Shielded Cable Transfer Impedance Measurements High frequency range

B. Démoulin, L. Koné
TELICE-IEMN Group, Université Lille 1, (France)

Introduction

In our previous paper [1] we described the setup for the measurement of the transfer impedance of shielded coaxial cables. We demonstrated that the conventional triaxial setup does not allow to reach frequencies higher than 100 MHz, mainly due to the onset of propagation phenomena. In fact, for a cable sample of 1 m in length, the maximum frequency attainable by means of a measurement of the near-end crosstalk voltage does not exceed 30 MHz. On the other hand, if we measure the far-end crosstalk voltage, the maximum attainable frequency can grow up to 100 MHz. However, the technology of the transfer impedance bench setup must be
modified, in order to be able to explore the bandwidth between 100 MHz and 1 GHz. It is a known fact that the accuracy produced by a purely computational compensation of propagation phenomena is not acceptable; hence, a more rational approach to the reduction of the propagation effects is the reduction of the physical length of the sample under test. The rule of proportionality with wavelength teaches us that we need to reduce the dimension by a factor of 10 (i.e., adopt samples of 10 cm), in order to reliably measure up to 1 GHz, without being affected by propagation. On the other hand, at higher frequencies, mismatches of the perturbation line
tend to amplify, especially at the line extremities where the signal source and the matched load are connected. The consequence of such defects is to produce an uncertain estimate of the perturbation current, which–in turn–results in a non-negligible error of the transfer impedance estimate. All the above reasons call for an
adjustment of the technology of the transfer impedance bench setup to the extension of the frequency range [2], [3]. The first Section of this article concerns the description of two transfer impedance setups, whose configurations have been devised in particular for reducing the mismatch defects for cable samples of 10 cm in length. We first describe the wire injection method, the construction principle of which resides in a perturbation line made by a thin ribbon conductor glued on the cable
insulating jacket. Then we introduce the shield discontinuity method, by which the sample under test represents the inner conductor of a coaxial cavity terminated by a short circuit. The second Section deals with the calibration of the above setups by

Register to see link

You are not allowed to view links. Register or Login