TABLE OF CONTENTS The process of a standard analog-to-digital (A/D) conversion is normally classified into two distinct operations: sampling and quantization. Either sampling (often accompanied by holding) or quantization can be used to transform a signal with a continuous amplitude into a set of discrete levels.
Sampling is a time-based operation that is typically controlled by a system clock, and it emphasizes capturing the input data at certain discrete moments for the convenience of subsequent operations (e.g., quantization), whereas quantization utilizes a reference (voltage or current) to categorize the input data into discrete groups, regardless of whether the data have been sampled. In other words, sampling acquires and preconditions the input signal, while quantization generates an expression of the input signal using a series of digital bits, based on the result of comparing the input signal's amplitude with the reference. Thus, sampling is often realized by sample-and-hold (S&H) circuits, while quantization is normally done by comparators.
Nevertheless, in some A/D converter (ADC) architectures, the dividing line between sampling and quantization fades out. As a result, sampling and quantization can be done simultaneously and no dedicated S&H circuits would be required. A well-known example is called the flash (or parallel) ADC architecture. As the name suggests, these ADCs operate at a very high speed since they flash out the digital bits in a parallel fashion.
Figures 5.6(a) to Figure 5.6(c) illustrate three different ways to implement a 2-bit flash ADC. All three circuits operate in...
