No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
Thru-Load-Delay: An Improved Technique for Calibrating the Dual Six-Port
Abstract
In some earlier papers the use of a "thru-short-delay" technique for calibrating the dual six-port was described. Another scheme required only a length of precision transmission line and a calibration circuit. The better features of these two somewhat different approaches have now been combined and the requirement for either a known short, or a "calibration circuit" eliminated. This paper will discuss this new procedure and also describe its application in a practical dual six-port system.
Recent advances in millimeter wave technologies, both in component and system design, in line with important size and cost reductions, have opened up new applications in ultra-high-speed wireless communications, radar and imaging sensors. The paper presents the evolution of millimeter wave circuit and modules fabrication and characterization technologies in the past decades. Novel planar low-cost fabrication technologies have been successfully developed in this period. In combination with the standard rectangular wave-guide technology, these offer great opportunities for prototyping and testing of future millimeter wave transceivers or front-ends, which integrate antenna arrays, down-converters, modulators, amplifiers, etc., in a compact fixture. The paper uses, as a suggestive example, the evolution of the multi-port interferometric front-ends implementation from millimeter wave bulky components and systems to miniaturized and high-efficient ones. Circuit and system designs are carefully done to avoid (as much as possible) complicated calibration methods or difficult post-processing of baseband data. This requires an increased effort in design and fabrication, but it allows miniaturization, low-power consumption, while keeping very good overall performances. Useful and straightforward laboratory characterization techniques of circuits and systems are described in detail.
A vector network analyzer (VNA) is a complex and versatile instrument that is used in the field of radio frequency (RF) engineering for accurate and precise measurements. The goal of this chapter is to review briefly the VNA architecture and the measurement errors. The VNA calibration standards and the calibration techniques will also be discussed. This chapter will conclude with an overview of the measurement uncertainty.
A VNA calibration procedure
characterizes and reduces the systematic errors of the instrumentation from the raw measurement
data. The measurement accuracy
mainly depends on the quality of the VNA
calibration. Therefore, the choice of the calibration procedure is very critical. The goal of this chapter is present a comparison of the state of the art 2-port VNA calibration techniques at millimetre frequencies. This comparison will be useful to analyze the efficiency of two different VNA calibration techniques.
This paper attempts to answer a number of questions that arise when using one or more lengths of precision coaxial transmission line to calibrate a dual 6-port automatic network analyzer, questions such as: How important is the quality of the test port relative to that of the line? What type connectors should the line standards have? What are the advantages of using two lines instead of one line and a through connection when test port imperfections are considered? How many lines are optimum from a quality control point of view? What should the lengths be? The answers to these questions appear to be: ¿ The quality of the line is much more important than that of the test port. A perfect line will calibrate out most imperfections in the test port. An example is given where 75-¿ test ports are calibrated with 50-¿ lines, and then used to measure reflection coefficient relative to 50 ¿ with very little error. ¿ Greatest accuracy is achieved with line standards having male connectors. ¿ Two lines get rid of many test port imperfections that one line cannot. Three lines will show up a problem if one line is bad. Five lines will identify which line is bad. Five is probably optimum. ¿ There may not be an optimum for the actual lengths of a set of lines, but there does appear to be an optimum difference in the lengths.
A 2- to 18-GHz dual six-port automatic network analyzer using diode power detectors is described. An analysis of the calibration technique is discussed in detail. Measurement accuracies of better than 0.1-dB up to 40-dB insertion loss from 2 to 7 GHz are reported. Possible causes of the errors observed at higher frequencies are given.
lnitial results of the performance of an experimental dual six-port automatic network analyzer operating in the 2-18-GHz range with thermistor-type power detectors are given. The imprecision in measuring reflection coefficients of one-port devices, or the scattering parameters of two-port devices is 4x10-5, excluding connector repeatability. At 3 GHz, the imprecision in measuring attenuation varies from 0.0003 dB at low values of attenuation to 0.15 dB at 60 dB. The systematic error in measuring attenuation appears to be Iess than the imprecision. The systematic error in measuring reflection coefficient appears to be less than 0.0004. Additional systematic errors caused by changes in the calibration constants over a 20-week period were observed to be less than 0.003 dB in attenuation and less than 0.002 in reflection coefficient.
The development of an explicit and exact algorithm, computing the equivalent error representation for a network analyzer system, makes accurate measurements possible upon non-connectable microwave networks, embedded in complex transmission line environments. Typical examples are measurements at non-standard impedance levels and upon non-TEM waveguide modes.
A new procedure is described, intended for system calibration and systematic error removal in automated S-parameter measurements. The new procedure uses only three non-critical reference standards, called the "Through", the "Short" and the "Delay" and leads to a set of equations for which a fast, non-iterative, explicit solution is known.
In a companion paper, in this digest, a scheme for reducing the (single) six-port calibration problem to that of an equivalent four-port reflectometer has been described. This now makes it possible to apply existing four-port calibration schemes. One such method is the "thru-short-delay" (TSD) procedure. This paper briefly outlines this calibration approach.
This paper describes a new technique for the simultaneous calibration of two six-port networks as reflectometers. Once calibrated, the complete S parameters of an unknown two-port network can be obtained from power measurements alone. The technique requires only a uniform transmission line, of unknown length and loss, and a short circuit as calibrating standards. Its application to an automated microwave measurement system is described and potential error sources are quantitatively discussed.
This paper describes a technique for calibrating a pair of 6-port reflectometers for measuring the complex reflection coefficient of 1-port devices, or the scattering parameters of reciprocal 2-port devices. The operations in the calibration consist of connecting the two 6-ports together, connecting each 6-port to a calibration circuit consisting of two unknown terminations and a leveling loop, and then connecting the standard. The standard can be one termination whose complex impedance is known, or a length of precision transmission line whose cross-sectional dimensions are known. The length and loss of the line are not required. The solution for the constants which characterize each 6-port is closed, requiring no iteration.
lnitial results of the performance of an experimental dual six-port automatic network analyzer operating in the 2-18-GHz range with thermistor-type power detectors are given. The imprecision in measuring reflection coefficients of one-port devices, or the scattering parameters of two-port devices is 4x10-5, excluding connector repeatability. At 3 GHz, the imprecision in measuring attenuation varies from 0.0003 dB at low values of attenuation to 0.15 dB at 60 dB. The systematic error in measuring attenuation appears to be Iess than the imprecision. The systematic error in measuring reflection coefficient appears to be less than 0.0004. Additional systematic errors caused by changes in the calibration constants over a 20-week period were observed to be less than 0.003 dB in attenuation and less than 0.002 in reflection coefficient.