Tensor decomposition-based multi-line TRL calibration method

Abstract

The invention discloses a tensor decomposition-based multi-line TRL calibration parameter calculation method. The method comprises the following steps that: synchronous CP tensor decomposition is performed on a plurality of redundant N transmission lines obtained in multi-line TRL calibration on all frequency points; characteristic roots obtained through decomposition are subjected to least squarefitting, so that a propagation constant is obtained; and required port calibration parameters are obtained by means of CP tensor decomposition. According to the technical schemes of the invention, the error of the classic multi-line TRL calibration can be decreased by 5 to 30 times, and therefore, the design requirements of the transmission line frequency coverage of the multi-line TRL calibration are improved. The method is suitable for the measurement calibration of an actual network analyzer or the calibration correction of simulation data.

Description

technical field [0001] The invention relates to a method for calibrating a network analyzer through a transmission line in the radio frequency field. In particular, it relates to an algorithm for universal radio frequency S-parameter calibration by decomposing the measurement results of multiple transmission lines into synchronous tensors to obtain propagation constants and calibration constants. Background technique [0002] In the field of radio frequency, S parameters reflect the frequency response characteristics of the network under test. However, due to the influence of non-ideal networks such as connecting lines, interfaces, adapters, and transmission lines, the measured S-parameter results obtained by vector network analyzers cannot accurately reflect the frequency response characteristics of the network to be tested, especially in the frequency range above GHz. The resulting parasitic and radiation effects can seriously interfere with the measurement results. In t...

AI Technical Summary

Problems solved by technology

However, the measured S-parameter results usually obtained by vector network analyzers cannot accurately reflect the frequency response characteristics of the network under test due to the influence of non-ideal networks such as connecting lines, interfaces, adapters, and transmission lines, especially in the frequency range above GHz. parasitic and radiation effects can seriously interfere with the measurement results

In the current general-purpose multi-line TRL calibration method, each frequency response covering the entire network to be tested is formed by using multiple, redundant transmission lines as the standard network and according to the selected two-two line pairs, and through the two-two line The weight factor in the error analysis model of the pair weights the influence of each line pair to obtain the optimal propagation constant estimation and calibration parameter estimation. The selection of two line pairs will bring repeated weights of associated errors, so that in Boundary jump error on common line selection

Images