Implementation and Validation of An Automated Experimental Platform for Vector Network Analysis Based on Software Defined Radio

Artikel i vetenskaplig tidskrift, 2026

Software defined radio (SDR) technology offers a programmable and reconfigurable foundation for custom vector network analyzer (VNA) platforms. Existing SDR-based VNA implementations, however, are often constrained by operating frequency range, sequential frequency-sweep speed, and insufficient characterization of measurement limitations. Motivated by these constraints, this article presents an automated SDR-based experimental platform for broadband transmission response (S21) characterization of passive radio frequency (RF) devices from 100 MHz to 7.2 GHz, implemented on a USRP X410 SDR. The platform uses a Zadoff-Chu (ZC) sequence with correlation processing to estimate the complex frequency response over one instantaneous sub-band from a broadband acquisition record, together with path selection controlled by RF switches and complex transmission response normalization. This approach replaces sequential point-by-point sweeping within one sub-band with a broadband correlation acquisition. However, the frequency spacing is still determined by the ZC sequence length and sampling rate, and random-noise reduction depends on the number of repeated ZC periods used for averaging. The present implementation is validated for passive device S21 measurements through comparison with a commercial VNA. It is not presented as a full substitute for a calibrated two-port VNA, because reflection measurements, full two-port error correction, metrology-grade traceability, and the dynamic range and long-term stability of commercial VNAs remain outside the demonstrated capability of this platform.