Signal rectification noise

There are various approaches to the data-reduction task. An often used one consists of computing the modulus of the complex phase-detector signal. This removes all offset imperfections as well as any receiver phase misad-justment, bringing us theoretically to what we would have by summing the outputs of two independent, ideal diode detectors. In this case, however, the original signals are still available and can be used to check various aspects of data quality, carry out additional corrections (such as removal of noise-rectification artifacts), or submitted to alternative evaluation algorithms. 

The performance of various rectification methods is compared for the noise-free underlying signal represented as a uniform distribution, non-stationary stochastic process, and data with deterministic features. 

The two constraints that make direct observation of weak absorption signals impracticable are the presence of pink noise, which contains a preponderance of low-frequency power compared with white noise, and the enhancement of this and other sources of noise by the rectification process through which MMW signals are detected. Both predicate the use of some kind of modulation at a frequency greater than the noise comer frequency and the use of a phase-coherent detector to convert the resulting modulated signal to a DC level suitable for display or for post-detection computer processing. 

It is always advisible to filter out the noise, e.g., noise caused by rectification, by means of a 30 Hz fourth-order analog low-pass filter of the Butterworth type. This is also recommended if analog signals are to be AD converted. 

---