5.1. Introduction

Sensors in instrumentation systems usually emit analog signals that must be conditioned before they are digitized (see Chapter 1). Analog filtering, which is indispensable in an electronic conditioning device, has two principal functions:

• improving the signal-to-noise ratio;

• eliminating any frequencies that might be aliased by sampling that precedes digital processing.

Analog filters are also used in the output stages of instrumentation systems when it is necessary to reconstruct an analog signal from the conversion of a digital signal.

Filters are complex and expensive mechanisms, with features that are often crucial to the overall performance of a system.

5.2. Technological Constraints

The role of a filter is to separate the useful frequencies in a signal (those that carry information) from unwanted frequencies, such as noise or other signals.

The basic circuit of a filter is the resonator, a mechanism whose performance varies very selectively according to the frequency. A resonator is for the most part modeled by a transfer function of the second order, with parameters that are the resonance frequency f ₀ and the quality coefficient Q. For example:

An efficient instrumentation filter must have the following qualities:

• it should have high quality coefficients Q to get a good frequency discrimination;

• it needs very precise and stable resonance pulsations, according to temperature and time. In particular, imprecise component values must have only a slight repercussion on the values of Q and ? ₀. If this condition is met, we can say...