In today’s technology, analog data is typically digitized prior to recording and archiving. Analog signals are converted into individual time slices or steps, called “samples.” Sampling causes a bit of the analog signal to be lost, meaning that the samples may not provide “exact” reproductions of the original signal. Additionally, aliasing effects caused by digitizing techniques can affect the overall quality of the data.

In assessing requirements for data acquisition systems, engineers apply sampling to analog-to-digital converters (ADCs), employing the Nyquist Sampling Theorem, which states that for an analog-to-digital conversion (ADC) to result in a faithful reproduction of a measured signal, the analog waveform must be sampled frequently and at a rate which is at least twice the highest frequency of interest. In general terms, higher sampling rates produce better signal definitions. If you can live with large data files, it is best to sample the data at the fastest possible rate. When sampling the data at 10, 20, or 50 times is not an option, however, limiting the bandwidth of the data becomes a viable option. This means the data must be filtered.

In this paper, we will explain some of the practical reasons why using analog filters in a data acquisition system may be necessary.