Written by Robert L. Kay

Published June 1996


Since the advent of inexpensive processing power the general engineering community has had opportunity to apply digital filtering techniques to signal processing applications. Filter functions that were not realizable in analog implementations came into practicality. Filters funtions that were not realizable in analog implementations came into practicality. Filters with very steep cutoffs or very narrow notches could be realized by software implementations. An amazing amount of signal to noise improvement became attainable with systems having spectrally separated and unique signal components. As digital filtering became more and more popular, unique integrated circuit chipsets and processors were developed to provide high speed, compact implementations. Complementary design tools were developed to speed development time and simplify the necessary software coding requirements.

However, this new tool still requires an understanding and some expertise in mathematics, most notably of Fourier and Laplace transforms. The progress in personal computer software tools has reduced the need to manipulate the mathematics symbolically. This makes application of software driven digital filters a straightforward number crunching task in most cases. Yet it is still crucial for the designer to understand the principles at work. The designer should know the basic principles behind filters, FFT's and numberical processes.

Medical products require a broad range of signal processing requirements. Generally speaking, medical analyzers present some of the most interesting and rigorous design challenges in the signal processing arena. Diagnostic devices are based upon instrumentation of one form or another. In the development of the majority of that type of instrumentation, the goal is to maximize signal to noise ratio. Interference from AC power, motors, and such can dramatically reduce the ability of low cost equipment to achieve its best performance. Digital filtering can provide a low cost solution to reducing noise beyond what can be achieved with typical analog solutions. This article provides the reader with a basic introduction to digital filters and provides a simple tool to allow the reader to explore these fascinating tools in greater detail.


In general, the need for filtering arises when the signal of interest is mixed with interfering signals or noise. Almost all noise or sources of interference can be analyzed from the frequency domain perspective. In order to determine which filter type and characteristics are required for any particular application, the spectral characteristics of the signal and its noise or interfering components is necessary. The designer must then define the required output SNR (signal to noise ratio), response time, overshoot and other parameters as may be deemed significant to the application. When that is complete, the filtering requirements can be established and a detailed design can commence.

In the past, analog filters were the primary means to performing the necessary filtering to separate the signal from the noise components. In some cases, the design of multipole analog filters required a number of relatively large components to achieve the required degree of filtering. The resulting filter was susceptible to minor variations in component values and often required additional amplifiers to perform buffering. The stability of these amplifiers required a reasonable degree of analysis which might be offset using cookbook designs and simulation software. Sharp filter functions were expensive and time consuming to implement. Predicting the response of these filters to all variations and scenarios was difficult and time consuming. Amplifiers saturated, inductors saturated, and capacitors exhibited dielectric absorption to name a few of the ill's that analog designers had to deal with. The reader must be fully aware that with digital filters, the basic concepts behind these problems have not changed and they now plague the digital engineer too.

©copyright 1996, Elite Engineering Corp.

To read specific aspects of this article click on the topics below:

Digital Filters - General Form
Computing The Coefficients
Additional Functions That Affect The Digital Filter
Summary And References
Digital Filter Program Download

To send e-mail to the author click here:

Main||Services||About Us||Product/Experience