FREQUENCY DOMAIN EFFECTS OF LOW RESOLUTION DIGITIZATION

Abstract

There is a requirement to digitize certain wide-band analog signals in telemetry applications. Typically, an analog-to-digital converter (ADC) with eight or more bits of resolution is used. The resulting signal requires a much larger transmission bandwidth than the original analog signal. The frequency domain information is of primary interest for many applications. In these cases, there are several methods for minimizing the transmitted bandwidth. One method is to perform fast fourier transforms (FFTs) on the signals and only transmit information about the largest signals. The disadvantages of this approach include: relatively long time delay before transmission, resolution bandwidth fixed when FFT performed (unless phase information is also transmitted), and extra complexity in the telemeter. This paper will discuss some effects of minimizing the transmitted bandwidth by quantizing to a small number of bits. The performance will also be compared with analog frequency modulation (FM). Measured performance will be presented for four different input signals and one-, three-, and eight-bit quantization. These signals are amplitude modulation, angle modulation, sum of sine waves, and frequency sweep. The test setup is shown in figure 1. The analyses presented in this paper were performed using either fast fourier transforms (FFTs) or a Kay DSP Sonagraph. The FFT length was 1024 points and a Hann (cosine) window was used. The analysis hardware used for these tests has an analog input, therefore, all digitized signals were converted to analog signals before analysis. The signals were low pass filtered before analysis to minimize aliasing in the analysis and display process.