dard match. Moreover, the individual S A21 (or S A12 ) can be obtained from the expression of S A12 S A21 by estimating their phases [8, 14]. A similar procedure can also be applied to the calculation of the scattering parameters of fixture B. Therefore, the full two-port calibration of asymmetric test fixtures A and B can be done using only one known reflection-calibration standard.
EXPERIMENTAL RESULTSIn order to check the accuracy of the improved TM method, an HP8720D network analyzer was calibrated using the HP85052D coaxial-calibration standards over the frequency range from 50 MHz to 20 GHz. The test ports 1 and 2 of the calibrated network analyzer were male and female, respectively. The reference planes were located in such a way that the measured S-parameter S T21 for the thru standard equals 1 ϩ j0. In order to minimize the random error resulting from repeat connections, three cascaded networks are connected, as shown in Figure 3. The fixtures A and B considered in the analysis were two 6-dB attenuators. The auxiliary fixture C consists of three two-ports, C1, C2, C3, where C1 and C3 are adapters of the HP85052D coaxial calibration kit for converting the port sex of fixture C, and C2 is a 3-dB coaxial attenuator. It should be noted that C2 was reversed in the cascaded networks AC and CB, respectively. The DUT consists of two coax-to-microstrip transition and two sections of 50⍀ transmission lines with a series-chip capacitor embedded in between. The fixtures are reciprocal and asymmetrical, and they can be measured directly using the calibrated network analyzer. Thus, a comparison between our calculated results and the directly measured results can be made. Figure 4 shows the term k obtained using the improved TM method, the TSM method, and the directly measured data. From Figure 4, it can be seen that the improved TM method is better than the TSM method due to the fact that fewer known standards are used in the calibration. Figure 5 shows the scattering parameters of the DUT obtained using the improved TM method and the directly measured data. Figures 4 and 5 show that there is a good agreement between the theory and experiment.
CONCLUSIONAn improved TM calibration method for asymmetric reciprocal test fixtures A and B has been proposed in this paper. The calibration process may be outlined as follows: firstly, two hypothetical symmetrical microwave networks were constructed based on "triple-through" method by means of an auxiliary fixture C. S A22 , S A12 S A21 , S B22 , and S B12 S B21 can be expressed as functions of S A11 and S B11 ; then, with the help of the measurements of test fixtures A and B terminated by the match standard, the scattering parameters S A11 and S B11 of fixtures A and B can be determined; finally, the individual S A21 and S B21 can be obtained from the expressions of S A12 S A21 and S B12 S B21 by estimating their phases. Hence, the full two-port calibration of asymmetric test fixtures A and B can be done.Compared with the other calibration techniques widely used, the improved ...
