattering coefficients of a two-port device from the readings taken by eight power sensors. The paper outlines the development of a novel nine-port network analyser that requires only six power sensors for the same measurement function. Test results confirm that the prototype instrument is capable of yielding reasonably good measurement accuracies.

Microwave noise characterization of two-port devices using an uncalibrated tuner

A novel noise-parameter and S-parameter measurement system is proposed. Any device under test (DUT) can be characterized using the proposed setup. Such characterization is performed without any preset conditions on the input impedance of the noise-receiver and the repeatability of the impedance tuning mechanism. The DUT is used as a standard for tuner calibration. Measurements were carried out on a general purpose GaAs-MESFET. The extracted transistor noise-parameters are in good agreement with the manufacturer's specified values over the operating frequency band (4-8 GHz).

Evaluation of sequential-in-random-out memory device

The buffering of data is a potential bottleneck to performance. In general, the buffer can be implemented either by FIFO or dual-port RAM, which are both two-port memory devices. In the Letter a classification of buffers is proposed. According to the classification, a new two-port memory device with the sequential-in-random-out feature is introduced. Simulation results show that 45% time, at most, could be saved, as compared to a FIFO buffer.

New technique for measuring thescattering coefficients of two-port devices

A new technique (based on a extension of the six-port reflectometer procedure) is used to measure the scattering coefficients of two-port devices. Only six power sensors are required, as opposed to a total of eight for the dual-reflectometer scheme. Tests confirm that the nine-port instrument based on this technique is able to yield reasonably good measurement accuracies.

Load test on superconducting transformer and fault current limiting devices for electric power system

A superconducting fault current limiting transformer has been investigated to find out its feasibility. In general, the following two types of superconducting fault-current limiters could be expected: noninductive two-port and four-port devices based on superconductive-to-normal state transition. The authors have investigated the transient phenomena in the transformer and then designed a 10 kVA superconducting fault current limiting transformer. First of all, in order to understand performances during quenching in the main coils of the transformer at power system faults, no-load and short-circuit tests in the auxiliary coils were performed. With the same test for the main coils and four windings, exciting conductance, susceptance and leakage impedance were estimated. Under a constant resistive load, actual load tests and sudden short-circuit tests were carried out. Test results are presented. >

Integrated shielding lines on multilayer polysilicon resonators

Abstract In this paper, multilayer polysilicon-silicon nitride beam resonators with integrated conductor and resistor patterns for driving and sensing are presented. In particular, the resonators contain ground lanes between the active signal lines for reducing electrical crosstalk - a major problem in two-port devices. The free-standing polysilicon structures have been fabricated using an improved localized oxidation of silicon (LOCOS) process for the sacrificial oxide, which requires no additional mask. Several different resonator designs are proposed, always using piezoresistive detection while the excitation is either thermal or capacitive. Preliminary frequency-response measurements of thermally excited resonators show promising characteristics; furthermore, the measurements point out the importance of shielding.

Basic Properties of an LC Filter Using a Thin-Film Inductor

This paper discusses the application of thin-film inductors to thin-film-based LC filters. We fabricated a monolithic thin-film LC filter by using rf sputtering and dry etching techniques. The filter consists of thin-film inductors and thin-film capacitors fabricated on a common substrate. This filter was designed as a second-order low-pass filter having Chebyshev characteristics, with a cut-off frequency of 50 MHz. We developed a technique for measuring this two-port device at high frequencies, and then measured the frequency characteristics. Here we discuss the merits of thin-film LC filters using soft magnetic thin films.

An improved calibration technique for on-wafer large-signal transistor characterization

The on-wafer measurement of complex quantities and absolute power levels of active devices is truly significant for nonlinear device characterization and modeling. An original procedure, which allows one to perform both the vector and the power calibrations at the RF wafer probe tips used for on-wafer measurement of two-port devices, is presented. The measurement system is based on an automatic vector network analyzer with coaxial directional couplers and RF coplanar wafer probes. A new error model of the dual directional coupler, which samples the power waves traveling at device output, allows one to take advantage of the coaxial section at the output of the measuring system for calibrating the power level up to the on-wafer probe tips. >

Calibration methods for time domain network analysis

Accurate calibration techniques for the characterization of general one- and two-port networks using time domain reflection/transmission (TDR/T) measurements are presented. Simple one-port open-short-match corrections formulated by W.M. Scott and G.S. Smith (1986) are generalized for three arbitrary known loads and extended to the two-port case. Known, general frequency-domain techniques are shown to be directly applicable to the time-domain measurements, including the use of redundancy to reduce the number of required calibration standards. A time-domain thru-match-short method similar to the TRL method of G.F. Engen and C.A. Hoer (1979) is presented. Examples of the measured results for typical one- and two-port devices are included and compared with vector network analyzer measurement results to validate the time domain network analysis algorithms.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>

Linear dynamic self-heating in SOI MOSFETs

A model for small-signal dynamic self-heating is derived for the general case of a two-port device and then specialized to the case of an SOI MOSFET. The model is fitted to measured data for an SOI MOSFET and shown to accurately describe the frequency dependence of the self-heating. For this device, three time constants of 0.25 mu s, 17 ns, and 90 ps adequately characterize the thermal response, showing that self-heating effects are active over a very wide frequency range. >

薄膜インダクタを用いたＬＣフィルタの基礎特性

This paper discusses the application of thin-film inductors to thin-film-based LC filters. We fabricated a monolithic thin-film LC filter by using the rf sputtering and dry-etching techniques. The filter consists of thin-film inductors and thin-fillm capacitors made on a common substrate. This filter was designed as a second-order low-pass filter having Chebyshev characteristics, with a cut-off frequency of 50 MHz. We developed a technique for measuring this two-port device at high frequencies, and then measured the frequency characteristics. Here, we discuss the merits of thin-film LC filters using soft magnetic thin films.

Invariance of S/sub 21//S/sub 12/ and K-factor under parallel operation of linear two-port devices

Based on a generalized circuit model for parallel-operated amplifiers with linear two-port devices, it has been proved that the S-parameter ratio S/sub 21//S/sub 12/ and hence MSG (maximum stable gain) are invariant as long as the devices have an identical value of S/sub 21//S/sub 12/ and the input and output networks are reciprocal. The invariance of K factor has been shown to hold for two cases: (i) devices are identical and input/output networks are lossless and symmetric with respect to each device, and (ii) identical admittances are added to the networks of case (i) so as to connect every device port with each other. Thus at least in these two cases, MAG (maximum available gain) and U (unilateral gain) are invariant as well as MSG under parallel operation of linear two-port devices. The invariance of S/sub 21//S/sub 12/ and hence MSG applies to a variety of parallel-operated amplifiers such as distributed amplifiers and linear power amplifiers.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>

A general noise de-embedding procedure for packaged two-port linear active devices

A method based on the noise correlation technique and its applications is described. The package, which need not be reciprocal, may consist of an arbitrary interconnection of linear passive elements at thermal equilibrium. Only the terminal admittance properties of the package need be known. However, in certain special cases which lead to singular submatrices of the admittance matrix, the method is inapplicable. This situation can occur when elements such as isolators are part of the package. The necessary theoretical foundation and experimental techniques to enable workers not familiar with the field to assemble the software and laboratory setup for two-port noise de-embedding is provided. The automated noise measurement system used for data acquisition and the mathematical basis for it are described in some detail. The validity of the de-embedding approach is established with extensive experimental data obtained on three MESFETs and a pseudomorphic HEMT.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>

A new method of noise parameter calculation using direct matrix analysis

A generalized matrix approach to noise analysis of multiport networks with arbitrary topology is presented. The network may be composed of any number of passive components and active two-port devices. The method is direct and easy to follow and has been implemented in a computer program. The program is based on admittance matrix formulation and noise current sources of each device and their correlations. For MESFET devices, one can supply the noise parameters of each device or its physical and geometrical parameters for evaluation of the noise sources and their correlations.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>

Three-probe microstrip measuring system for S -parameter measurements

The three-probe scheme for input impedance measurement was modified for measuring the S-parameters of a two-port device. The method is simple, low cost and can be easily scaled to millimetre-wave frequencies. The measured results were verified by an automatic network analyser. >

DC SQUID small signal analysis

A small signal analysis for the DC SQUID (superconducting quantum interference device) is presented. This analysis involves the linearization of the SQUID equations for a given flux and current bias. Considering the SQUID to be a two-port device, an impedance matrix can be determined for a SQUID coupled to a general input circuit. The two intrinsic noise sources referred to the input are computed. One advantage of this analysis over other models of the DC SQUID is that an analytic expression can be determined for the cross-correlation power spectral density of the noise sources. These results should be useful in determining the signal-to-noise ratio of low-noise SQUIDs for a given application. Other differences and similarities of the small signal analysis to other DC SQUID models are discussed.

Characterization of diode detectors used in six-port reflectometers

A method for characterizing in situ the diodes used as power detectors in six-port network analyzers is presented. The method requires only two unknown loads which are connected at the measurement ports. Measurement results obtained for one-port and two-port devices with a dual six-port realized in the 2-18-GHz frequency range are reported. In order to show the influence of frequency, these results are compared to those obtained using a midband characterization of the diode detectors.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>

20 GHz optical analog link using an external modulator

An optical analog link operating at 20 GHz with a 3-dB bandwidth of 5.2 GHz was demonstrated. This link has an intermodulation-free dynamic range of 108 dB-Hz/sup 2/3/ and a maximum signal-to-noise ratio (SNR) of 155 dB-Hz. Previously reported direct modulation analog links have not been able to operate at 20 GHz and, even in the 10-GHz frequency range, have had 20-30-dB-poorer SNRs; externally modulated links are capable of high-frequency operation but have demonstrated 30-50-dB-poorer SNRs than the present link. The link consists of a CW laser, an integrated-optical intensity modulator, and a p-i-n diode detector connected by fibers; it contains no amplifiers. The link performance is evaluated at the RF ports, i.e. it is treated as a two-port RF device. It is noted that by simply using a detector with no rolloff and by reducing the modulator excess optical insertion loss to 3 dB, the RF insertion loss can be improved to -25 dB and the noise figure to 33 dB. >