Statistical Measurement Techniques for Equivalent Source Mismatch of 1.85 mm Power Splitter

Abstract

The equivalent source mismatch (Gamma <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">G</sub> ) of a 1.85 mm coaxial power splitter was characterized by use of two statistical measurement techniques. The first technique, originally described by Juroshek uses a modified one-port calibration method to determine Gamma <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">G</sub> . The second method uses two-port measurements of the splitter with one of the ports loaded with a series of calibration standards. This second, ldquoindirectrdquo method provides measurements of S-parameters for the three-port device that can subsequently be used to calculate Gamma <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">G</sub> . Measurements of Gamma <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">G</sub> made with the two techniques are in good agreement. This demonstrates that the value of Gamma <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">G</sub> of a splitter can be determined by statistical measurement techniques, thus providing the possibility of exploiting redundant measurements to reduce the effect of random measurement errors. Analysis of repeated measurements of Gamma <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">G</sub> shows that the effect of random measurement errors is lower for the indirect method than for the Juroshek method.