Reduction of the Methodological Error in Measuring the Frequency Response of the Impedance of Twisted Pairs

• Published in: Measurement techniques Vol. 65; no. 9; pp. 679 - 685
• Main Authors: Kocherov, A. V., Semenov, A. B., Rudenko, V. I., Smirnova, V. V.
• Format: Journal Article
• Language: English
• Published: New York Springer US 01-12-2022; Springer; Springer Nature B.V
• Subjects: Analytical Chemistry; Cables; Calibration; Characterization and Evaluation of Materials; Error analysis; Frequency response; Hypotheses; Impedance; Measurement; Measurement Science and Instrumentation; Measuring instruments; Modernization; Network analysers; Physical Chemistry; Physics; Physics and Astronomy; Radio Engineering Measurements; Wave resistance; twisted pair; wave resistance; vector network analyzer; LAN cable; impedance; return loss; distributed load
• ISSN: 0543-1972; 1573-8906
• DOI: 10.1007/s11018-023-02139-3

A brief overview is given of the standards required for serial production of LAN cables of various categories. It is shown that for twisted-pair LAN cables category 5e and higher, it is necessary to complete development of the procedure for measurement of key frequency responses (in particular, determining the frequency dependence of the actual value of wave resistance), regulated by the operative GOST R 54429-2011. The reason for the errors of network analyzers are revealed as the effect of the extended cutting section; and the hypothesis that it is impossible to eliminate it within the framework of existing models is substantiated. A hypothesis is formulated about the interference of signals reflected from the beginning of the twisted pair and from the place of its entry into the shell as the primary reason for the distortion of the results of practical measurements. Experimental data confirming the correctness of the proposed assumptions is presented. Modernization of the calibration procedure of the measurement system is proposed, without updating the hardware and software of the vector network analyzer (VNA). The required result is achieved by using a distributed calibration load, individualized for each type of cable, based on a segment of a twisted pair that is manufactured in a production laboratory. It is shown using calculated data and measurement results that the use of such a load reduces from 14 to 3% the methodological measurement error of the impedance of the widespread category 5e cables. The results obtained are invariant to the type of measurement system and can be used in production testing of modern LAN cables of higher categories.