Efficient Dithering Technique With Periodic Waveforms for RF Test and Characterization

Abstract

A dithering technique based on ensemble averaging of waveforms is proposed. In the proposed method, we perform a single measurement of multiple periods of a sine-dithered periodic waveform, whereas in the existing ensemble averaging methods several different measurements with dithering additive noise are performed. The measured waveform is then segmented and the segments are averaged to obtain a noise-reduced post-processed waveform. This technique provides improved efficiency in test and characterization applications where accuracy higher than that of the built-in data converters of the measurement equipment is required. Moreover, we show that by using large dithering signals in the proposed technique, we can reduce the nonideal effects found in real applications. As an example, we show in a simulated example that by using the approach with an ensemble size of five in an ideal quantizer, we are able to reduce the mean-square error and adjacent channel distortion by 11 dB. A reduction of 6 dB is achieved when we use the same ensemble size in a nonideal quantizer. An experiment is then conducted to evaluate the reduction in the adjacent channel power of a modulated signal emanating from an arbitrary waveform, by using the proposed technique. The experimental results show that even for a modulated signal with 160-MHz bandwidth at 4.5 GHz, by using an ensemble size of four, we are able to reduce the adjacent channel distortion by 5 dB.