Correcting for systematic errors in one-port calibration-standards

This paper describes how to correct for systematic errors in one-port vector network analyzer calibrations that are caused by modeling errors in the one-port calibration standards. The paper shows how to correct an open-short-load calibration that was made with a broadband load whose reflection coefficient is not zero. The method can also be used to correct for modeling errors in the open and short reflects if those errors are known. The method is demonstrated by performing an open-short-load calibration with a broadband load whose reflection coefficient is as large as 0.035 at 50 GHz, and then correcting that calibration to produce measurement accuracies comparable to those from a thru-reflect-line calibration.