This work proposes an improved technique for accurate, contactless measurement of the absolute voltage waveforms of microwave circuits, employing a passive electric-field sensing probe. The proposed technique uses an electromagnetic model of the interaction between the probe and a device under test, to allow the extraction of the coupling capacitance variation versus frequency. Employing these information the measurement accuracy is improved, especially for higher (i.e., harmonic) frequencies, yielding enhanced waveform fidelity. The proposed method is validated on a microstrip line carrying waveforms with rich harmonic content. The accuracy of the proposed technique is benchmarked against a conventional thru-reflect-line (TRL) de-embedding approach by a nonlinear vector network analyzer (NVNA). Measurement results show that the root-mean square (RMS) error can be improved by 3 percentage points (from 8% to 5%) compared to the prior arts over the frequency range from 1 to 5 GHz.@en

This paper introduces an improved nonintrusive near-field technique for in situ characterization of distributed effects in GaN high-power transistors. Compared with previous passive probing approaches which sense electric fields induced by drain bondwires, the proposed method employs the position-signal difference method to measure the E-field fluctuations induced by transistor fingers. This allows a robust and detailed identification of in-circuit electrical quantities, such as voltages, currents, loading impedance, and output power, spatially distributed over individual transistor cells and fingers. The E-fields needed for determining the distributed phenomena have been measured in situ above the fingers of a 100-W GaN power transistor at fundamental and second-harmonic frequencies, while the device operates under realistic loading conditions. The deduced in-circuit quantities are compared with their counterparts from an independently developed distributed in-house model of the same device for validation. The practical value of the proposed method is further demonstrated by uniquely identifying device damage at the finger level (enforced by laser cutting).@en