Overview

Up until now, we've been discussing the real (mostly analog) world of signals. Now it's time to digitize those signals for use in LabVIEW. By definition, analog signals are continuous-time, continuous-value functions. That means they can take on any possible value and are defined over all possible time resolutions. (By the way, don't think that digital pulses are special; they're just analog signals that happen to be square waves. If you look closely, they have all kinds of ringing, noise, and slew rate limits all the characteristics of analog signals.)

An analog-to-digital converter (ADC) samples your analog signals on a regular basis and converts the amplitude at each sample time to a digital value with finite resolution. These are termed discrete-time, discrete-value functions. Unlike their analog counterparts, discrete functions are defined only at times specified by the sample interval and may only have values determined by the resolution of the ADC. In other words, when you digitize an analog signal, you have to approximate. How much you can throw out depends on your signal and your specifications for data analysis. Is 1 percent resolution acceptable? Or is 0.0001 percent required? And how fine does the temporal resolution need to be? One second? Or 1 ns? Please be realistic. Additional amplitude and temporal resolution can be expensive. To answer these questions, we need to look at this business of sampling more closely.

Sampling Theorem

A fundamental rule of sampled data systems is that the...