In the text of this guide, important terms are included in a glossary where they are explained. (The glossary can be found at our website in the Resource Library or you can request it.)

We make analog filters which discriminate between the components of an electrical input signal according to their frequencies, and alter the amounts of these frequencies present at the output. The ranges of input frequency over which the filter produces substantial output are the passbands, while the ranges that the filter suppresses are called the stopbands. The simplest types of filter are: lowpass filters, whose passbands are at lower frequencies than the stopbands, and highpass filters -which are the other way round!

People doing data acquisition often ask 'Why bother about unwanted frequencies in the input signal? Surely I can remove them from my data with data analysis software if they get in the way?' Well, our advice is always to remember the early computing maxim: 'garbage in, garbage out'. The data acquisition process causes errors which cannot be compensated for by any process applied to the data. Filtering is often needed to alleviate the problems at the front end of the DAQ system.

Analog filters remain a quick, cost effective solution to many signal related problems. Small noisy signals of various types are often the cause of many problems in a complete range of measurement and control. Simply stated, put noise in and you will have noise in your data.

Small Noisy Signals

A combined filter/amplifier from Kemo installed near the transducer or signal source can greatly reduce problems. This can be especially helpful where the signal cables are longer than 10 feet. On the production floor machine control and post-production testing can be greatly enhanced. See our Din Rail Filters as one example of product to help with these applications.


Probably the most important use of our filters concerns aliasing.   Anti-aliasing filters are applied so that when you inspect your set of sample points, you know which input frequency components were present. Otherwise, there are many possible input frequencies (the aliases) all of which can produce the same data points. Anti-aliasing filters (usually, but not always, a lowpass filter) remove all but the wanted input frequency.

Swamping or Dynamic Range

This is another problem that filters overcome. Data acquisition systems have a limited dynamic range; the ratio between the largest and smallest signals they can record. Long before these limits are reached, the ability of the system to discriminate between different components of a mixed input is compromised. Small important signals can be mixed up with large but irrelevant signals, which can cause erroneous data capture. A filtering process can be applied to the data once in digital form, but the wanted signal that remains will be corrupted with residues of the unwanted signal caused by the failure of the acquisition process. Removing unwanted signals which are larger than the wanted ones enables you to 'turn up the volume' on the desired signal and capture it with greater security, whatever acquisition or recording system is being used.

In Depth Filter Notes –a different series of notes - are available on the following list of subjects that may be of interest to the reader: NOTE 1 DYNAMIC RANGE; NOTE 2 WAVEFORM DISTORTION; NOTE 3 ALIASING; NOTE 4 SWITCHED CAPACITOR FILTERS; NOTE 5 SETTLING TIME; NOTE 6 RECONSTRUCTION. You can find these at our website in the RESOURCE LIBRARY.