Novel Sensor Model Calibration Method for Resistively Loaded Diode Detectors

Abstract

Miniaturized radio-frequency electromagnetic near-field broadband probes based on resistively loaded diode detectors have a limited linear dynamic range of $\approx$ 30 dB. Generic linearization schemes have been proposed and applied for continuous wave like or pulse-modulated signals to extend the dynamic range to over 50 dB. Modulation specific linearization has been proposed to also enable precise measurements of current wireless systems with complex modulation schemes. However, calibrations that require an experimental amplitude sweep for each signal have become impractical due to the growing number of wireless communication systems. In this paper, a novel sensor model calibration method has been developed and validated. It is based on calibration of the optimized sensor equivalent circuit model that was derived from the sensor response as a function of bandwidth, duty cycle, modulation schemes, data rate, and statistical distribution. It is shown that all the elements of the equivalent circuit model can be sufficiently accurate determined by the dynamic response to a set of ten generic signals. This model is then used to numerically determine the linearization parameters for any digitized communication signal. The method was tested on various probes for over 200 modulations, resulting in a linearity uncertainty of less than $k$ = 2) for a dynamic range of >50 dB. The proposed method will improve the precision of measurements, reduce calibration costs, increase the flexibility for application of diode-loaded sensors, and enable the use of real-time information for automated probe linearization during or after measurements.