James P. Randa

Noise characterization of multiport amplifiers

This paper addresses the issue of the definition and measurement of the noise figure and parameters to characterize multiport devices, particularly differential amplifiers. A parametrization in terms of the noise matrix appears to be the most practical. The noise figure for a given output port is defined and related to the noise matrix and scattering parameters of the device, as well as the correlations between different input noise waves. The degradation of the signal-to-noise ratio is obtained from a special choice of the input correlation function. Two examples are considered in detail: a three-port differential amplifier and a four-port mixed-mode amplifier, both with reflectionless terminations. The noise figures, effective input temperatures, and gains are related to the results of a series of hot-cold measurements, as in the familiar two-port case.

Multiple-source, multiple-frequency error of an electric field meter

Electric field meters (EFMs) are typically calibrated using single-frequency, single-source standard fields. The response to multiple sources or nonsinusoidal time dependence may be different, however. Possible errors in a multiple-source, multiple-frequency environment are analyzed for an EFM consisting of an electrically short dipole antenna with a diode load and a radio-frequency (RF) filter transmission line. Also considered are errors in the assumption of equal electric and magnetic energy densities in a multiple plane-wave environment. Typical errors of field-intensity measurements are about one to 3 dB, but in some circumstances they can exceed 10 dB.

Comparison of adapter characterization methods

We review and compare three different methods for characterization of precision adapters. Two of the methods are one-port techniques using two different reflective terminations in the one case and a matched load and multiple lines with reflective terminations in the other. The third technique is a conventional two-port adapter-removal technique. The intrinsic efficiencies of several different adapters are measured with each technique, and the results are compared. The results usually agree within about 0.005 for efficiencies near one. In all cases, the differences are consistent with the estimated uncertainties of the techniques, which range from about 0.002 to about 0.012, depending on the method, connectors, and frequency.

Errors resulting from the reflectivity of calibration targets

For a microwave total-power radiometer, we consider the error introduced by neglecting the difference in the antenna reflection coefficient between when it views a distant scene and when it views a nearby calibration target. An approximate expression is presented for the error, and measurements are described that enable one to estimate the resulting uncertainty in the measured brightness temperature. The measurement results are presented for several combinations of antenna and calibration target. The resulting uncertainty ranges from about 0.1 K to several kelvins for the representative cases considered.

On-Wafer Measurement of Transistor Noise Parameters at NIST

The National Institute of Standards and Technology has developed the capability to measure noise parameters on a wafer in the 1-12.4-GHz range. The authors describe the measurement method and the uncertainty analysis and present results of measurements on a highly reflective transistor. Typical standard uncertainties are within the range of 20-25 K in Tmin, which is the minimum transistor noise temperature, and about 0.03 in the magnitude of Gammaopt, which is the reflection coefficient for which Tmin occurs

Proposed Development of a National Standard for Microwave Brightness Temperature

We review the advantages of a national standard for microwave brightness temperature and outline our proposed approach toward developing such a standard. The proposal is a combined standard that would comprise both a standard radiometer, traceable to primary noise standards, and a fully characterized standard target. Keywords-brightness temperature; microwave radiometry; radiometer calibration; remote sensing; standards

Precision measurement method for cryogenic amplifier noise temperatures below 5 K

We report precision measurements of the effective input noise temperature of a cryogenic (liquid-helium temperature) monolithic-microwave integrated-circuit amplifier at the amplifier reference planes within the cryostat. A method is given for characterizing and removing the effect of the transmission lines between the amplifier reference planes and the input and output connectors of the cryostat. In conjunction with careful noise measurements, this method enables us to measure amplifier noise temperatures below 5 K with an uncertainty of 0.3 K. The particular amplifier that was measured exhibits a noise temperature below 5.5 K from 1 to 11 GHz, attaining a minimum value of 2.3 K/spl plusmn/0.3 K at 7 GHz. This corresponds to a noise figure of 0.034 dB/spl plusmn/0.004 dB. The measured amplifier gain is between 33.4 dB/spl plusmn/0.3 dB and 35.8 dB/spl plusmn/0.3 dB over the 1-12-GHz range.

Uncertainty Analysis for Noise-Parameter Measurements at NIST

An uncertainty analysis is presented for the National Institute of Standards and Technology (NIST) measurements of the noise parameters of amplifiers and transistors in both connectorized (coaxial) and on-wafer environments. We treat both the X-parameters, which are based on the wave representation of the noise correlation matrix, and the traditional IEEE noise parameters. The type A uncertainties are obtained from the fit that computes the noise parameters from an overdetermined system of equations, and the type B uncertainties are computed by a Monte Carlo program. Some complications that are explicitly discussed include the effect of an output attenuator or probe, physical bounds, and the occurrence of unphysical results. Some sample results are given.

Stability measurements on noise sources

We report results of stability and repeatability measurements performed on a selection of different noise sources for selected frequencies between 12 GHz and 26.5 GHz. Measurements cover intervals classified as intermediate term (about 1 week) and long term (about 1 year or more). Noise sources measured include a commercial diode source, a gas-discharge source constructed by NIST, a specially modified commercial diode source, and a variable-temperature FET-based source. All sources exhibit excellent stability, typically consistent with zero drift in noise temperature within the uncertainty of the tests.

Amplifier noise measurements at NIST

We have recently measured the noise characteristics of two low-noise commercial amplifiers in the 2.0-4.0 GHz frequency range. The tests were part of a program to develop and validate measurement methods for a noise-figure measurement service. Measured noise figures were about 0.5/spl plusmn/0.04 dB. We present the results and the accompanying uncertainties. We also describe the measurement method and summarize the many checks that were used to validate the method.