This paper presents a comprehensive study of the Microstrip T-Resonator (MTR) for dielectric characterization of PCB substrates. Due to conflicting assumptions in existing models, we evaluate two MTR configurations: MTR-S, corresponding to the short-substrate configuration, and MTR-E, corresponding to the extended-substrate configuration. Microstrip discontinuities, particularly the open end, significantly affect dielectric characterization. Specifically, the effective length used for relative permittivity extraction relies on an open-end correction derived for extended substrates, making MTR-E the appropriate choice for this parameter. Loss tangent estimation, in turn, depends on radiation loss models that assume the substrate terminates at the open end of the resonator, as in the MTR-S configuration. Thus, this work investigates the MTR methodology and the effects of substrate extension on dielectric characterization accuracy. Relative permittivity and loss tangent are estimated using measured and simulated data, and the results are employed to assess the error and deviation of each configuration. The analysis highlights that aligning device geometry with model assumptions significantly improves estimation accuracy, particularly for loss tangent, which exhibits greater variation with substrate termination.